dataset
stringclasses 6
values | model
stringclasses 13
values | seed
int64 42
46
| n_topics
int64 10
50
| topic_descriptions
sequencelengths 1
50
| runtime_s
float64 0.93
52.3k
| encoder
stringclasses 5
values | diversity
float64 0.09
1
| c_npmi
float64 -0.38
0.21
| wec_ex
float64 0.11
0.49
| wec_in
float64 0.07
0.94
|
|---|---|---|---|---|---|---|---|---|---|---|
ArXiv ML Papers | KeyNMF | 44 | 10 | [
[
"learned",
"training",
"representations",
"knowledge",
"learning",
"machine",
"tasks",
"learn",
"supervised",
"deep"
],
[
"neural",
"network",
"deep",
"training",
"convolutional",
"networks",
"trained",
"layers",
"layer",
"cnns"
],
[
"attack",
"robustness",
"adversarial",
"attacks",
"robust",
"defense",
"security",
"examples",
"training",
"detection"
],
[
"accuracy",
"classifier",
"classifiers",
"features",
"detection",
"classification",
"class",
"supervised",
"feature",
"classes"
],
[
"models",
"data",
"modeling",
"prediction",
"inference",
"model",
"predictions",
"bayesian",
"forecasting",
"predictive"
],
[
"algorithm",
"algorithms",
"gradient",
"complexity",
"optimization",
"descent",
"sampling",
"stochastic",
"convex",
"optimal"
],
[
"accuracy",
"dataset",
"training",
"trained",
"datasets",
"features",
"labeled",
"data",
"feature",
"supervised"
],
[
"reward",
"learning",
"reinforcement",
"agents",
"agent",
"policy",
"tasks",
"control",
"policies",
"rl"
],
[
"image",
"generative",
"gans",
"adversarial",
"generation",
"gan",
"images",
"autoencoder",
"generate",
"variational"
],
[
"node",
"networks",
"network",
"nodes",
"graph",
"graphs",
"embedding",
"embeddings",
"representations",
"clustering"
]
] | 278.633199 | all-MiniLM-L6-v2 | 0.82 | 0.080636 | 0.173903 | 0.776905 |
ArXiv ML Papers | KeyNMF | 45 | 10 | [
[
"learned",
"training",
"representations",
"knowledge",
"learning",
"machine",
"tasks",
"learn",
"supervised",
"deep"
],
[
"networks",
"neural",
"deep",
"network",
"training",
"convolutional",
"trained",
"layer",
"layers",
"cnns"
],
[
"attack",
"robustness",
"adversarial",
"attacks",
"robust",
"defense",
"security",
"examples",
"training",
"detection"
],
[
"accuracy",
"classifier",
"classifiers",
"features",
"detection",
"classification",
"class",
"supervised",
"feature",
"classes"
],
[
"models",
"data",
"modeling",
"prediction",
"inference",
"model",
"predictions",
"bayesian",
"forecasting",
"predictive"
],
[
"algorithm",
"algorithms",
"gradient",
"complexity",
"optimization",
"descent",
"sampling",
"stochastic",
"convex",
"optimal"
],
[
"accuracy",
"dataset",
"training",
"trained",
"datasets",
"features",
"labeled",
"data",
"feature",
"supervised"
],
[
"reward",
"learning",
"reinforcement",
"agents",
"agent",
"policy",
"tasks",
"control",
"policies",
"rl"
],
[
"image",
"generative",
"gans",
"adversarial",
"generation",
"gan",
"images",
"autoencoder",
"generate",
"variational"
],
[
"node",
"networks",
"network",
"nodes",
"graph",
"graphs",
"embedding",
"embeddings",
"representations",
"clustering"
]
] | 306.181129 | all-MiniLM-L6-v2 | 0.82 | 0.080636 | 0.173903 | 0.777719 |
ArXiv ML Papers | KeyNMF | 46 | 10 | [
[
"learned",
"training",
"representations",
"knowledge",
"learning",
"machine",
"tasks",
"learn",
"supervised",
"deep"
],
[
"neural",
"network",
"deep",
"training",
"convolutional",
"networks",
"trained",
"layers",
"layer",
"cnns"
],
[
"attack",
"robustness",
"adversarial",
"attacks",
"robust",
"defense",
"security",
"examples",
"training",
"detection"
],
[
"accuracy",
"classifier",
"classifiers",
"features",
"detection",
"classification",
"class",
"supervised",
"feature",
"classes"
],
[
"models",
"data",
"modeling",
"prediction",
"inference",
"model",
"predictions",
"bayesian",
"forecasting",
"predictive"
],
[
"algorithm",
"algorithms",
"gradient",
"complexity",
"optimization",
"descent",
"sampling",
"stochastic",
"convex",
"optimal"
],
[
"accuracy",
"dataset",
"training",
"trained",
"datasets",
"features",
"labeled",
"data",
"feature",
"supervised"
],
[
"reward",
"learning",
"reinforcement",
"agents",
"agent",
"policy",
"tasks",
"control",
"policies",
"rl"
],
[
"image",
"generative",
"gans",
"adversarial",
"generation",
"gan",
"images",
"autoencoder",
"generate",
"variational"
],
[
"node",
"networks",
"network",
"nodes",
"graph",
"graphs",
"embedding",
"embeddings",
"representations",
"clustering"
]
] | 278.553499 | all-MiniLM-L6-v2 | 0.82 | 0.080636 | 0.173903 | 0.772516 |
ArXiv ML Papers | KeyNMF | 43 | 20 | [
[
"generalization",
"learned",
"learn",
"learns",
"this",
"machine",
"knowledge",
"learning",
"adaptation",
"ensemble"
],
[
"neural",
"layer",
"networks",
"architectures",
"layers",
"network",
"recurrent",
"neurons",
"architecture",
"convolutional"
],
[
"attacks",
"robust",
"detection",
"adversarial",
"attack",
"defense",
"robustness",
"examples",
"security",
"attacker"
],
[
"classifier",
"classification",
"classifiers",
"accuracy",
"class",
"label",
"detection",
"classes",
"feature",
"features"
],
[
"models",
"model",
"modeling",
"forecasting",
"predictions",
"machine",
"predictive",
"forecast",
"language",
"parameters"
],
[
"clustering",
"matrix",
"algorithm",
"algorithms",
"efficient",
"complexity",
"sparse",
"performance",
"clusters",
"search"
],
[
"benchmark",
"datasets",
"features",
"dataset",
"bias",
"imagenet",
"regularization",
"accuracy",
"labels",
"feature"
],
[
"embedding",
"graph",
"network",
"representations",
"networks",
"edges",
"graphs",
"nodes",
"embeddings",
"node"
],
[
"generative",
"images",
"gan",
"adversarial",
"gans",
"image",
"generate",
"networks",
"generation",
"generator"
],
[
"features",
"attention",
"representations",
"feature",
"tasks",
"language",
"recognition",
"task",
"learn",
"learns"
],
[
"agent",
"reward",
"reinforcement",
"learning",
"agents",
"policy",
"control",
"tasks",
"rl",
"policies"
],
[
"cnn",
"image",
"detection",
"convolutional",
"images",
"cnns",
"segmentation",
"vision",
"convolution",
"network"
],
[
"train",
"tasks",
"training",
"trained",
"loss",
"performance",
"transfer",
"network",
"task",
"generalization"
],
[
"optimization",
"convex",
"gradient",
"descent",
"stochastic",
"convergence",
"algorithm",
"regularization",
"minimization",
"sgd"
],
[
"prediction",
"predictive",
"predicting",
"predict",
"performance",
"predictions",
"regression",
"forecasting",
"predicted",
"lstm"
],
[
"supervised",
"self",
"unsupervised",
"clustering",
"labels",
"label",
"regularization",
"labeled",
"unlabeled",
"semi"
],
[
"dnn",
"dnns",
"deep",
"accuracy",
"network",
"image",
"benchmarks",
"learning",
"segmentation",
"mri"
],
[
"information",
"privacy",
"dataset",
"data",
"knowledge",
"statistical",
"dimensional",
"empirical",
"causal",
"private"
],
[
"inference",
"bayesian",
"variational",
"autoencoders",
"autoencoder",
"latent",
"gaussian",
"posterior",
"stochastic",
"prior"
],
[
"optimal",
"bandit",
"exploration",
"stochastic",
"regret",
"sampling",
"bandits",
"reward",
"complexity",
"rewards"
]
] | 304.47764 | all-MiniLM-L6-v2 | 0.76 | 0.006613 | 0.17139 | 0.803993 |
ArXiv ML Papers | KeyNMF | 44 | 20 | [
[
"generalization",
"learned",
"learn",
"learns",
"this",
"machine",
"knowledge",
"learning",
"adaptation",
"ensemble"
],
[
"neural",
"layer",
"networks",
"architectures",
"layers",
"network",
"recurrent",
"neurons",
"architecture",
"convolutional"
],
[
"attacks",
"robust",
"detection",
"adversarial",
"attack",
"defense",
"robustness",
"examples",
"security",
"attacker"
],
[
"classifier",
"classification",
"classifiers",
"accuracy",
"class",
"label",
"detection",
"classes",
"feature",
"features"
],
[
"models",
"model",
"modeling",
"forecasting",
"predictions",
"machine",
"predictive",
"forecast",
"language",
"parameters"
],
[
"clustering",
"matrix",
"algorithm",
"algorithms",
"efficient",
"complexity",
"sparse",
"performance",
"clusters",
"search"
],
[
"benchmark",
"datasets",
"features",
"dataset",
"bias",
"imagenet",
"regularization",
"accuracy",
"labels",
"feature"
],
[
"embedding",
"graph",
"network",
"representations",
"networks",
"edges",
"graphs",
"nodes",
"embeddings",
"node"
],
[
"generative",
"images",
"gan",
"adversarial",
"gans",
"image",
"generate",
"networks",
"generation",
"generator"
],
[
"features",
"attention",
"representations",
"feature",
"tasks",
"language",
"recognition",
"task",
"learn",
"learns"
],
[
"agent",
"reward",
"reinforcement",
"learning",
"agents",
"policy",
"control",
"tasks",
"rl",
"policies"
],
[
"cnn",
"image",
"detection",
"convolutional",
"images",
"cnns",
"segmentation",
"vision",
"convolution",
"network"
],
[
"train",
"tasks",
"training",
"trained",
"loss",
"performance",
"transfer",
"network",
"task",
"generalization"
],
[
"optimization",
"convex",
"gradient",
"descent",
"stochastic",
"convergence",
"algorithm",
"regularization",
"minimization",
"sgd"
],
[
"prediction",
"predictive",
"predicting",
"predict",
"performance",
"predictions",
"regression",
"forecasting",
"predicted",
"lstm"
],
[
"supervised",
"self",
"unsupervised",
"clustering",
"labels",
"label",
"regularization",
"labeled",
"unlabeled",
"semi"
],
[
"dnn",
"dnns",
"deep",
"accuracy",
"network",
"image",
"benchmarks",
"learning",
"segmentation",
"mri"
],
[
"information",
"privacy",
"dataset",
"data",
"knowledge",
"statistical",
"dimensional",
"empirical",
"causal",
"private"
],
[
"inference",
"bayesian",
"variational",
"autoencoders",
"autoencoder",
"latent",
"gaussian",
"posterior",
"stochastic",
"prior"
],
[
"optimal",
"bandit",
"exploration",
"stochastic",
"regret",
"sampling",
"bandits",
"reward",
"complexity",
"rewards"
]
] | 373.919204 | all-MiniLM-L6-v2 | 0.76 | 0.006613 | 0.17139 | 0.800443 |
ArXiv ML Papers | KeyNMF | 45 | 20 | [
[
"generalization",
"learned",
"learn",
"learns",
"this",
"machine",
"knowledge",
"learning",
"adaptation",
"ensemble"
],
[
"neural",
"layer",
"networks",
"architectures",
"layers",
"network",
"recurrent",
"neurons",
"architecture",
"convolutional"
],
[
"attacks",
"robust",
"detection",
"adversarial",
"attack",
"defense",
"robustness",
"examples",
"security",
"attacker"
],
[
"classifier",
"classification",
"classifiers",
"accuracy",
"class",
"label",
"detection",
"classes",
"feature",
"features"
],
[
"models",
"model",
"modeling",
"forecasting",
"predictions",
"machine",
"predictive",
"forecast",
"language",
"parameters"
],
[
"clustering",
"matrix",
"algorithm",
"algorithms",
"efficient",
"complexity",
"sparse",
"performance",
"clusters",
"search"
],
[
"benchmark",
"datasets",
"features",
"dataset",
"bias",
"imagenet",
"regularization",
"accuracy",
"labels",
"feature"
],
[
"embedding",
"graph",
"network",
"representations",
"networks",
"edges",
"graphs",
"nodes",
"embeddings",
"node"
],
[
"generative",
"images",
"gan",
"adversarial",
"gans",
"image",
"generate",
"networks",
"generation",
"generator"
],
[
"features",
"attention",
"representations",
"feature",
"tasks",
"language",
"recognition",
"task",
"learn",
"learns"
],
[
"agent",
"reward",
"reinforcement",
"learning",
"agents",
"policy",
"control",
"tasks",
"rl",
"policies"
],
[
"cnn",
"image",
"detection",
"convolutional",
"images",
"cnns",
"segmentation",
"vision",
"convolution",
"network"
],
[
"train",
"tasks",
"training",
"trained",
"loss",
"performance",
"transfer",
"network",
"task",
"generalization"
],
[
"optimization",
"convex",
"gradient",
"descent",
"stochastic",
"convergence",
"algorithm",
"regularization",
"minimization",
"sgd"
],
[
"prediction",
"predictive",
"predicting",
"predict",
"performance",
"predictions",
"regression",
"forecasting",
"predicted",
"lstm"
],
[
"supervised",
"self",
"unsupervised",
"clustering",
"labels",
"label",
"regularization",
"labeled",
"unlabeled",
"semi"
],
[
"dnn",
"dnns",
"deep",
"accuracy",
"network",
"image",
"benchmarks",
"learning",
"segmentation",
"mri"
],
[
"information",
"privacy",
"dataset",
"data",
"knowledge",
"statistical",
"dimensional",
"empirical",
"causal",
"private"
],
[
"inference",
"bayesian",
"variational",
"autoencoders",
"autoencoder",
"latent",
"gaussian",
"posterior",
"stochastic",
"prior"
],
[
"optimal",
"bandit",
"regret",
"sampling",
"complexity",
"reward",
"bandits",
"exploration",
"stochastic",
"rewards"
]
] | 376.953101 | all-MiniLM-L6-v2 | 0.76 | 0.006613 | 0.17139 | 0.804307 |
ArXiv ML Papers | KeyNMF | 46 | 20 | [
[
"generalization",
"learned",
"learn",
"learns",
"this",
"machine",
"knowledge",
"learning",
"adaptation",
"ensemble"
],
[
"neural",
"layer",
"networks",
"architectures",
"layers",
"network",
"recurrent",
"neurons",
"architecture",
"convolutional"
],
[
"attacks",
"robust",
"detection",
"adversarial",
"attack",
"defense",
"robustness",
"examples",
"security",
"attacker"
],
[
"classifier",
"classification",
"classifiers",
"accuracy",
"class",
"label",
"detection",
"classes",
"feature",
"features"
],
[
"models",
"model",
"modeling",
"forecasting",
"predictions",
"machine",
"predictive",
"forecast",
"language",
"parameters"
],
[
"clustering",
"matrix",
"algorithm",
"algorithms",
"efficient",
"complexity",
"sparse",
"performance",
"clusters",
"search"
],
[
"benchmark",
"datasets",
"features",
"dataset",
"bias",
"imagenet",
"regularization",
"accuracy",
"labels",
"feature"
],
[
"embedding",
"graph",
"network",
"representations",
"networks",
"edges",
"graphs",
"nodes",
"embeddings",
"node"
],
[
"generative",
"images",
"gan",
"adversarial",
"gans",
"image",
"generate",
"networks",
"generation",
"generator"
],
[
"features",
"attention",
"representations",
"feature",
"tasks",
"language",
"recognition",
"task",
"learn",
"learns"
],
[
"agent",
"reward",
"reinforcement",
"learning",
"agents",
"policy",
"control",
"tasks",
"rl",
"policies"
],
[
"cnn",
"image",
"detection",
"convolutional",
"images",
"cnns",
"segmentation",
"vision",
"convolution",
"network"
],
[
"train",
"tasks",
"training",
"trained",
"loss",
"performance",
"transfer",
"network",
"task",
"generalization"
],
[
"optimization",
"convex",
"gradient",
"descent",
"stochastic",
"convergence",
"algorithm",
"regularization",
"minimization",
"sgd"
],
[
"prediction",
"predictive",
"predicting",
"predict",
"performance",
"predictions",
"regression",
"forecasting",
"predicted",
"lstm"
],
[
"supervised",
"self",
"unsupervised",
"clustering",
"labels",
"label",
"regularization",
"labeled",
"unlabeled",
"semi"
],
[
"dnn",
"dnns",
"deep",
"accuracy",
"network",
"image",
"benchmarks",
"learning",
"segmentation",
"mri"
],
[
"information",
"privacy",
"dataset",
"data",
"knowledge",
"statistical",
"dimensional",
"empirical",
"causal",
"private"
],
[
"inference",
"bayesian",
"variational",
"autoencoders",
"autoencoder",
"latent",
"gaussian",
"posterior",
"stochastic",
"prior"
],
[
"optimal",
"bandit",
"exploration",
"stochastic",
"regret",
"sampling",
"bandits",
"reward",
"complexity",
"rewards"
]
] | 300.56525 | all-MiniLM-L6-v2 | 0.76 | 0.006613 | 0.17139 | 0.798926 |
ArXiv ML Papers | KeyNMF | 43 | 30 | [
[
"learn",
"generalization",
"learning",
"machine",
"learns",
"knowledge",
"learned",
"this",
"adaptation",
"ensemble"
],
[
"architectures",
"recurrent",
"layer",
"network",
"neural",
"networks",
"neurons",
"input",
"layers",
"neuron"
],
[
"robust",
"robustness",
"adversarial",
"defense",
"attacks",
"attack",
"examples",
"security",
"attacker",
"detection"
],
[
"classes",
"classification",
"class",
"ensemble",
"accuracy",
"label",
"recognition",
"classify",
"benchmark",
"detection"
],
[
"models",
"predictions",
"model",
"forecasting",
"modeling",
"machine",
"predictive",
"forecast",
"parameters",
"systems"
],
[
"algorithm",
"algorithms",
"search",
"methods",
"complexity",
"efficient",
"matrix",
"efficiently",
"minimization",
"adaptive"
],
[
"datasets",
"benchmark",
"tensor",
"overfitting",
"labels",
"label",
"regularization",
"bias",
"imagenet",
"labeled"
],
[
"networks",
"graphs",
"nodes",
"graph",
"network",
"representations",
"edges",
"node",
"embeddings",
"embedding"
],
[
"gan",
"gans",
"generative",
"adversarial",
"generation",
"generate",
"networks",
"generating",
"generator",
"generated"
],
[
"attention",
"tasks",
"representations",
"language",
"task",
"learn",
"learns",
"embeddings",
"text",
"representation"
],
[
"reinforcement",
"agent",
"agents",
"policy",
"reward",
"learning",
"control",
"rl",
"policies",
"action"
],
[
"cnns",
"filters",
"network",
"convolutional",
"convolution",
"cnn",
"imagenet",
"layers",
"recognition",
"model"
],
[
"examples",
"train",
"loss",
"generalization",
"tasks",
"training",
"transfer",
"normalization",
"weights",
"federated"
],
[
"convex",
"gradient",
"minimization",
"stochastic",
"sgd",
"descent",
"optimization",
"convergence",
"constraints",
"algorithm"
],
[
"regression",
"forecasting",
"predicted",
"predictive",
"prediction",
"predictions",
"predict",
"predicting",
"forecast",
"forecasts"
],
[
"supervised",
"unlabeled",
"labeled",
"self",
"labels",
"label",
"regularization",
"labeling",
"unsupervised",
"semi"
],
[
"dnn",
"accuracy",
"dnns",
"deep",
"network",
"learning",
"mri",
"benchmarks",
"crucial",
"medical"
],
[
"information",
"data",
"knowledge",
"statistical",
"dimensional",
"causal",
"empirical",
"health",
"temporal",
"labeled"
],
[
"visual",
"attention",
"selection",
"features",
"feature",
"interpretability",
"importance",
"explanations",
"dimensionality",
"tree"
],
[
"reward",
"regret",
"optimal",
"exploration",
"bandits",
"sampling",
"bandit",
"stochastic",
"rewards",
"strategy"
],
[
"variational",
"autoencoder",
"autoencoders",
"latent",
"generative",
"decoder",
"encoder",
"representation",
"vae",
"model"
],
[
"detection",
"accuracy",
"dataset",
"traffic",
"challenge",
"results",
"detecting",
"detect",
"mnist",
"identification"
],
[
"object",
"images",
"image",
"visual",
"vision",
"detection",
"segmentation",
"imaging",
"objects",
"3d"
],
[
"memory",
"compression",
"efficient",
"network",
"performance",
"quantization",
"accuracy",
"imagenet",
"hardware",
"lstm"
],
[
"trained",
"knowledge",
"task",
"distillation",
"layer",
"network",
"input",
"train",
"ai",
"tasks"
],
[
"regularization",
"sparse",
"matrix",
"tensor",
"sparsity",
"dimensional",
"factorization",
"linear",
"gaussian",
"thresholding"
],
[
"differentially",
"communication",
"private",
"decentralized",
"distributed",
"federated",
"complexity",
"privacy",
"information",
"differential"
],
[
"accuracy",
"classifier",
"classifiers",
"detection",
"svm",
"classes",
"class",
"classifying",
"fairness",
"performance"
],
[
"prior",
"stochastic",
"gaussian",
"bayesian",
"inference",
"posterior",
"sampling",
"estimation",
"variational",
"likelihood"
],
[
"cluster",
"clusters",
"algorithm",
"clustering",
"unsupervised",
"instances",
"analysis",
"patterns",
"hierarchical",
"segmentation"
]
] | 331.399801 | all-MiniLM-L6-v2 | 0.743333 | -0.026298 | 0.178553 | 0.824749 |
ArXiv ML Papers | KeyNMF | 44 | 30 | [
[
"learning",
"machine",
"this",
"learn",
"knowledge",
"learned",
"ml",
"computational",
"generalization",
"adaptation"
],
[
"networks",
"neural",
"network",
"layer",
"layers",
"architectures",
"neurons",
"neuron",
"input",
"weights"
],
[
"attack",
"adversarial",
"robustness",
"robust",
"attacks",
"defense",
"attacker",
"examples",
"adversarially",
"security"
],
[
"accuracy",
"classification",
"classes",
"ensemble",
"labels",
"label",
"classify",
"class",
"recognition",
"benchmark"
],
[
"models",
"machine",
"predictions",
"modeling",
"predictive",
"forecasting",
"forecast",
"model",
"parameters",
"probabilistic"
],
[
"algorithm",
"complexity",
"efficiently",
"matrix",
"performance",
"efficient",
"search",
"algorithms",
"methods",
"computing"
],
[
"imagenet",
"bias",
"overfitting",
"sampling",
"labels",
"benchmark",
"regularization",
"datasets",
"label",
"labeled"
],
[
"graphs",
"networks",
"graph",
"nodes",
"embedding",
"node",
"network",
"edges",
"embeddings",
"gnns"
],
[
"gans",
"generation",
"gan",
"image",
"adversarial",
"images",
"generative",
"generate",
"networks",
"generator"
],
[
"learn",
"task",
"knowledge",
"tasks",
"learned",
"ai",
"robot",
"human",
"language",
"forgetting"
],
[
"policy",
"policies",
"reinforcement",
"control",
"action",
"learning",
"rl",
"agent",
"reward",
"agents"
],
[
"image",
"segmentation",
"cnns",
"convolutional",
"cnn",
"images",
"convolution",
"imagenet",
"vision",
"layers"
],
[
"training",
"trained",
"train",
"loss",
"network",
"generalization",
"transfer",
"performance",
"weights",
"imagenet"
],
[
"convergence",
"stochastic",
"gradient",
"descent",
"sgd",
"batch",
"generalization",
"convex",
"regularization",
"boosting"
],
[
"predict",
"forecasting",
"predictions",
"performance",
"predictive",
"regression",
"predicted",
"lstm",
"predicting",
"prediction"
],
[
"supervised",
"self",
"unsupervised",
"labeled",
"labels",
"label",
"unlabeled",
"semi",
"labeling",
"regularization"
],
[
"deep",
"dnn",
"dnns",
"network",
"accuracy",
"image",
"segmentation",
"benchmarks",
"mri",
"learning"
],
[
"information",
"knowledge",
"statistical",
"data",
"causal",
"dimensional",
"temporal",
"health",
"empirical",
"samples"
],
[
"feature",
"features",
"selection",
"visual",
"interpretability",
"importance",
"attention",
"explanations",
"dimensionality",
"tree"
],
[
"optimal",
"bandit",
"sampling",
"regret",
"bandits",
"reward",
"exploration",
"stochastic",
"rewards",
"complexity"
],
[
"anomaly",
"detect",
"detecting",
"attacks",
"detection",
"object",
"image",
"images",
"intrusion",
"outlier"
],
[
"dataset",
"accuracy",
"images",
"trained",
"challenge",
"segmentation",
"image",
"results",
"bias",
"traffic"
],
[
"optimization",
"objective",
"algorithm",
"constraints",
"convex",
"minimization",
"optimizing",
"optimal",
"optimize",
"search"
],
[
"language",
"encoder",
"recurrent",
"attention",
"recognition",
"speech",
"memory",
"text",
"decoder",
"rnn"
],
[
"semantic",
"learn",
"embeddings",
"visual",
"learned",
"objects",
"representations",
"representation",
"embedding",
"learns"
],
[
"matrix",
"factorization",
"sparsity",
"tensor",
"gaussian",
"dimensional",
"regularization",
"sparse",
"linear",
"thresholding"
],
[
"complexity",
"federated",
"distributed",
"privacy",
"private",
"information",
"decentralized",
"communication",
"differentially",
"differential"
],
[
"accuracy",
"class",
"classifiers",
"classifier",
"svm",
"classes",
"classifying",
"performance",
"fairness",
"labels"
],
[
"variational",
"inference",
"bayesian",
"autoencoders",
"posterior",
"latent",
"likelihood",
"autoencoder",
"gaussian",
"prior"
],
[
"algorithm",
"unsupervised",
"clustering",
"cluster",
"analysis",
"hierarchical",
"clusters",
"instances",
"patterns",
"segmentation"
]
] | 273.848338 | all-MiniLM-L6-v2 | 0.746667 | -0.028393 | 0.174224 | 0.827824 |
ArXiv ML Papers | KeyNMF | 45 | 30 | [
[
"knowledge",
"learn",
"learning",
"machine",
"this",
"generalization",
"learned",
"adaptation",
"ensemble",
"quantum"
],
[
"neural",
"network",
"layer",
"networks",
"recurrent",
"neurons",
"layers",
"neuron",
"architectures",
"input"
],
[
"robustness",
"adversarial",
"defense",
"attacks",
"robust",
"attack",
"examples",
"attacker",
"security",
"detection"
],
[
"classes",
"recognition",
"class",
"ensemble",
"classification",
"label",
"accuracy",
"classify",
"benchmark",
"detection"
],
[
"predictive",
"predictions",
"model",
"forecasting",
"models",
"modeling",
"machine",
"forecast",
"parameters",
"inference"
],
[
"algorithms",
"algorithm",
"search",
"matrix",
"efficiently",
"methods",
"efficient",
"complexity",
"factorization",
"sampling"
],
[
"tensor",
"imagenet",
"bias",
"overfitting",
"labels",
"benchmark",
"regularization",
"datasets",
"label",
"labeled"
],
[
"edges",
"node",
"gnns",
"nodes",
"graphs",
"embedding",
"embeddings",
"network",
"graph",
"networks"
],
[
"networks",
"adversarial",
"generation",
"gans",
"generator",
"generative",
"generate",
"gan",
"generating",
"generated"
],
[
"learn",
"learns",
"embedding",
"learned",
"representations",
"representation",
"embeddings",
"objects",
"visual",
"recognition"
],
[
"policy",
"action",
"rl",
"agents",
"reinforcement",
"control",
"reward",
"learning",
"agent",
"policies"
],
[
"layers",
"cnns",
"model",
"filters",
"convolutional",
"imagenet",
"convolution",
"cnn",
"network",
"recognition"
],
[
"training",
"trained",
"task",
"loss",
"train",
"transfer",
"tasks",
"generalization",
"distillation",
"network"
],
[
"convex",
"gradient",
"stochastic",
"optimization",
"descent",
"algorithm",
"minimization",
"regularization",
"convergence",
"constraints"
],
[
"predicting",
"predict",
"predictions",
"prediction",
"predictive",
"regression",
"forecasting",
"predicted",
"forecast",
"forecasts"
],
[
"unsupervised",
"supervised",
"labeled",
"labels",
"unlabeled",
"label",
"semi",
"regularization",
"self",
"labeling"
],
[
"accuracy",
"crucial",
"network",
"dnns",
"learning",
"medical",
"mri",
"deep",
"benchmarks",
"dnn"
],
[
"temporal",
"health",
"dimensional",
"statistical",
"information",
"knowledge",
"causal",
"data",
"empirical",
"dimensionality"
],
[
"features",
"feature",
"matrix",
"selection",
"visual",
"sparse",
"attention",
"interpretability",
"dimensionality",
"factorization"
],
[
"bandits",
"regret",
"bandit",
"exploration",
"stochastic",
"reward",
"rewards",
"strategy",
"sampling",
"exploitation"
],
[
"latent",
"autoencoders",
"autoencoder",
"generative",
"variational",
"encoder",
"decoder",
"vae",
"model",
"learned"
],
[
"detect",
"detection",
"traffic",
"dataset",
"accuracy",
"trained",
"challenge",
"results",
"detecting",
"identification"
],
[
"image",
"images",
"visual",
"vision",
"segmentation",
"detection",
"imaging",
"object",
"resolution",
"3d"
],
[
"tasks",
"attention",
"language",
"text",
"task",
"recurrent",
"encoder",
"speech",
"languages",
"decoder"
],
[
"memory",
"network",
"efficient",
"performance",
"compression",
"accuracy",
"quantization",
"imagenet",
"pruning",
"hardware"
],
[
"optimal",
"algorithm",
"sample",
"minimax",
"markov",
"complexity",
"generalization",
"optimality",
"estimation",
"bound"
],
[
"distributed",
"federated",
"private",
"privacy",
"information",
"communication",
"complexity",
"differentially",
"decentralized",
"differential"
],
[
"accuracy",
"classifiers",
"classifier",
"detection",
"classes",
"class",
"svm",
"classifying",
"performance",
"fairness"
],
[
"gaussian",
"sparse",
"prior",
"inference",
"bayesian",
"posterior",
"variational",
"sampling",
"stochastic",
"estimation"
],
[
"cluster",
"clustering",
"clusters",
"algorithm",
"unsupervised",
"sparse",
"matrix",
"analysis",
"patterns",
"hierarchical"
]
] | 315.3753 | all-MiniLM-L6-v2 | 0.73 | -0.020482 | 0.177637 | 0.823951 |
ArXiv ML Papers | KeyNMF | 46 | 30 | [
[
"machine",
"learning",
"learn",
"knowledge",
"this",
"generalization",
"adaptation",
"learned",
"ensemble",
"quantum"
],
[
"networks",
"neural",
"network",
"architectures",
"layers",
"layer",
"recurrent",
"neurons",
"neuron",
"input"
],
[
"robust",
"attacks",
"robustness",
"adversarial",
"attack",
"defense",
"examples",
"attacker",
"adversarially",
"security"
],
[
"class",
"classification",
"classes",
"classify",
"binary",
"label",
"accuracy",
"recognition",
"ensemble",
"benchmark"
],
[
"predictions",
"models",
"inference",
"forecasting",
"machine",
"predictive",
"forecast",
"model",
"parameters",
"modeling"
],
[
"algorithm",
"algorithms",
"matrix",
"search",
"efficient",
"minimization",
"complexity",
"factorization",
"efficiently",
"methods"
],
[
"overfitting",
"imagenet",
"datasets",
"bias",
"labels",
"label",
"regularization",
"tensor",
"labeled",
"benchmark"
],
[
"graph",
"graphs",
"nodes",
"networks",
"network",
"node",
"embeddings",
"embedding",
"edges",
"gnns"
],
[
"gans",
"networks",
"generation",
"gan",
"adversarial",
"generate",
"generator",
"generative",
"generating",
"generated"
],
[
"learn",
"representations",
"learns",
"learned",
"visual",
"embedding",
"embeddings",
"objects",
"representation",
"semantic"
],
[
"reinforcement",
"policy",
"learning",
"agent",
"agents",
"rl",
"reward",
"control",
"policies",
"action"
],
[
"recognition",
"cnn",
"convolutional",
"filters",
"network",
"cnns",
"layers",
"imagenet",
"convolution",
"vision"
],
[
"training",
"loss",
"tasks",
"generalization",
"train",
"transfer",
"weights",
"federated",
"examples",
"normalization"
],
[
"descent",
"stochastic",
"gradient",
"convex",
"optimization",
"convergence",
"algorithm",
"minimization",
"regularization",
"sgd"
],
[
"predicting",
"predictive",
"forecasting",
"regression",
"predictions",
"prediction",
"predict",
"predicted",
"forecast",
"forecasts"
],
[
"labeled",
"unsupervised",
"self",
"supervised",
"labels",
"label",
"semi",
"unlabeled",
"regularization",
"labeling"
],
[
"dnns",
"deep",
"dnn",
"accuracy",
"benchmarks",
"network",
"medical",
"mri",
"learning",
"crucial"
],
[
"knowledge",
"data",
"information",
"statistical",
"temporal",
"regression",
"health",
"dimensional",
"causal",
"empirical"
],
[
"selection",
"feature",
"attention",
"sparse",
"features",
"visual",
"matrix",
"factorization",
"interpretability",
"dimensionality"
],
[
"regret",
"exploration",
"optimal",
"complexity",
"sampling",
"reward",
"stochastic",
"bandit",
"bandits",
"rewards"
],
[
"detect",
"detection",
"anomaly",
"anomalies",
"detecting",
"outlier",
"attacks",
"intrusion",
"object",
"detector"
],
[
"dataset",
"accuracy",
"expert",
"identification",
"results",
"traffic",
"features",
"mnist",
"biases",
"challenge"
],
[
"images",
"image",
"segmentation",
"vision",
"imaging",
"visual",
"resolution",
"3d",
"pixel",
"pixels"
],
[
"tasks",
"attention",
"language",
"task",
"recurrent",
"text",
"encoder",
"languages",
"speech",
"decoder"
],
[
"trained",
"knowledge",
"layer",
"tasks",
"distillation",
"task",
"network",
"ai",
"transfer",
"input"
],
[
"bayesian",
"inference",
"posterior",
"autoencoders",
"autoencoder",
"variational",
"stochastic",
"latent",
"gaussian",
"prior"
],
[
"information",
"privacy",
"decentralized",
"private",
"distributed",
"federated",
"complexity",
"communication",
"differentially",
"differential"
],
[
"classifiers",
"accuracy",
"svm",
"classifier",
"classes",
"classifying",
"class",
"fairness",
"performance",
"labels"
],
[
"unsupervised",
"analysis",
"algorithm",
"cluster",
"sparse",
"matrix",
"clusters",
"hierarchical",
"clustering",
"instances"
],
[
"hardware",
"accuracy",
"imagenet",
"efficient",
"compression",
"quantization",
"performance",
"memory",
"network",
"pruning"
]
] | 315.064199 | all-MiniLM-L6-v2 | 0.753333 | -0.043129 | 0.176423 | 0.825853 |
ArXiv ML Papers | KeyNMF | 43 | 40 | [
[
"generalization",
"this",
"machine",
"learning",
"quantum",
"challenges",
"learners",
"ml",
"fairness",
"computational"
],
[
"neurons",
"recurrent",
"neuron",
"neural",
"networks",
"architectures",
"layers",
"layer",
"input",
"gradient"
],
[
"attack",
"defense",
"robust",
"adversarial",
"robustness",
"attacks",
"examples",
"attacker",
"adversarially",
"security"
],
[
"label",
"classification",
"accuracy",
"classes",
"class",
"classify",
"ensemble",
"labels",
"benchmark",
"binary"
],
[
"model",
"predictions",
"predictive",
"models",
"machine",
"modeling",
"parameters",
"systems",
"inference",
"interpretable"
],
[
"complexity",
"matrix",
"efficient",
"algorithm",
"search",
"algorithms",
"efficiently",
"methods",
"sampling",
"adaptive"
],
[
"benchmark",
"labeled",
"bias",
"imagenet",
"labels",
"label",
"overfitting",
"datasets",
"tensor",
"sampling"
],
[
"graph",
"networks",
"nodes",
"graphs",
"node",
"structure",
"edges",
"gnns",
"vertices",
"gnn"
],
[
"adversarial",
"gan",
"gans",
"generative",
"generation",
"networks",
"generate",
"generator",
"generating",
"generated"
],
[
"reinforcement",
"policy",
"agent",
"reward",
"agents",
"rl",
"control",
"learning",
"policies",
"action"
],
[
"markov",
"bound",
"optimality",
"estimation",
"optimal",
"sample",
"algorithm",
"complexity",
"generalization",
"minimax"
],
[
"convolutional",
"cnn",
"cnns",
"convolution",
"imagenet",
"layers",
"filters",
"vision",
"pooling",
"filter"
],
[
"train",
"distillation",
"layer",
"trained",
"input",
"classify",
"brain",
"inputs",
"softmax",
"objects"
],
[
"algorithm",
"minimization",
"descent",
"gradient",
"convergence",
"stochastic",
"optimization",
"constraints",
"convex",
"sgd"
],
[
"predictions",
"predict",
"predicting",
"lstm",
"regression",
"performance",
"prediction",
"predictors",
"predictive",
"predicted"
],
[
"unlabeled",
"supervised",
"self",
"labeling",
"labels",
"label",
"annotation",
"unsupervised",
"semi",
"labeled"
],
[
"accuracy",
"dnns",
"mri",
"deep",
"crucial",
"medical",
"benchmarks",
"learned",
"dnn",
"net"
],
[
"statistical",
"privacy",
"private",
"data",
"information",
"dimensional",
"causal",
"empirical",
"health",
"temporal"
],
[
"visual",
"selection",
"features",
"feature",
"interpretability",
"explanations",
"extraction",
"dimensionality",
"importance",
"tree"
],
[
"posterior",
"bayesian",
"gaussian",
"sampling",
"inference",
"stochastic",
"prior",
"variational",
"likelihood",
"estimation"
],
[
"latent",
"variational",
"autoencoders",
"autoencoder",
"encoder",
"decoder",
"model",
"vae",
"generative",
"representation"
],
[
"accuracy",
"mnist",
"dataset",
"biases",
"results",
"traffic",
"challenge",
"bias",
"imbalance",
"identification"
],
[
"visual",
"image",
"segmentation",
"images",
"vision",
"resolution",
"imaging",
"3d",
"pixels",
"pixel"
],
[
"computational",
"compression",
"accuracy",
"hardware",
"imagenet",
"memory",
"quantization",
"efficient",
"processing",
"performance"
],
[
"transfer",
"examples",
"generalization",
"federated",
"train",
"weights",
"loss",
"normalization",
"training",
"samples"
],
[
"sparsity",
"tensor",
"matrix",
"sparse",
"factorization",
"dimensional",
"gaussian",
"linear",
"thresholding",
"matrices"
],
[
"layers",
"model",
"distributed",
"layer",
"network",
"pruning",
"architecture",
"flow",
"architectures",
"communication"
],
[
"accuracy",
"classifiers",
"svm",
"classifier",
"class",
"classes",
"fairness",
"classifying",
"adversarially",
"performance"
],
[
"clustering",
"cluster",
"algorithm",
"unsupervised",
"instances",
"clusters",
"analysis",
"hierarchical",
"segmentation",
"patterns"
],
[
"visual",
"speech",
"face",
"recognition",
"audio",
"accuracy",
"facial",
"bias",
"acoustic",
"recognize"
],
[
"transfer",
"target",
"unsupervised",
"learned",
"adaptation",
"knowledge",
"domain",
"reasoning",
"distillation",
"student"
],
[
"reward",
"regret",
"exploration",
"bandits",
"bandit",
"rewards",
"exploitation",
"strategy",
"stochastic",
"sampling"
],
[
"learned",
"visual",
"learn",
"learns",
"representations",
"objects",
"representation",
"object",
"agent",
"scene"
],
[
"regularization",
"loss",
"overfitting",
"generalization",
"regularized",
"bias",
"lasso",
"dnns",
"label",
"robust"
],
[
"vision",
"sequence",
"neural",
"encoder",
"recurrent",
"decoder",
"attention",
"transformer",
"architecture",
"visual"
],
[
"detect",
"detection",
"anomaly",
"object",
"detecting",
"attacks",
"outlier",
"intrusion",
"anomalies",
"detector"
],
[
"human",
"challenge",
"ai",
"task",
"robotic",
"representation",
"robot",
"tasks",
"labeling",
"forgetting"
],
[
"dimensional",
"representation",
"space",
"nodes",
"vectors",
"embeddings",
"semantic",
"representations",
"similarity",
"embedding"
],
[
"forecast",
"forecasting",
"model",
"ensemble",
"boosting",
"accuracy",
"forecasts",
"weather",
"predicting",
"predictions"
],
[
"languages",
"language",
"recurrent",
"text",
"nlp",
"corpus",
"speech",
"words",
"semantic",
"word"
]
] | 327.944701 | all-MiniLM-L6-v2 | 0.7325 | -0.06031 | 0.174484 | 0.851345 |
ArXiv ML Papers | KeyNMF | 44 | 40 | [
[
"machine",
"learning",
"knowledge",
"adaptation",
"ml",
"automated",
"this",
"generalization",
"fairness",
"computational"
],
[
"neurons",
"recurrent",
"networks",
"neural",
"neuron",
"layer",
"layers",
"input",
"architectures",
"gradient"
],
[
"robustness",
"attack",
"attacks",
"defense",
"examples",
"adversarial",
"robust",
"attacker",
"adversarially",
"security"
],
[
"accuracy",
"ensemble",
"class",
"classify",
"labels",
"label",
"binary",
"classes",
"classification",
"benchmark"
],
[
"interpretable",
"predictions",
"bias",
"language",
"models",
"machine",
"inference",
"ai",
"many",
"likelihood"
],
[
"matrix",
"methods",
"efficiently",
"adaptive",
"algorithm",
"efficient",
"algorithms",
"complexity",
"search",
"minimization"
],
[
"overfitting",
"imagenet",
"benchmark",
"bias",
"labels",
"datasets",
"regularization",
"label",
"labeled",
"sampling"
],
[
"graphs",
"graph",
"nodes",
"edges",
"networks",
"node",
"gnns",
"vertices",
"knowledge",
"structure"
],
[
"adversarial",
"generate",
"generated",
"generation",
"networks",
"generator",
"generative",
"generating",
"gan",
"gans"
],
[
"objects",
"learn",
"learns",
"flow",
"visual",
"imitation",
"examples",
"representations",
"dynamics",
"learned"
],
[
"planning",
"rl",
"policy",
"control",
"reward",
"learning",
"action",
"policies",
"actions",
"reinforcement"
],
[
"convolution",
"convolutional",
"cnn",
"cnns",
"layers",
"imagenet",
"filters",
"vision",
"pooling",
"filter"
],
[
"knowledge",
"classify",
"brain",
"train",
"layer",
"inputs",
"distillation",
"trained",
"softmax",
"input"
],
[
"convex",
"optimization",
"minimization",
"descent",
"constraints",
"algorithm",
"stochastic",
"convergence",
"sgd",
"gradient"
],
[
"lstm",
"predict",
"predicting",
"predictions",
"regression",
"prediction",
"predictive",
"predicted",
"performance",
"predictors"
],
[
"labeled",
"unlabeled",
"supervised",
"unsupervised",
"labels",
"semi",
"label",
"self",
"labeling",
"annotation"
],
[
"crucial",
"accuracy",
"dnns",
"mri",
"deep",
"benchmarks",
"dnn",
"medical",
"learning",
"generalization"
],
[
"knowledge",
"statistical",
"data",
"information",
"dimensional",
"health",
"empirical",
"causal",
"labeled",
"temporal"
],
[
"visual",
"feature",
"features",
"interpretability",
"explanations",
"selection",
"importance",
"extraction",
"dimensionality",
"tree"
],
[
"bandit",
"bandits",
"exploration",
"optimal",
"regret",
"reward",
"complexity",
"rewards",
"sampling",
"stochastic"
],
[
"autoencoders",
"autoencoder",
"variational",
"generative",
"decoder",
"latent",
"encoder",
"vae",
"learned",
"representation"
],
[
"traffic",
"accuracy",
"imbalance",
"challenge",
"results",
"mnist",
"dataset",
"biases",
"expert",
"identification"
],
[
"images",
"image",
"visual",
"segmentation",
"imaging",
"vision",
"resolution",
"pixels",
"3d",
"pixel"
],
[
"acoustic",
"audio",
"visual",
"face",
"speech",
"recognition",
"accuracy",
"facial",
"recognize",
"bias"
],
[
"generalization",
"training",
"loss",
"train",
"transfer",
"examples",
"weights",
"federated",
"normalization",
"samples"
],
[
"dimensional",
"thresholding",
"sparse",
"sparsity",
"linear",
"lasso",
"regularization",
"matrix",
"factorization",
"gaussian"
],
[
"pruning",
"network",
"layers",
"flow",
"layer",
"distributed",
"relu",
"architectures",
"communication",
"architecture"
],
[
"classifiers",
"classifier",
"accuracy",
"svm",
"class",
"classifying",
"classes",
"fairness",
"labels",
"generalization"
],
[
"gaussian",
"bayesian",
"stochastic",
"inference",
"posterior",
"likelihood",
"estimation",
"sampling",
"prior",
"variational"
],
[
"cluster",
"analysis",
"instances",
"algorithm",
"unsupervised",
"clusters",
"clustering",
"hierarchical",
"patterns",
"segmentation"
],
[
"human",
"robot",
"task",
"nlp",
"challenge",
"information",
"tasks",
"language",
"knowledge",
"forgetting"
],
[
"actions",
"strategies",
"agents",
"ai",
"communication",
"environments",
"human",
"games",
"game",
"agent"
],
[
"semantic",
"embeddings",
"representation",
"representations",
"embedding",
"language",
"word",
"words",
"similarity",
"dimensional"
],
[
"tensor",
"matrices",
"rank",
"matrix",
"factorization",
"tensors",
"decomposition",
"completion",
"complexity",
"quantum"
],
[
"distributed",
"privacy",
"complexity",
"information",
"private",
"federated",
"decentralized",
"differentially",
"differential",
"security"
],
[
"intrusion",
"detect",
"detection",
"object",
"attacks",
"detecting",
"anomaly",
"outlier",
"detector",
"anomalies"
],
[
"forecast",
"accuracy",
"forecasting",
"boosting",
"forecasts",
"ensemble",
"weather",
"predicting",
"predictions",
"forest"
],
[
"efficient",
"accuracy",
"quantization",
"memory",
"compression",
"imagenet",
"performance",
"hardware",
"computational",
"computing"
],
[
"sequence",
"attention",
"language",
"decoder",
"encoder",
"rnn",
"recurrent",
"rnns",
"text",
"neural"
],
[
"predictive",
"parameters",
"speech",
"architecture",
"systems",
"predictions",
"modeling",
"dynamics",
"example",
"model"
]
] | 320.776981 | all-MiniLM-L6-v2 | 0.7325 | -0.05883 | 0.17023 | 0.852586 |
ArXiv ML Papers | KeyNMF | 45 | 40 | [
[
"machine",
"fairness",
"ml",
"generalization",
"learning",
"learn",
"computational",
"quantum",
"challenges",
"this"
],
[
"recurrent",
"neural",
"neuron",
"input",
"architectures",
"networks",
"neurons",
"gradient",
"layers",
"layer"
],
[
"attacks",
"adversarial",
"robustness",
"robust",
"defense",
"examples",
"attack",
"attacker",
"adversarially",
"security"
],
[
"classification",
"class",
"classes",
"classify",
"labels",
"label",
"benchmark",
"ensemble",
"binary",
"accuracy"
],
[
"predictive",
"modeling",
"models",
"inference",
"model",
"machine",
"predictions",
"parameters",
"systems",
"graphical"
],
[
"efficient",
"algorithms",
"algorithm",
"search",
"efficiently",
"adaptive",
"matrix",
"methods",
"complexity",
"minimization"
],
[
"datasets",
"benchmark",
"overfitting",
"sampling",
"labels",
"label",
"imagenet",
"labeled",
"bias",
"samples"
],
[
"graph",
"nodes",
"node",
"networks",
"graphs",
"embedding",
"edges",
"embeddings",
"gnns",
"vertices"
],
[
"gans",
"generate",
"generation",
"adversarial",
"gan",
"generative",
"networks",
"generator",
"generating",
"generated"
],
[
"forgetting",
"robotic",
"tasks",
"challenge",
"robot",
"human",
"learned",
"task",
"learn",
"labeling"
],
[
"policy",
"agent",
"reinforcement",
"reward",
"agents",
"control",
"policies",
"action",
"rl",
"games"
],
[
"cnn",
"convolutional",
"convolution",
"cnns",
"layers",
"imagenet",
"vision",
"filters",
"pooling",
"filter"
],
[
"trained",
"input",
"layer",
"classify",
"brain",
"inputs",
"softmax",
"distillation",
"labels",
"train"
],
[
"convergence",
"optimization",
"convex",
"stochastic",
"gradient",
"descent",
"algorithm",
"minimization",
"sgd",
"constraints"
],
[
"regression",
"prediction",
"predictions",
"predicted",
"predictive",
"predicting",
"predict",
"lstm",
"performance",
"predictors"
],
[
"unsupervised",
"supervised",
"labeled",
"unlabeled",
"labels",
"semi",
"label",
"self",
"labeling",
"annotation"
],
[
"benchmarks",
"medical",
"dnn",
"dnns",
"mri",
"deep",
"crucial",
"accuracy",
"optimal",
"learning"
],
[
"empirical",
"data",
"temporal",
"analysis",
"health",
"causal",
"dimensional",
"information",
"statistical",
"dimensionality"
],
[
"features",
"interpretability",
"selection",
"visual",
"feature",
"explanations",
"importance",
"dimensionality",
"extraction",
"tree"
],
[
"stochastic",
"sampling",
"bayesian",
"posterior",
"gaussian",
"inference",
"prior",
"estimation",
"variational",
"likelihood"
],
[
"encoder",
"variational",
"autoencoder",
"autoencoders",
"latent",
"decoder",
"generative",
"vae",
"model",
"learned"
],
[
"dataset",
"accuracy",
"challenge",
"results",
"imbalance",
"bias",
"traffic",
"mnist",
"biases",
"identification"
],
[
"images",
"image",
"segmentation",
"visual",
"vision",
"imaging",
"3d",
"resolution",
"pixels",
"pixel"
],
[
"network",
"layer",
"pruning",
"model",
"layers",
"flow",
"distributed",
"architecture",
"architectures",
"communication"
],
[
"train",
"generalization",
"loss",
"training",
"transfer",
"examples",
"federated",
"weights",
"normalization",
"samples"
],
[
"factorization",
"sparsity",
"sparse",
"matrix",
"thresholding",
"dimensional",
"linear",
"gaussian",
"complexity",
"coding"
],
[
"privacy",
"private",
"federated",
"decentralized",
"distributed",
"information",
"differentially",
"complexity",
"differential",
"communication"
],
[
"accuracy",
"classes",
"class",
"fairness",
"svm",
"classifier",
"classifying",
"classifiers",
"adversarially",
"bayes"
],
[
"analysis",
"clusters",
"cluster",
"algorithm",
"clustering",
"unsupervised",
"instances",
"hierarchical",
"patterns",
"segmentation"
],
[
"speech",
"facial",
"visual",
"audio",
"face",
"bias",
"accuracy",
"acoustic",
"recognize",
"recognition"
],
[
"reasoning",
"distillation",
"learned",
"knowledge",
"unsupervised",
"domain",
"transfer",
"adaptation",
"target",
"student"
],
[
"bandit",
"optimal",
"bandits",
"exploration",
"sampling",
"rewards",
"reward",
"regret",
"stochastic",
"strategy"
],
[
"ensemble",
"model",
"forecasting",
"forecasts",
"forecast",
"boosting",
"weather",
"predicting",
"accuracy",
"predictions"
],
[
"matrix",
"factorization",
"matrices",
"tensors",
"completion",
"tensor",
"decomposition",
"quantum",
"complexity",
"rank"
],
[
"architecture",
"neural",
"recurrent",
"sequence",
"vision",
"transformer",
"attention",
"encoder",
"decoder",
"rnns"
],
[
"detect",
"detection",
"detecting",
"anomaly",
"object",
"attacks",
"outlier",
"anomalies",
"intrusion",
"detector"
],
[
"representations",
"learns",
"embeddings",
"learn",
"embedding",
"representation",
"objects",
"visual",
"learned",
"semantic"
],
[
"imagenet",
"accuracy",
"efficient",
"memory",
"quantization",
"computational",
"compression",
"hardware",
"machine",
"performance"
],
[
"languages",
"corpus",
"text",
"words",
"nlp",
"speech",
"language",
"recurrent",
"semantic",
"word"
],
[
"regularized",
"generalization",
"regularization",
"overfitting",
"bias",
"lasso",
"loss",
"label",
"robust",
"dnns"
]
] | 281.200305 | all-MiniLM-L6-v2 | 0.72 | -0.057154 | 0.177938 | 0.847863 |
ArXiv ML Papers | KeyNMF | 46 | 40 | [
[
"ml",
"learn",
"this",
"machine",
"generalization",
"learning",
"fairness",
"quantum",
"challenges",
"computational"
],
[
"networks",
"layer",
"neural",
"network",
"neurons",
"recurrent",
"layers",
"neuron",
"input",
"architectures"
],
[
"attacks",
"adversarial",
"robustness",
"attack",
"robust",
"defense",
"examples",
"attacker",
"adversarially",
"security"
],
[
"benchmark",
"accuracy",
"classify",
"classes",
"labels",
"label",
"classification",
"class",
"ensemble",
"binary"
],
[
"model",
"machine",
"predictions",
"models",
"modeling",
"predictive",
"parameters",
"systems",
"inference",
"graphical"
],
[
"efficiently",
"algorithm",
"search",
"algorithms",
"efficient",
"methods",
"sampling",
"adaptive",
"selection",
"complexity"
],
[
"benchmark",
"overfitting",
"labeled",
"imagenet",
"labels",
"label",
"datasets",
"bias",
"sampling",
"samples"
],
[
"graphs",
"networks",
"nodes",
"network",
"graph",
"node",
"edges",
"gnns",
"embeddings",
"embedding"
],
[
"gans",
"generation",
"gan",
"adversarial",
"generative",
"generate",
"generating",
"generator",
"networks",
"generated"
],
[
"learn",
"task",
"learned",
"robot",
"robotic",
"tasks",
"human",
"forgetting",
"challenge",
"ai"
],
[
"policy",
"reward",
"reinforcement",
"agents",
"agent",
"learning",
"control",
"rl",
"policies",
"action"
],
[
"layers",
"convolutional",
"network",
"cnns",
"convolution",
"cnn",
"imagenet",
"filters",
"model",
"vision"
],
[
"trained",
"layer",
"train",
"input",
"network",
"distillation",
"softmax",
"classify",
"layers",
"brain"
],
[
"stochastic",
"optimization",
"gradient",
"descent",
"convergence",
"convex",
"algorithm",
"constraints",
"minimization",
"sgd"
],
[
"predicted",
"predictive",
"performance",
"predictors",
"lstm",
"regression",
"predictions",
"prediction",
"predicting",
"predict"
],
[
"unlabeled",
"unsupervised",
"self",
"labeling",
"labels",
"label",
"supervised",
"annotation",
"semi",
"labeled"
],
[
"dnn",
"dnns",
"deep",
"accuracy",
"benchmarks",
"network",
"learning",
"mri",
"crucial",
"medical"
],
[
"dimensionality",
"data",
"dimensional",
"health",
"information",
"temporal",
"causal",
"empirical",
"analysis",
"statistical"
],
[
"feature",
"features",
"selection",
"dimensionality",
"explanations",
"visual",
"interpretability",
"importance",
"extraction",
"tree"
],
[
"sampling",
"bayesian",
"likelihood",
"inference",
"gaussian",
"stochastic",
"estimation",
"posterior",
"variational",
"prior"
],
[
"autoencoders",
"encoder",
"autoencoder",
"latent",
"variational",
"decoder",
"model",
"generative",
"vae",
"learned"
],
[
"challenge",
"accuracy",
"dataset",
"traffic",
"results",
"mnist",
"biases",
"bias",
"imbalance",
"identification"
],
[
"vision",
"visual",
"imaging",
"3d",
"image",
"images",
"pixels",
"segmentation",
"resolution",
"pixel"
],
[
"memory",
"network",
"efficient",
"performance",
"compression",
"quantization",
"accuracy",
"pruning",
"imagenet",
"hardware"
],
[
"transfer",
"training",
"generalization",
"loss",
"train",
"federated",
"weights",
"examples",
"normalization",
"samples"
],
[
"sparsity",
"sparse",
"dimensional",
"factorization",
"thresholding",
"matrix",
"linear",
"gaussian",
"compressed",
"coding"
],
[
"privacy",
"distributed",
"private",
"federated",
"information",
"decentralized",
"complexity",
"differentially",
"differential",
"communication"
],
[
"svm",
"accuracy",
"classifiers",
"classifier",
"classes",
"fairness",
"class",
"classifying",
"performance",
"adversarially"
],
[
"clusters",
"clustering",
"hierarchical",
"analysis",
"cluster",
"unsupervised",
"algorithm",
"instances",
"patterns",
"segmentation"
],
[
"face",
"audio",
"recognition",
"acoustic",
"bias",
"speech",
"facial",
"accuracy",
"visual",
"recognize"
],
[
"distillation",
"transfer",
"learned",
"knowledge",
"reasoning",
"unsupervised",
"adaptation",
"domain",
"student",
"target"
],
[
"rewards",
"exploration",
"reward",
"bandits",
"bandit",
"regret",
"stochastic",
"strategy",
"exploitation",
"sampling"
],
[
"learns",
"learn",
"embeddings",
"semantic",
"embedding",
"learned",
"representations",
"objects",
"visual",
"representation"
],
[
"completion",
"tensors",
"tensor",
"matrices",
"rank",
"factorization",
"decomposition",
"matrix",
"quantum",
"complexity"
],
[
"attention",
"recurrent",
"decoder",
"encoder",
"sequence",
"vision",
"neural",
"architecture",
"transformer",
"rnns"
],
[
"detection",
"detect",
"intrusion",
"detecting",
"anomaly",
"anomalies",
"attacks",
"object",
"outlier",
"detector"
],
[
"bias",
"regularized",
"loss",
"overfitting",
"lasso",
"label",
"regularization",
"generalization",
"dnns",
"robust"
],
[
"sample",
"algorithm",
"complexity",
"minimax",
"optimal",
"optimality",
"bound",
"generalization",
"estimation",
"markov"
],
[
"forecasts",
"forecast",
"ensemble",
"forecasting",
"boosting",
"model",
"accuracy",
"weather",
"predicting",
"predictions"
],
[
"recurrent",
"speech",
"language",
"words",
"nlp",
"languages",
"text",
"corpus",
"semantic",
"word"
]
] | 327.913438 | all-MiniLM-L6-v2 | 0.7225 | -0.046772 | 0.174806 | 0.842258 |
ArXiv ML Papers | KeyNMF | 43 | 50 | [
[
"learning",
"machine",
"this",
"generalization",
"ml",
"fairness",
"computational",
"challenges",
"quantum",
"automated"
],
[
"neural",
"networks",
"input",
"recurrent",
"neurons",
"layer",
"architectures",
"neuron",
"weights",
"backpropagation"
],
[
"attacks",
"adversarial",
"attack",
"examples",
"security",
"robustness",
"defense",
"attacker",
"dnns",
"vulnerable"
],
[
"benchmark",
"class",
"classification",
"binary",
"forest",
"classes",
"accuracy",
"ensemble",
"classify",
"svm"
],
[
"interpretability",
"models",
"many",
"inference",
"predictions",
"interpretable",
"likelihood",
"ai",
"bias",
"machine"
],
[
"algorithm",
"complexity",
"adaptive",
"efficient",
"sample",
"edge",
"samples",
"search",
"distributed",
"bounds"
],
[
"datasets",
"imagenet",
"benchmark",
"bias",
"samples",
"overfitting",
"sampling",
"annotation",
"fairness",
"sentiment"
],
[
"edges",
"graph",
"nodes",
"networks",
"graphs",
"node",
"gnns",
"structure",
"gnn",
"vertices"
],
[
"generation",
"adversarial",
"gan",
"generative",
"generate",
"gans",
"networks",
"generating",
"generator",
"generated"
],
[
"robotic",
"robot",
"task",
"tasks",
"challenge",
"vision",
"human",
"challenging",
"forgetting",
"representation"
],
[
"reinforcement",
"policy",
"reward",
"learning",
"rl",
"control",
"policies",
"action",
"planning",
"games"
],
[
"layers",
"depth",
"segmentation",
"image",
"convolution",
"convolutional",
"convolutions",
"pooling",
"layer",
"audio"
],
[
"layer",
"train",
"classify",
"inputs",
"distillation",
"input",
"trained",
"brain",
"softmax",
"objects"
],
[
"minimization",
"optimizing",
"convex",
"constraints",
"optimization",
"objective",
"optimize",
"constrained",
"search",
"optimizer"
],
[
"predict",
"prediction",
"predictions",
"predicting",
"predictive",
"regression",
"predicted",
"predictors",
"performance",
"future"
],
[
"self",
"supervised",
"labeled",
"unlabeled",
"unsupervised",
"semi",
"annotation",
"annotated",
"transformations",
"imagenet"
],
[
"benchmarks",
"deep",
"mri",
"dnn",
"dnns",
"accuracy",
"medical",
"net",
"crucial",
"batch"
],
[
"information",
"data",
"statistical",
"dimensional",
"empirical",
"causal",
"temporal",
"health",
"analysis",
"dimensionality"
],
[
"features",
"visual",
"interpretability",
"feature",
"selection",
"explanations",
"dimensionality",
"importance",
"tree",
"extraction"
],
[
"gaussian",
"posterior",
"inference",
"bayesian",
"sampling",
"prior",
"likelihood",
"variational",
"estimation",
"distributions"
],
[
"encoder",
"latent",
"decoder",
"autoencoders",
"autoencoder",
"variational",
"vae",
"generative",
"learned",
"representation"
],
[
"results",
"mnist",
"dataset",
"challenge",
"accuracy",
"traffic",
"bias",
"identification",
"biases",
"imbalance"
],
[
"image",
"images",
"segmentation",
"vision",
"visual",
"imaging",
"3d",
"pixels",
"resolution",
"pixel"
],
[
"recognition",
"accuracy",
"visual",
"speech",
"face",
"audio",
"facial",
"bias",
"acoustic",
"recognize"
],
[
"transfer",
"weights",
"normalization",
"generalization",
"federated",
"train",
"training",
"loss",
"examples",
"samples"
],
[
"generalization",
"minimax",
"complexity",
"bound",
"estimation",
"optimal",
"sample",
"optimality",
"markov",
"approximation"
],
[
"architectures",
"architecture",
"distributed",
"layers",
"pruning",
"flow",
"layer",
"communication",
"network",
"relu"
],
[
"svm",
"accuracy",
"class",
"classifiers",
"classifier",
"classes",
"classifying",
"bayes",
"fairness",
"performance"
],
[
"cluster",
"clusters",
"clustering",
"unsupervised",
"analysis",
"instances",
"patterns",
"hierarchical",
"segmentation",
"nodes"
],
[
"machine",
"algorithms",
"search",
"performance",
"sampling",
"minimization",
"methods",
"efficiently",
"trees",
"matrix"
],
[
"distillation",
"reasoning",
"knowledge",
"transfer",
"unsupervised",
"domain",
"learned",
"adaptation",
"student",
"target"
],
[
"rewards",
"exploitation",
"regret",
"bandits",
"stochastic",
"reward",
"strategy",
"exploration",
"bandit",
"sampling"
],
[
"embeddings",
"representation",
"embedding",
"semantic",
"dimensional",
"representations",
"space",
"retrieval",
"vectors",
"objects"
],
[
"sparsity",
"sparse",
"matrix",
"linear",
"thresholding",
"dimensional",
"coding",
"gaussian",
"factorization",
"compressed"
],
[
"privacy",
"federated",
"information",
"private",
"distributed",
"decentralized",
"differentially",
"complexity",
"differential",
"security"
],
[
"detection",
"object",
"detect",
"anomaly",
"detecting",
"intrusion",
"attacks",
"outlier",
"anomalies",
"detector"
],
[
"sequence",
"neural",
"transformer",
"encoder",
"attention",
"decoder",
"recurrent",
"vision",
"architecture",
"rnns"
],
[
"performance",
"computational",
"lstm",
"efficient",
"memory",
"hardware",
"quantization",
"compression",
"processing",
"computing"
],
[
"forecasting",
"forecasts",
"predicting",
"ensemble",
"forecast",
"weather",
"boosting",
"accuracy",
"predictions",
"forest"
],
[
"model",
"predictions",
"modeling",
"systems",
"architecture",
"parameters",
"dynamics",
"example",
"predictive",
"speech"
],
[
"cnns",
"cnn",
"imagenet",
"pruning",
"accuracy",
"filter",
"pruned",
"filters",
"convolution",
"vision"
],
[
"stochastic",
"gradient",
"batch",
"hessian",
"sgd",
"descent",
"convex",
"boosting",
"convergence",
"gradients"
],
[
"language",
"speech",
"nlp",
"words",
"recurrent",
"text",
"languages",
"corpus",
"semantic",
"word"
],
[
"regularization",
"generalization",
"overfitting",
"lasso",
"bias",
"dnns",
"regularized",
"loss",
"examples",
"benchmark"
],
[
"games",
"communication",
"strategies",
"agents",
"agent",
"environments",
"ai",
"game",
"multi",
"autonomous"
],
[
"noise",
"labeled",
"noisy",
"challenge",
"labels",
"label",
"corpus",
"labeling",
"multi",
"unlabeled"
],
[
"matrix",
"decomposition",
"factorization",
"rank",
"tensor",
"tensors",
"matrices",
"spectral",
"quantum",
"completion"
],
[
"accuracy",
"robust",
"estimate",
"estimation",
"uncertainty",
"safety",
"noise",
"imagenet",
"robustness",
"adversarially"
],
[
"learns",
"recommender",
"objects",
"learned",
"flow",
"examples",
"recommendations",
"robots",
"efficiency",
"uncertainty"
],
[
"dynamics",
"representations",
"learned",
"imitation",
"learn",
"self",
"3d",
"visual",
"class",
"forgetting"
]
] | 338.251861 | all-MiniLM-L6-v2 | 0.7 | -0.088703 | 0.162947 | 0.860016 |
ArXiv ML Papers | KeyNMF | 44 | 50 | [
[
"challenges",
"learning",
"generalization",
"meta",
"this",
"other",
"active",
"iterative",
"efficiently",
"several"
],
[
"neuron",
"backpropagation",
"layer",
"networks",
"input",
"neurons",
"architectures",
"recurrent",
"neural",
"weights"
],
[
"robustness",
"adversarial",
"attacks",
"attack",
"examples",
"robust",
"attacker",
"defense",
"adversarially",
"security"
],
[
"class",
"classification",
"classes",
"labels",
"accuracy",
"label",
"classify",
"benchmark",
"binary",
"svm"
],
[
"bias",
"models",
"predictions",
"estimation",
"interpretable",
"likelihood",
"many",
"modeling",
"overfitting",
"interpretability"
],
[
"optimization",
"convex",
"constraints",
"algorithm",
"objective",
"optimizing",
"minimization",
"optimize",
"search",
"constrained"
],
[
"benchmark",
"datasets",
"imagenet",
"overfitting",
"labels",
"bias",
"tensor",
"labeled",
"label",
"samples"
],
[
"gnns",
"networks",
"vertices",
"edges",
"graph",
"graphs",
"node",
"nodes",
"structure",
"gnn"
],
[
"networks",
"generate",
"adversarial",
"gan",
"gans",
"generative",
"generation",
"generator",
"generating",
"generated"
],
[
"machines",
"human",
"systems",
"computational",
"fairness",
"ai",
"machine",
"quantum",
"automated",
"ml"
],
[
"reward",
"reinforcement",
"learning",
"policies",
"policy",
"rl",
"control",
"action",
"planning",
"games"
],
[
"segmentation",
"layers",
"convolutions",
"convolution",
"pooling",
"convolutional",
"depth",
"layer",
"image",
"cnns"
],
[
"distillation",
"trained",
"input",
"layer",
"train",
"ai",
"brain",
"inputs",
"softmax",
"classify"
],
[
"gradient",
"stochastic",
"iterative",
"convex",
"convergence",
"batch",
"hessian",
"gradients",
"sgd",
"descent"
],
[
"predictive",
"predictions",
"predicted",
"predicting",
"predict",
"prediction",
"regression",
"predictors",
"future",
"performance"
],
[
"unlabeled",
"unsupervised",
"self",
"labeling",
"labels",
"label",
"annotation",
"semi",
"labeled",
"supervised"
],
[
"dnn",
"deep",
"accuracy",
"benchmarks",
"mri",
"dnns",
"crucial",
"net",
"batch",
"medical"
],
[
"dimensional",
"temporal",
"statistical",
"information",
"causal",
"data",
"health",
"empirical",
"labeled",
"dimensionality"
],
[
"feature",
"features",
"visual",
"explanations",
"selection",
"interpretability",
"dimensionality",
"focus",
"extraction",
"importance"
],
[
"prior",
"sampling",
"likelihood",
"bayesian",
"estimation",
"gaussian",
"posterior",
"priors",
"samples",
"uncertainty"
],
[
"latent",
"decoder",
"encoder",
"autoencoder",
"generative",
"variational",
"autoencoders",
"representation",
"learned",
"vae"
],
[
"challenge",
"accuracy",
"results",
"dataset",
"traffic",
"bias",
"imbalance",
"mnist",
"biases",
"identification"
],
[
"images",
"image",
"visual",
"vision",
"segmentation",
"imaging",
"resolution",
"3d",
"pixels",
"pixel"
],
[
"recognition",
"accuracy",
"speech",
"face",
"facial",
"audio",
"visual",
"bias",
"acoustic",
"performance"
],
[
"training",
"loss",
"train",
"normalization",
"transfer",
"generalization",
"weights",
"federated",
"examples",
"samples"
],
[
"bound",
"optimality",
"markov",
"minimax",
"estimation",
"sample",
"optimal",
"algorithm",
"generalization",
"complexity"
],
[
"layer",
"distributed",
"architectures",
"flow",
"layers",
"pruning",
"network",
"architecture",
"relu",
"communication"
],
[
"fairness",
"classes",
"metrics",
"adversarially",
"classifiers",
"classifying",
"svm",
"class",
"classifier",
"accuracy"
],
[
"hierarchical",
"algorithm",
"clustering",
"clusters",
"unsupervised",
"cluster",
"instances",
"analysis",
"patterns",
"matrix"
],
[
"sampling",
"matrix",
"complexity",
"adaptive",
"algorithm",
"algorithms",
"efficient",
"methods",
"search",
"efficiently"
],
[
"predictive",
"predictions",
"modeling",
"model",
"parameters",
"example",
"systems",
"architecture",
"dynamics",
"speech"
],
[
"bandit",
"regret",
"bandits",
"reward",
"exploration",
"strategy",
"rewards",
"exploitation",
"stochastic",
"and"
],
[
"embedding",
"search",
"nodes",
"similarity",
"space",
"metric",
"embeddings",
"vectors",
"dimensional",
"encoder"
],
[
"complexity",
"sparsity",
"thresholding",
"tensor",
"sparse",
"dimensional",
"gaussian",
"linear",
"matrix",
"factorization"
],
[
"privacy",
"private",
"federated",
"information",
"complexity",
"distributed",
"differentially",
"decentralized",
"differential",
"security"
],
[
"attacks",
"object",
"intrusion",
"detector",
"detecting",
"detect",
"detection",
"anomalies",
"anomaly",
"outlier"
],
[
"attention",
"vision",
"neural",
"encoder",
"transformer",
"recurrent",
"decoder",
"sequence",
"architecture",
"rnns"
],
[
"memory",
"performance",
"lstm",
"compression",
"quantization",
"processing",
"efficient",
"hardware",
"computation",
"computing"
],
[
"forecasts",
"forecasting",
"weather",
"predictions",
"forecast",
"series",
"probabilistic",
"predicting",
"ensemble",
"lstm"
],
[
"language",
"recurrent",
"corpus",
"languages",
"speech",
"text",
"nlp",
"words",
"semantic",
"word"
],
[
"cnns",
"imagenet",
"filters",
"cnn",
"pruned",
"pruning",
"convolution",
"accuracy",
"filter",
"vision"
],
[
"accuracy",
"tree",
"forest",
"performance",
"predictive",
"regression",
"trees",
"ensemble",
"boosting",
"gradient"
],
[
"generalization",
"overfitting",
"regularized",
"regularization",
"loss",
"lasso",
"label",
"examples",
"bias",
"dnns"
],
[
"labeling",
"task",
"challenge",
"tasks",
"forgetting",
"robot",
"robotic",
"challenging",
"information",
"vision"
],
[
"learned",
"knowledge",
"reasoning",
"transfer",
"distillation",
"unsupervised",
"target",
"student",
"domain",
"adaptation"
],
[
"parameters",
"probabilistic",
"latent",
"distributions",
"graphical",
"variational",
"flow",
"causal",
"stochastic",
"inference"
],
[
"learns",
"learned",
"recommender",
"examples",
"flow",
"objects",
"recommendations",
"uncertainty",
"robots",
"efficiency"
],
[
"agent",
"agents",
"communication",
"environments",
"game",
"games",
"multi",
"strategies",
"autonomous",
"distributed"
],
[
"objects",
"embeddings",
"representations",
"imagenet",
"representation",
"contrastive",
"semantic",
"visual",
"hierarchical",
"object"
],
[
"learn",
"imitation",
"learned",
"dynamics",
"input",
"class",
"3d",
"forgetting",
"predictors",
"self"
]
] | 345.798477 | all-MiniLM-L6-v2 | 0.688 | -0.078824 | 0.170102 | 0.855549 |
ArXiv ML Papers | KeyNMF | 45 | 50 | [
[
"ml",
"computational",
"machine",
"challenges",
"generalization",
"this",
"learning",
"quantum",
"other",
"fairness"
],
[
"neurons",
"recurrent",
"networks",
"neural",
"neuron",
"architectures",
"input",
"layer",
"backpropagation",
"weights"
],
[
"examples",
"attacks",
"attack",
"adversarial",
"robustness",
"robust",
"defense",
"attacker",
"adversarially",
"security"
],
[
"class",
"classes",
"classify",
"classification",
"accuracy",
"ensemble",
"benchmark",
"svm",
"binary",
"boosting"
],
[
"inference",
"many",
"predictions",
"interpretable",
"machine",
"likelihood",
"models",
"ai",
"interpretability",
"bias"
],
[
"algorithm",
"optimization",
"convex",
"constraints",
"objective",
"minimization",
"optimize",
"optimizing",
"constrained",
"search"
],
[
"benchmark",
"annotation",
"samples",
"fairness",
"overfitting",
"bias",
"imagenet",
"datasets",
"sampling",
"sentiment"
],
[
"graph",
"graphs",
"node",
"structure",
"edges",
"nodes",
"gnns",
"networks",
"vertices",
"gnn"
],
[
"networks",
"generate",
"generation",
"generative",
"generator",
"gan",
"gans",
"generated",
"generating",
"adversarial"
],
[
"robotic",
"task",
"tasks",
"robot",
"challenge",
"challenging",
"forgetting",
"human",
"vision",
"information"
],
[
"reward",
"policy",
"reinforcement",
"policies",
"learning",
"rl",
"control",
"action",
"planning",
"games"
],
[
"convolutions",
"convolution",
"segmentation",
"image",
"depth",
"convolutional",
"pooling",
"layer",
"layers",
"audio"
],
[
"layer",
"inputs",
"distillation",
"train",
"trained",
"input",
"classify",
"softmax",
"brain",
"objects"
],
[
"descent",
"gradient",
"stochastic",
"convergence",
"sgd",
"boosting",
"convex",
"batch",
"hessian",
"gradients"
],
[
"predicted",
"predict",
"predictive",
"performance",
"future",
"prediction",
"predictors",
"predictions",
"regression",
"predicting"
],
[
"semi",
"supervised",
"labeled",
"unlabeled",
"self",
"unsupervised",
"annotation",
"transformations",
"laplacian",
"annotated"
],
[
"deep",
"dnn",
"dnns",
"accuracy",
"mri",
"benchmarks",
"crucial",
"medical",
"net",
"batch"
],
[
"statistical",
"health",
"information",
"dimensional",
"data",
"causal",
"temporal",
"empirical",
"analysis",
"samples"
],
[
"interpretability",
"feature",
"selection",
"visual",
"features",
"explanations",
"tree",
"importance",
"dimensionality",
"extraction"
],
[
"bayesian",
"inference",
"prior",
"posterior",
"sampling",
"likelihood",
"distributions",
"gaussian",
"estimation",
"variational"
],
[
"decoder",
"generative",
"autoencoder",
"encoder",
"autoencoders",
"latent",
"variational",
"vae",
"representation",
"learned"
],
[
"bias",
"challenge",
"imbalance",
"dataset",
"results",
"identification",
"biases",
"traffic",
"mnist",
"accuracy"
],
[
"segmentation",
"image",
"images",
"visual",
"vision",
"resolution",
"3d",
"imaging",
"pixels",
"pixel"
],
[
"face",
"accuracy",
"speech",
"recognition",
"audio",
"bias",
"visual",
"acoustic",
"facial",
"recognize"
],
[
"training",
"loss",
"weights",
"transfer",
"train",
"federated",
"normalization",
"generalization",
"examples",
"samples"
],
[
"minimax",
"complexity",
"optimality",
"optimal",
"sample",
"algorithm",
"bound",
"generalization",
"estimation",
"markov"
],
[
"layer",
"layers",
"pruning",
"network",
"flow",
"distributed",
"architectures",
"communication",
"architecture",
"relu"
],
[
"classifier",
"accuracy",
"classifying",
"class",
"classes",
"fairness",
"bayes",
"machine",
"method",
"nearest"
],
[
"clustering",
"cluster",
"algorithm",
"clusters",
"analysis",
"unsupervised",
"instances",
"hierarchical",
"patterns",
"segmentation"
],
[
"efficient",
"algorithms",
"efficiently",
"adaptive",
"search",
"methods",
"algorithm",
"sampling",
"matrix",
"complexity"
],
[
"knowledge",
"domain",
"adaptation",
"unsupervised",
"transfer",
"reasoning",
"learned",
"distillation",
"student",
"target"
],
[
"rewards",
"exploration",
"reward",
"regret",
"bandits",
"bandit",
"exploitation",
"sampling",
"strategy",
"stochastic"
],
[
"representations",
"representation",
"visual",
"imagenet",
"objects",
"semantic",
"object",
"embeddings",
"hierarchical",
"contrastive"
],
[
"sparse",
"factorization",
"thresholding",
"linear",
"matrix",
"sparsity",
"dimensional",
"gaussian",
"coding",
"compressed"
],
[
"privacy",
"complexity",
"federated",
"distributed",
"information",
"differential",
"decentralized",
"differentially",
"private",
"security"
],
[
"detector",
"attacks",
"intrusion",
"anomaly",
"detect",
"object",
"detecting",
"detection",
"outlier",
"anomalies"
],
[
"attention",
"encoder",
"recurrent",
"decoder",
"vision",
"sequence",
"transformer",
"neural",
"architecture",
"rnns"
],
[
"memory",
"performance",
"quantization",
"efficient",
"compression",
"computational",
"lstm",
"accuracy",
"hardware",
"processing"
],
[
"speech",
"recurrent",
"language",
"nlp",
"text",
"corpus",
"languages",
"semantic",
"word",
"words"
],
[
"parameters",
"predictive",
"predictions",
"modeling",
"model",
"systems",
"example",
"dynamics",
"architecture",
"speech"
],
[
"cnns",
"imagenet",
"cnn",
"pruning",
"filters",
"convolution",
"filter",
"pruned",
"vision",
"accuracy"
],
[
"forecasting",
"forecast",
"ensemble",
"weather",
"forecasts",
"accuracy",
"predicting",
"boosting",
"predictions",
"forest"
],
[
"embedding",
"embeddings",
"space",
"dimensional",
"nodes",
"vectors",
"metric",
"search",
"encoder",
"similarity"
],
[
"dnns",
"bias",
"lasso",
"robust",
"overfitting",
"loss",
"regularization",
"generalization",
"regularized",
"examples"
],
[
"agent",
"agents",
"environments",
"ai",
"communication",
"games",
"game",
"strategies",
"autonomous",
"multi"
],
[
"label",
"labeled",
"multi",
"labeling",
"labels",
"noisy",
"challenge",
"corpus",
"noise",
"unlabeled"
],
[
"decomposition",
"factorization",
"matrices",
"tensors",
"rank",
"matrix",
"completion",
"tensor",
"quantum",
"complexity"
],
[
"accuracy",
"kernel",
"security",
"bias",
"svm",
"classifiers",
"recognize",
"dimensionality",
"metrics",
"ai"
],
[
"learn",
"learned",
"class",
"imitation",
"self",
"3d",
"forgetting",
"input",
"dynamics",
"predictors"
],
[
"objects",
"examples",
"robots",
"uncertainty",
"flow",
"learned",
"learns",
"recommender",
"recommendations",
"efficiency"
]
] | 314.850515 | all-MiniLM-L6-v2 | 0.694 | -0.089638 | 0.163796 | 0.858691 |
ArXiv ML Papers | KeyNMF | 46 | 50 | [
[
"machine",
"generalization",
"learning",
"this",
"ml",
"challenges",
"quantum",
"fairness",
"automated",
"ensemble"
],
[
"neuron",
"architectures",
"neural",
"neurons",
"recurrent",
"networks",
"layer",
"layers",
"input",
"backpropagation"
],
[
"attack",
"attacks",
"robust",
"robustness",
"adversarial",
"defense",
"attacker",
"adversarially",
"examples",
"security"
],
[
"benchmark",
"ensemble",
"classification",
"class",
"classes",
"accuracy",
"classify",
"binary",
"svm",
"forest"
],
[
"inference",
"machine",
"ai",
"interpretable",
"bias",
"predictions",
"models",
"many",
"likelihood",
"interpretability"
],
[
"sample",
"distributed",
"efficient",
"adaptive",
"search",
"edge",
"algorithm",
"complexity",
"minimizing",
"computing"
],
[
"benchmark",
"fairness",
"datasets",
"imagenet",
"sentiment",
"annotation",
"overfitting",
"bias",
"algorithmic",
"world"
],
[
"graph",
"graphs",
"node",
"structure",
"edges",
"networks",
"gnns",
"nodes",
"vertices",
"gnn"
],
[
"gan",
"generation",
"generating",
"generative",
"networks",
"generator",
"generate",
"adversarial",
"gans",
"generated"
],
[
"objects",
"robotic",
"videos",
"robot",
"autonomous",
"object",
"planning",
"driving",
"visual",
"robotics"
],
[
"reward",
"learning",
"policy",
"reinforcement",
"rl",
"policies",
"action",
"control",
"games",
"atari"
],
[
"image",
"visual",
"segmentation",
"pixels",
"vision",
"images",
"resolution",
"pixel",
"imaging",
"maps"
],
[
"layer",
"trained",
"input",
"train",
"brain",
"inputs",
"distillation",
"classify",
"softmax",
"layers"
],
[
"convergence",
"gradient",
"stochastic",
"boosting",
"sgd",
"descent",
"convex",
"batch",
"hessian",
"gradients"
],
[
"predictions",
"predict",
"prediction",
"predictive",
"predicted",
"predicting",
"regression",
"performance",
"predictors",
"lstm"
],
[
"unlabeled",
"unsupervised",
"labeled",
"supervised",
"semi",
"annotation",
"self",
"annotated",
"transformations",
"laplacian"
],
[
"dnns",
"accuracy",
"dnn",
"batch",
"benchmarks",
"crucial",
"deep",
"medical",
"mri",
"net"
],
[
"data",
"dimensional",
"statistical",
"causal",
"health",
"information",
"temporal",
"analysis",
"dimensionality",
"empirical"
],
[
"feature",
"selection",
"visual",
"interpretability",
"features",
"explanations",
"importance",
"dimensionality",
"extraction",
"tree"
],
[
"variational",
"inference",
"likelihood",
"gaussian",
"prior",
"bayesian",
"posterior",
"probabilistic",
"distributions",
"estimation"
],
[
"variational",
"decoder",
"autoencoders",
"autoencoder",
"vae",
"learned",
"generative",
"representation",
"encoder",
"latent"
],
[
"dataset",
"challenge",
"mnist",
"results",
"accuracy",
"bias",
"traffic",
"biases",
"imbalance",
"identification"
],
[
"convex",
"optimizing",
"constrained",
"optimize",
"constraints",
"minimization",
"objective",
"optimization",
"optimizer",
"search"
],
[
"accuracy",
"face",
"bias",
"speech",
"recognition",
"facial",
"audio",
"acoustic",
"visual",
"recognize"
],
[
"train",
"transfer",
"weights",
"training",
"generalization",
"loss",
"federated",
"normalization",
"examples",
"improvements"
],
[
"optimality",
"optimal",
"minimax",
"complexity",
"estimation",
"generalization",
"bound",
"approximation",
"sample",
"markov"
],
[
"layer",
"flow",
"layers",
"network",
"pruning",
"distributed",
"architecture",
"communication",
"architectures",
"relu"
],
[
"classifier",
"classifiers",
"accuracy",
"svm",
"class",
"fairness",
"classes",
"adversarially",
"classifying",
"performance"
],
[
"clusters",
"clustering",
"cluster",
"unsupervised",
"analysis",
"segmentation",
"instances",
"nodes",
"patterns",
"hierarchical"
],
[
"efficiently",
"search",
"algorithms",
"methods",
"performance",
"machine",
"minimization",
"matrix",
"trees",
"boosting"
],
[
"tasks",
"forgetting",
"task",
"information",
"ai",
"challenge",
"representation",
"challenging",
"human",
"labeling"
],
[
"rewards",
"reward",
"stochastic",
"strategy",
"exploration",
"regret",
"bandit",
"bandits",
"exploitation",
"and"
],
[
"contrastive",
"imagenet",
"hierarchical",
"objects",
"semantic",
"representations",
"representation",
"embeddings",
"learns",
"visual"
],
[
"sparsity",
"thresholding",
"sparse",
"dimensional",
"factorization",
"matrix",
"linear",
"coding",
"compressed",
"gaussian"
],
[
"privacy",
"private",
"distributed",
"federated",
"information",
"decentralized",
"complexity",
"differentially",
"differential",
"security"
],
[
"detection",
"detect",
"intrusion",
"anomaly",
"detecting",
"object",
"attacks",
"outlier",
"anomalies",
"detector"
],
[
"vision",
"sequence",
"attention",
"architecture",
"decoder",
"encoder",
"neural",
"recurrent",
"transformer",
"rnns"
],
[
"hardware",
"quantization",
"efficient",
"compression",
"accuracy",
"performance",
"memory",
"lstm",
"imagenet",
"computational"
],
[
"recurrent",
"language",
"text",
"languages",
"nlp",
"corpus",
"speech",
"words",
"semantic",
"word"
],
[
"model",
"modeling",
"predictive",
"parameters",
"predictions",
"example",
"dynamics",
"architecture",
"systems",
"speech"
],
[
"convolution",
"convolutional",
"cnn",
"cnns",
"layers",
"vision",
"pooling",
"imagenet",
"filters",
"filter"
],
[
"forecasting",
"ensemble",
"accuracy",
"forecasts",
"forecast",
"weather",
"predicting",
"boosting",
"predictions",
"forest"
],
[
"metric",
"nodes",
"embedding",
"vectors",
"space",
"search",
"dimensional",
"embeddings",
"similarity",
"encoder"
],
[
"regularized",
"lasso",
"overfitting",
"dnns",
"loss",
"bias",
"regularization",
"generalization",
"robust",
"examples"
],
[
"reasoning",
"transfer",
"domain",
"student",
"unsupervised",
"knowledge",
"distillation",
"adaptation",
"learned",
"target"
],
[
"labeling",
"noisy",
"challenge",
"labeled",
"labels",
"label",
"multi",
"noise",
"corpus",
"unlabeled"
],
[
"sampling",
"sample",
"monte",
"random",
"samples",
"markov",
"carlo",
"mcmc",
"adaptive",
"computational"
],
[
"communication",
"games",
"environments",
"agent",
"game",
"agents",
"ai",
"strategies",
"multi",
"distributed"
],
[
"matrix",
"quantum",
"factorization",
"tensors",
"spectral",
"matrices",
"decomposition",
"tensor",
"completion",
"rank"
],
[
"learn",
"imitation",
"dynamics",
"learning",
"examples",
"learns",
"input",
"learned",
"expert",
"flow"
]
] | 298.769343 | all-MiniLM-L6-v2 | 0.706 | -0.083808 | 0.167206 | 0.861095 |
ArXiv ML Papers | FASTopic | 43 | 10 | [
[
"the",
"and",
"to",
"in",
"of",
"we",
"on",
"for",
"that",
"is"
],
[
"policy",
"reinforcement",
"control",
"agent",
"rl",
"agents",
"reward",
"decision",
"regret",
"policies"
],
[
"disease",
"patient",
"health",
"forecasting",
"day",
"software",
"monitoring",
"ml",
"forecast",
"forest"
],
[
"stochastic",
"convergence",
"linear",
"matrix",
"bounds",
"complexity",
"bound",
"regression",
"optimization",
"sample"
],
[
"object",
"visual",
"video",
"segmentation",
"image",
"images",
"3d",
"reconstruction",
"brain",
"shape"
],
[
"wireless",
"device",
"frequency",
"hardware",
"transmission",
"separation",
"gpu",
"library",
"accelerators",
"throughput"
],
[
"text",
"audio",
"graph",
"speech",
"embeddings",
"graphs",
"language",
"word",
"node",
"embedding"
],
[
"tasks",
"task",
"learn",
"representation",
"multi",
"domain",
"attention",
"knowledge",
"representations",
"level"
],
[
"architecture",
"layer",
"train",
"adversarial",
"against",
"robustness",
"convolutional",
"attacks",
"architectures",
"examples"
],
[
"uncertainty",
"variational",
"distributions",
"bayesian",
"estimation",
"likelihood",
"quantum",
"probabilistic",
"density",
"processes"
]
] | 257.922492 | all-MiniLM-L6-v2 | 1 | -0.0561 | 0.173073 | 0.857041 |
ArXiv ML Papers | FASTopic | 44 | 10 | [
[
"speaker",
"gpu",
"character",
"audio",
"speech",
"fusion",
"acoustic",
"asr",
"enhancement",
"patch"
],
[
"interpretable",
"latent",
"bias",
"gans",
"generative",
"interpretability",
"autoencoders",
"biases",
"vae",
"explanations"
],
[
"stochastic",
"convergence",
"linear",
"algorithm",
"sample",
"optimal",
"optimization",
"complexity",
"function",
"estimation"
],
[
"to",
"the",
"and",
"of",
"in",
"on",
"that",
"this",
"we",
"is"
],
[
"supervised",
"representations",
"text",
"language",
"human",
"domain",
"images",
"representation",
"attention",
"source"
],
[
"agent",
"reinforcement",
"policy",
"rl",
"control",
"agents",
"reward",
"action",
"policies",
"environment"
],
[
"forecasting",
"monitoring",
"covid",
"series",
"forecast",
"sensors",
"day",
"forest",
"health",
"media"
],
[
"node",
"users",
"privacy",
"quantum",
"user",
"federated",
"graph",
"clustering",
"graphs",
"nodes"
],
[
"neural",
"networks",
"attacks",
"layer",
"architecture",
"adversarial",
"network",
"convolutional",
"memory",
"robustness"
],
[
"norm",
"covariance",
"kernel",
"lipschitz",
"integer",
"subspace",
"principal",
"mathbb",
"uniform",
"accelerated"
]
] | 256.159314 | all-MiniLM-L6-v2 | 1 | -0.098339 | 0.175223 | 0.837138 |
ArXiv ML Papers | FASTopic | 45 | 10 | [
[
"inference",
"variational",
"distributions",
"probabilistic",
"variables",
"bayesian",
"likelihood",
"estimation",
"latent",
"density"
],
[
"agents",
"agent",
"environment",
"reward",
"control",
"action",
"policy",
"rl",
"reinforcement",
"policies"
],
[
"disease",
"recommendation",
"social",
"user",
"health",
"patients",
"users",
"patient",
"cancer",
"relational"
],
[
"segmentation",
"neural",
"layers",
"networks",
"network",
"convolutional",
"layer",
"architectures",
"architecture",
"memory"
],
[
"day",
"svm",
"asr",
"eeg",
"speech",
"audio",
"covid",
"speaker",
"fusion",
"acoustic"
],
[
"optimal",
"algorithm",
"linear",
"gradient",
"optimization",
"function",
"stochastic",
"complexity",
"sample",
"problems"
],
[
"data",
"different",
"machine",
"adversarial",
"prediction",
"model",
"models",
"system",
"performance",
"to"
],
[
"pruning",
"invariant",
"teacher",
"smoothing",
"tensors",
"compressed",
"pruned",
"aggregate",
"deeper",
"orthogonal"
],
[
"domain",
"representations",
"language",
"supervised",
"attention",
"representation",
"text",
"tasks",
"source",
"feature"
],
[
"bound",
"regret",
"minimax",
"bounds",
"sqrt",
"bounded",
"epsilon",
"convex",
"private",
"bandit"
]
] | 255.266811 | all-MiniLM-L6-v2 | 1 | -0.04863 | 0.159003 | 0.856226 |
ArXiv ML Papers | FASTopic | 46 | 10 | [
[
"policy",
"reinforcement",
"agent",
"rl",
"reward",
"control",
"agents",
"action",
"quantum",
"policies"
],
[
"dataset",
"different",
"model",
"system",
"task",
"machine",
"models",
"accuracy",
"data",
"detection"
],
[
"kernels",
"integer",
"covariance",
"characterization",
"regime",
"relaxation",
"squared",
"lasso",
"exchange",
"arbitrarily"
],
[
"words",
"text",
"language",
"content",
"word",
"natural",
"style",
"bert",
"languages",
"media"
],
[
"robustness",
"attacks",
"attack",
"adversarial",
"against",
"perturbations",
"examples",
"robust",
"defense",
"security"
],
[
"convolutional",
"attention",
"neural",
"architecture",
"images",
"generative",
"image",
"layer",
"self",
"train"
],
[
"protein",
"series",
"health",
"patients",
"disease",
"was",
"forecasting",
"clinical",
"patient",
"forest"
],
[
"graphs",
"graph",
"user",
"clustering",
"metric",
"privacy",
"feature",
"label",
"classifier",
"users"
],
[
"linear",
"sample",
"optimization",
"optimal",
"function",
"stochastic",
"algorithm",
"functions",
"gradient",
"convergence"
],
[
"traffic",
"devices",
"shape",
"3d",
"driving",
"mobile",
"hardware",
"sensor",
"vehicle",
"motion"
]
] | 254.301818 | all-MiniLM-L6-v2 | 1 | -0.07566 | 0.152227 | 0.871059 |
ArXiv ML Papers | FASTopic | 43 | 20 | [
[
"network",
"neural",
"deep",
"architecture",
"memory",
"accuracy",
"system",
"power",
"systems",
"networks"
],
[
"task",
"tasks",
"knowledge",
"domain",
"learn",
"transfer",
"target",
"domains",
"learned",
"source"
],
[
"the",
"of",
"and",
"to",
"in",
"this",
"that",
"on",
"as",
"are"
],
[
"recommendation",
"causal",
"recommender",
"fair",
"outcomes",
"fairness",
"user",
"users",
"market",
"item"
],
[
"facial",
"camera",
"invariance",
"eye",
"localization",
"optical",
"recorded",
"svhn",
"detector",
"concern"
],
[
"text",
"language",
"audio",
"speech",
"word",
"words",
"style",
"speaker",
"media",
"languages"
],
[
"estimation",
"bayesian",
"uncertainty",
"variational",
"distributions",
"inference",
"likelihood",
"posterior",
"density",
"monte"
],
[
"gnn",
"physics",
"equations",
"quantum",
"chemical",
"molecular",
"scientific",
"dynamical",
"formalism",
"directed"
],
[
"perturbations",
"attacks",
"adversarial",
"examples",
"against",
"attack",
"dnns",
"robustness",
"mnist",
"overfitting"
],
[
"algorithm",
"linear",
"function",
"optimization",
"sample",
"problems",
"complexity",
"bounds",
"regression",
"gradient"
],
[
"width",
"sgd",
"accelerated",
"distributed",
"varepsilon",
"pac",
"delta",
"descent",
"ascent",
"hessian"
],
[
"imbalance",
"split",
"forgetting",
"setup",
"gradually",
"proposals",
"coverage",
"updated",
"suite",
"scratch"
],
[
"test",
"machine",
"classifier",
"dataset",
"labels",
"label",
"bias",
"research",
"classification",
"ml"
],
[
"tensor",
"factorization",
"trees",
"hundreds",
"tensors",
"accuracies",
"mining",
"millions",
"streams",
"percentage"
],
[
"gan",
"convolutional",
"images",
"video",
"segmentation",
"object",
"vae",
"generative",
"visual",
"image"
],
[
"engineering",
"programs",
"logic",
"practices",
"explanations",
"explanation",
"software",
"project",
"explaining",
"explainable"
],
[
"graph",
"series",
"graphs",
"clustering",
"structure",
"node",
"nodes",
"feature",
"embedding",
"metric"
],
[
"subspace",
"rank",
"sparse",
"matrix",
"recovery",
"sensing",
"subspaces",
"laplacian",
"projections",
"principal"
],
[
"agents",
"agent",
"policy",
"control",
"reinforcement",
"rl",
"reward",
"action",
"regret",
"policies"
],
[
"was",
"cancer",
"patients",
"patient",
"clinical",
"covid",
"day",
"eeg",
"forest",
"ct"
]
] | 3,251.426311 | all-MiniLM-L6-v2 | 1 | -0.133431 | 0.155734 | 0.856229 |
ArXiv ML Papers | FASTopic | 44 | 20 | [
[
"trained",
"training",
"image",
"supervised",
"images",
"train",
"domain",
"loss",
"segmentation",
"generative"
],
[
"imputation",
"compliance",
"fairness",
"bias",
"interventions",
"fair",
"biases",
"individuals",
"credit",
"candidates"
],
[
"perturbations",
"attacks",
"attack",
"robustness",
"gans",
"against",
"adversarial",
"defense",
"attacker",
"perturbation"
],
[
"any",
"optimal",
"algorithms",
"optimization",
"algorithm",
"constraints",
"problems",
"function",
"gradient",
"sample"
],
[
"sparse",
"regression",
"clustering",
"dimensional",
"linear",
"random",
"kernel",
"low",
"selection",
"noise"
],
[
"coding",
"cell",
"decoder",
"vae",
"labeling",
"pruned",
"par",
"encoders",
"imbalance",
"prominent"
],
[
"boosting",
"forecasting",
"forest",
"forecast",
"forecasts",
"market",
"weather",
"xgboost",
"year",
"price"
],
[
"posterior",
"bayesian",
"variational",
"distributions",
"mutual",
"mcmc",
"bandits",
"nonparametric",
"delta",
"interval"
],
[
"energy",
"memory",
"privacy",
"channel",
"power",
"distributed",
"devices",
"traffic",
"efficiency",
"federated"
],
[
"language",
"text",
"content",
"word",
"words",
"natural",
"style",
"languages",
"bert",
"media"
],
[
"convolutional",
"video",
"resolution",
"object",
"convolution",
"cnn",
"3d",
"scene",
"pooling",
"net"
],
[
"detection",
"prediction",
"was",
"research",
"system",
"human",
"machine",
"dataset",
"were",
"systems"
],
[
"agents",
"rl",
"policy",
"action",
"games",
"actions",
"reward",
"reinforcement",
"agent",
"environment"
],
[
"signals",
"cancer",
"speech",
"audio",
"frequency",
"speaker",
"eeg",
"acoustic",
"separation",
"magnetic"
],
[
"entries",
"projections",
"principal",
"nonnegative",
"completion",
"tensor",
"recovery",
"sketch",
"overlapping",
"algebra"
],
[
"patient",
"reports",
"clinical",
"calibration",
"patients",
"health",
"disease",
"assessment",
"explanation",
"diagnosis"
],
[
"flow",
"neurons",
"physics",
"quantum",
"inverse",
"dynamical",
"flows",
"equations",
"normalizing",
"chemical"
],
[
"representations",
"representation",
"structure",
"graphs",
"latent",
"sequence",
"graph",
"node",
"learn",
"topological"
],
[
"with",
"this",
"in",
"for",
"to",
"and",
"the",
"on",
"of",
"is"
],
[
"descent",
"convex",
"convergence",
"sqrt",
"regime",
"max",
"minimax",
"sgd",
"tilde",
"proximal"
]
] | 3,194.900422 | all-MiniLM-L6-v2 | 1 | -0.102096 | 0.157628 | 0.876176 |
ArXiv ML Papers | FASTopic | 45 | 20 | [
[
"music",
"biology",
"cognitive",
"dirichlet",
"association",
"preference",
"observational",
"described",
"graphical",
"inferring"
],
[
"text",
"speech",
"language",
"representations",
"semantic",
"word",
"audio",
"attention",
"source",
"translation"
],
[
"combinatorial",
"bandits",
"armed",
"greedy",
"tight",
"bandit",
"theoretic",
"ascent",
"oracle",
"exchange"
],
[
"bit",
"compression",
"compressed",
"quantization",
"device",
"resnet",
"gpu",
"scaling",
"distortion",
"pruned"
],
[
"dynamical",
"equations",
"forward",
"quantum",
"activation",
"physics",
"flows",
"neuron",
"relu",
"neurons"
],
[
"user",
"machine",
"how",
"techniques",
"systems",
"human",
"system",
"research",
"ml",
"privacy"
],
[
"temporal",
"recurrent",
"rnn",
"lstm",
"acoustic",
"eeg",
"conversion",
"urban",
"year",
"rnns"
],
[
"attacks",
"against",
"robustness",
"examples",
"adversarial",
"attack",
"detection",
"supervised",
"samples",
"robust"
],
[
"algorithm",
"convergence",
"optimal",
"optimization",
"stochastic",
"gradient",
"linear",
"complexity",
"bounds",
"sample"
],
[
"bayesian",
"prediction",
"variational",
"inference",
"estimation",
"series",
"distribution",
"uncertainty",
"variables",
"latent"
],
[
"cluster",
"trees",
"distance",
"clusters",
"metric",
"nearest",
"clustering",
"outlier",
"mining",
"neighbor"
],
[
"images",
"image",
"object",
"video",
"segmentation",
"visual",
"3d",
"imaging",
"vision",
"scene"
],
[
"deep",
"memory",
"architecture",
"networks",
"neural",
"network",
"convolutional",
"layer",
"end",
"accuracy"
],
[
"action",
"agent",
"rl",
"control",
"reinforcement",
"policies",
"policy",
"reward",
"agents",
"environments"
],
[
"traffic",
"distributed",
"decentralized",
"server",
"intelligent",
"internet",
"vehicle",
"wireless",
"smart",
"centralized"
],
[
"health",
"patient",
"software",
"disease",
"clinical",
"forest",
"reports",
"industry",
"assessment",
"healthcare"
],
[
"fairness",
"learners",
"decision",
"label",
"biases",
"bias",
"vs",
"statistically",
"explaining",
"rarely"
],
[
"for",
"of",
"the",
"we",
"to",
"and",
"in",
"that",
"is",
"this"
],
[
"graph",
"graphs",
"nodes",
"relational",
"gnns",
"node",
"edges",
"neighborhood",
"gnn",
"vertex"
],
[
"covariance",
"kernel",
"matrix",
"sparse",
"norm",
"rank",
"dimensionality",
"tensor",
"subspace",
"subspaces"
]
] | 3,218.907335 | all-MiniLM-L6-v2 | 1 | -0.116578 | 0.169421 | 0.873183 |
ArXiv ML Papers | FASTopic | 46 | 20 | [
[
"explanations",
"bias",
"explanation",
"medicine",
"explaining",
"practices",
"care",
"imputation",
"diseases",
"genetic"
],
[
"relational",
"gnn",
"links",
"topics",
"graph",
"molecular",
"vertices",
"vertex",
"passing",
"graphs"
],
[
"segmentation",
"3d",
"brain",
"imaging",
"eeg",
"separation",
"mri",
"magnetic",
"eye",
"localization"
],
[
"gan",
"autoencoders",
"generative",
"generation",
"image",
"gans",
"autoencoder",
"augmentation",
"unlabeled",
"vae"
],
[
"policy",
"control",
"agent",
"agents",
"rl",
"decision",
"reinforcement",
"reward",
"action",
"environments"
],
[
"nodes",
"sparse",
"low",
"distributed",
"node",
"rank",
"matrix",
"clustering",
"tensor",
"centralized"
],
[
"the",
"of",
"to",
"and",
"on",
"as",
"this",
"in",
"that",
"are"
],
[
"convolutional",
"neural",
"architecture",
"networks",
"layer",
"network",
"deep",
"architectures",
"layers",
"train"
],
[
"representations",
"domain",
"knowledge",
"source",
"tasks",
"transfer",
"representation",
"task",
"target",
"visual"
],
[
"fair",
"pac",
"bandits",
"sqrt",
"mathcal",
"bandit",
"regret",
"tilde",
"exchange",
"tight"
],
[
"speech",
"transformer",
"acoustic",
"audio",
"rnn",
"video",
"transformers",
"speaker",
"frame",
"gpu"
],
[
"robotic",
"reaching",
"robots",
"forgetting",
"robotics",
"imitation",
"updated",
"navigation",
"inefficient",
"encourages"
],
[
"media",
"word",
"text",
"answering",
"language",
"words",
"nlp",
"bert",
"sentences",
"counterfactual"
],
[
"mcmc",
"normalizing",
"monte",
"gps",
"inverse",
"algebra",
"sketch",
"carlo",
"equations",
"approximated"
],
[
"dataset",
"detection",
"classifier",
"classification",
"test",
"label",
"features",
"feature",
"were",
"accuracy"
],
[
"series",
"uncertainty",
"modeling",
"inference",
"variables",
"estimation",
"latent",
"bayesian",
"prediction",
"complex"
],
[
"iot",
"traffic",
"consumption",
"device",
"smart",
"library",
"wireless",
"mobile",
"sensors",
"market"
],
[
"bounds",
"stochastic",
"convergence",
"optimization",
"sampling",
"gradient",
"linear",
"function",
"sample",
"functions"
],
[
"relu",
"sgd",
"initialization",
"subspace",
"accelerated",
"norm",
"width",
"mathbb",
"quadratic",
"hessian"
],
[
"attack",
"robustness",
"adversarial",
"against",
"perturbations",
"attacks",
"security",
"defense",
"perturbation",
"attacker"
]
] | 3,199.345201 | all-MiniLM-L6-v2 | 1 | -0.121864 | 0.15734 | 0.873345 |
ArXiv ML Papers | FASTopic | 43 | 30 | [
[
"events",
"series",
"forecasting",
"lstm",
"forecast",
"weather",
"forecasts",
"market",
"anomalies",
"sensors"
],
[
"security",
"attack",
"defense",
"gans",
"gan",
"robustness",
"attacks",
"perturbations",
"against",
"adversarial"
],
[
"combinatorial",
"varepsilon",
"epsilon",
"mathcal",
"minimax",
"sqrt",
"delta",
"bandit",
"convex",
"regret"
],
[
"reconstructing",
"backpropagation",
"solvers",
"hessian",
"inverse",
"width",
"derivatives",
"equations",
"differentiation",
"optimisation"
],
[
"bayesian",
"gradient",
"sample",
"function",
"distributions",
"distribution",
"estimation",
"functions",
"bounds",
"stochastic"
],
[
"industry",
"platform",
"developers",
"platforms",
"things",
"software",
"library",
"technology",
"intrusion",
"engine"
],
[
"gaussian",
"estimators",
"kernels",
"uniform",
"kernel",
"regime",
"regression",
"covariance",
"nonparametric",
"squared"
],
[
"distributed",
"communication",
"privacy",
"federated",
"transmission",
"private",
"server",
"decentralized",
"wireless",
"messages"
],
[
"to",
"we",
"the",
"is",
"of",
"and",
"for",
"in",
"that",
"on"
],
[
"recommender",
"bias",
"social",
"ranking",
"communities",
"individuals",
"fairness",
"fair",
"item",
"interventions"
],
[
"segmentation",
"scene",
"pixel",
"images",
"convolutions",
"3d",
"gpu",
"patches",
"accelerators",
"segment"
],
[
"words",
"audio",
"language",
"text",
"speech",
"media",
"speaker",
"word",
"bert",
"acoustic"
],
[
"patients",
"protein",
"health",
"covid",
"patient",
"clinical",
"cancer",
"disease",
"cell",
"ct"
],
[
"manifold",
"clustering",
"rank",
"dimensional",
"matrix",
"distance",
"tensor",
"sparse",
"clusters",
"dimensionality"
],
[
"additive",
"assisted",
"python",
"hyper",
"streams",
"compress",
"parallel",
"plus",
"encourages",
"factorization"
],
[
"supervised",
"training",
"image",
"labels",
"datasets",
"loss",
"trained",
"images",
"pre",
"labeled"
],
[
"human",
"user",
"systems",
"research",
"how",
"about",
"system",
"machine",
"knowledge",
"world"
],
[
"vehicles",
"driving",
"dynamics",
"motion",
"trajectory",
"control",
"robot",
"dynamical",
"trajectories",
"imitation"
],
[
"physics",
"quantum",
"logic",
"chemical",
"materials",
"imposes",
"describes",
"promise",
"formalism",
"specify"
],
[
"learn",
"task",
"representations",
"representation",
"generative",
"inference",
"models",
"modeling",
"latent",
"tasks"
],
[
"orthogonal",
"transform",
"frequencies",
"variation",
"spectrum",
"spatially",
"sketch",
"frequency",
"phases",
"outliers"
],
[
"graph",
"node",
"vertex",
"nodes",
"relational",
"gnn",
"gnns",
"graphs",
"attributed",
"vertices"
],
[
"transfer",
"video",
"object",
"visual",
"domain",
"re",
"videos",
"objects",
"supervision",
"attribute"
],
[
"architecture",
"layer",
"network",
"networks",
"deep",
"neural",
"architectures",
"convolutional",
"layers",
"input"
],
[
"queries",
"selection",
"nearest",
"query",
"bayes",
"tail",
"pac",
"subsets",
"percentage",
"tree"
],
[
"questions",
"explanation",
"counterfactual",
"explanations",
"explainable",
"explaining",
"intended",
"credit",
"compliance",
"consequences"
],
[
"translation",
"rnn",
"sequence",
"recurrent",
"decoder",
"vae",
"decoding",
"rnns",
"autoregressive",
"translate"
],
[
"test",
"classifier",
"feature",
"features",
"detection",
"classification",
"prediction",
"ensemble",
"was",
"accuracy"
],
[
"reward",
"reinforcement",
"agent",
"agents",
"policy",
"rl",
"action",
"actions",
"games",
"game"
],
[
"up",
"efficiency",
"memory",
"efficient",
"search",
"consumption",
"cost",
"energy",
"iot",
"computational"
]
] | 3,296.720755 | all-MiniLM-L6-v2 | 0.996667 | -0.140893 | 0.166276 | 0.882703 |
ArXiv ML Papers | FASTopic | 44 | 30 | [
[
"matrix",
"rank",
"graphs",
"graph",
"clustering",
"structure",
"nodes",
"node",
"sparse",
"completion"
],
[
"distributions",
"variational",
"estimation",
"bayesian",
"inference",
"uncertainty",
"posterior",
"likelihood",
"monte",
"carlo"
],
[
"user",
"recommendation",
"ranking",
"topic",
"sentences",
"item",
"answering",
"media",
"nlp",
"recommender"
],
[
"protein",
"health",
"patients",
"disease",
"19",
"clinical",
"medical",
"patient",
"explanation",
"covid"
],
[
"convolutional",
"layer",
"network",
"neural",
"networks",
"architecture",
"deep",
"layers",
"architectures",
"cnn"
],
[
"mining",
"subspace",
"curse",
"kernel",
"regression",
"partitioning",
"kernels",
"neighbor",
"nearest",
"centers"
],
[
"facial",
"voice",
"asr",
"eeg",
"character",
"speech",
"speaker",
"acoustic",
"95",
"28"
],
[
"links",
"communities",
"message",
"molecular",
"nonnegative",
"passing",
"localized",
"directed",
"relational",
"vertices"
],
[
"trust",
"biases",
"serious",
"claim",
"susceptible",
"malicious",
"bias",
"outlier",
"assigned",
"measuring"
],
[
"labels",
"imbalance",
"supervision",
"unlabeled",
"supervised",
"annotation",
"label",
"labeled",
"labeling",
"patches"
],
[
"regret",
"fairness",
"communication",
"fair",
"bandit",
"learner",
"server",
"decentralized",
"centralized",
"bandits"
],
[
"control",
"dynamics",
"equations",
"physics",
"dynamical",
"particle",
"chemical",
"controller",
"derivatives",
"periodic"
],
[
"symbolic",
"rnns",
"scenes",
"paths",
"backbone",
"transformers",
"operate",
"express",
"segment",
"scratch"
],
[
"linear",
"function",
"convergence",
"functions",
"stochastic",
"optimal",
"optimization",
"gradient",
"convex",
"algorithm"
],
[
"weather",
"market",
"financial",
"day",
"daily",
"developers",
"software",
"boosting",
"year",
"forecast"
],
[
"classifier",
"detection",
"feature",
"classification",
"test",
"sensors",
"features",
"ensemble",
"class",
"vehicle"
],
[
"research",
"series",
"predictions",
"time",
"prediction",
"models",
"complex",
"modeling",
"ml",
"forecasting"
],
[
"intelligence",
"industrial",
"ai",
"interface",
"advanced",
"actively",
"forgetting",
"explainable",
"inform",
"discusses"
],
[
"reward",
"reinforcement",
"action",
"agent",
"actions",
"rl",
"agents",
"policies",
"policy",
"games"
],
[
"decision",
"framework",
"algorithms",
"learning",
"multi",
"machine",
"systems",
"knowledge",
"learn",
"problems"
],
[
"backpropagation",
"quantized",
"neuron",
"relu",
"pruned",
"degradation",
"normalization",
"activations",
"drop",
"architectural"
],
[
"bandwidth",
"besides",
"execution",
"wireless",
"implementations",
"gpu",
"transmission",
"latency",
"accelerators",
"demands"
],
[
"high",
"efficiency",
"computational",
"energy",
"cost",
"memory",
"low",
"efficient",
"scale",
"power"
],
[
"optical",
"3d",
"localization",
"reconstructed",
"mri",
"brain",
"segmentation",
"resonance",
"magnetic",
"association"
],
[
"outcome",
"pac",
"interventions",
"assumed",
"know",
"bayes",
"observational",
"concrete",
"interval",
"criterion"
],
[
"smoothness",
"regime",
"mathbb",
"varepsilon",
"lipschitz",
"lasso",
"exchange",
"inequality",
"differentially",
"convexity"
],
[
"privacy",
"adversarial",
"robustness",
"generative",
"attacks",
"against",
"robust",
"attack",
"gan",
"perturbations"
],
[
"of",
"in",
"the",
"to",
"and",
"is",
"we",
"for",
"that",
"this"
],
[
"domain",
"attention",
"representations",
"task",
"language",
"source",
"transfer",
"visual",
"pre",
"tasks"
],
[
"coefficients",
"geometry",
"coordinates",
"trick",
"orthogonal",
"penalty",
"optimizer",
"unconstrained",
"dictionary",
"symmetric"
]
] | 3,308.084736 | all-MiniLM-L6-v2 | 1 | -0.163399 | 0.153418 | 0.883532 |
ArXiv ML Papers | FASTopic | 45 | 30 | [
[
"bound",
"minimax",
"bounds",
"delta",
"varepsilon",
"mathbb",
"frac",
"tight",
"lipschitz",
"pac"
],
[
"they",
"use",
"model",
"models",
"modeling",
"inference",
"trained",
"predictions",
"prediction",
"datasets"
],
[
"item",
"users",
"user",
"recommendation",
"ranking",
"recommendations",
"recommender",
"aspect",
"personalized",
"twitter"
],
[
"server",
"decentralized",
"activation",
"neurons",
"relu",
"wireless",
"gnns",
"width",
"neuron",
"pruned"
],
[
"power",
"series",
"weather",
"time",
"consumption",
"forecasts",
"temporal",
"forecast",
"energy",
"channel"
],
[
"descent",
"stochastic",
"convergence",
"proximal",
"gradient",
"equations",
"sgd",
"accelerated",
"solvers",
"hessian"
],
[
"topics",
"authors",
"summary",
"answers",
"names",
"developers",
"document",
"scientific",
"proposals",
"topic"
],
[
"convolutional",
"cnn",
"brain",
"imaging",
"convolution",
"cnns",
"pooling",
"normalization",
"nas",
"mri"
],
[
"to",
"and",
"this",
"of",
"are",
"in",
"from",
"the",
"based",
"that"
],
[
"attack",
"attacks",
"adversarial",
"robustness",
"against",
"privacy",
"gan",
"robust",
"perturbations",
"gans"
],
[
"human",
"transfer",
"domain",
"research",
"knowledge",
"target",
"ai",
"reasoning",
"intelligence",
"driving"
],
[
"problems",
"non",
"dimensional",
"linear",
"optimization",
"functions",
"function",
"noise",
"points",
"regression"
],
[
"bias",
"fairness",
"fair",
"explanations",
"counterfactual",
"practices",
"biases",
"gender",
"explanation",
"explaining"
],
[
"speech",
"text",
"language",
"word",
"audio",
"words",
"languages",
"bert",
"speaker",
"labeling"
],
[
"clinical",
"patient",
"detected",
"diagnosis",
"inspection",
"anomalies",
"intrusion",
"technologies",
"spread",
"cancer"
],
[
"tensor",
"entries",
"rank",
"projections",
"subspace",
"matrix",
"matrices",
"dictionary",
"orthogonal",
"principal"
],
[
"updated",
"acquire",
"forgetting",
"chemical",
"navigation",
"look",
"possibilities",
"robotics",
"experiences",
"reaching"
],
[
"transportation",
"urban",
"localization",
"remote",
"transient",
"spatiotemporal",
"association",
"analytics",
"spatio",
"acquired"
],
[
"pruning",
"memory",
"hardware",
"library",
"scaling",
"bit",
"gpu",
"accelerators",
"bandwidth",
"throughput"
],
[
"classified",
"recurrent",
"sound",
"possibility",
"acoustic",
"rnns",
"28",
"aid",
"express",
"internal"
],
[
"latent",
"generative",
"autoencoders",
"ignore",
"discriminator",
"vae",
"reconstructing",
"encoders",
"reconstructed",
"supervision"
],
[
"structure",
"representation",
"clustering",
"graphs",
"graph",
"nodes",
"embedding",
"molecular",
"vertex",
"node"
],
[
"partition",
"integer",
"exchange",
"relaxation",
"vertices",
"cycles",
"laplacian",
"localized",
"graphical",
"recovered"
],
[
"policy",
"agent",
"reward",
"rl",
"reinforcement",
"action",
"control",
"agents",
"actions",
"policies"
],
[
"uncertainty",
"variational",
"bayesian",
"distributions",
"probabilistic",
"estimation",
"posterior",
"inference",
"carlo",
"monte"
],
[
"neural",
"images",
"image",
"art",
"object",
"visual",
"deep",
"training",
"attention",
"self"
],
[
"xgboost",
"screening",
"days",
"healthy",
"air",
"eeg",
"temperature",
"material",
"year",
"uncertainties"
],
[
"algorithm",
"algorithms",
"online",
"optimal",
"regret",
"distributed",
"bandit",
"cost",
"search",
"communication"
],
[
"test",
"svm",
"labels",
"classifier",
"label",
"classification",
"class",
"feature",
"features",
"machine"
],
[
"offs",
"interpolation",
"losses",
"adjust",
"alternatives",
"ad",
"ensures",
"understood",
"hyperparameters",
"choices"
]
] | 3,325.879388 | all-MiniLM-L6-v2 | 0.996667 | -0.158244 | 0.143302 | 0.87984 |
ArXiv ML Papers | FASTopic | 46 | 30 | [
[
"communication",
"federated",
"private",
"regret",
"bandit",
"bandits",
"centralized",
"tilde",
"server",
"decentralized"
],
[
"languages",
"nlp",
"speech",
"bert",
"audio",
"language",
"word",
"text",
"words",
"speaker"
],
[
"recommender",
"users",
"recommendation",
"recurrent",
"user",
"attention",
"item",
"video",
"videos",
"multimodal"
],
[
"fairness",
"bias",
"fair",
"preferences",
"outcomes",
"treatment",
"biases",
"know",
"criteria",
"imbalance"
],
[
"memory",
"architectures",
"layer",
"pruning",
"layers",
"architecture",
"convolution",
"hardware",
"distillation",
"scaling"
],
[
"200",
"updating",
"speedups",
"stored",
"pytorch",
"exhaustive",
"pruned",
"asynchronous",
"storing",
"speedup"
],
[
"95",
"developers",
"94",
"xgboost",
"forest",
"boosting",
"97",
"days",
"trend",
"infrastructure"
],
[
"dimensional",
"sparse",
"kernel",
"functions",
"function",
"optimization",
"regression",
"parameters",
"linear",
"nonlinear"
],
[
"transforms",
"coefficients",
"28",
"distortion",
"normalization",
"reconstructing",
"denoising",
"blind",
"reconstructed",
"fitting"
],
[
"interactive",
"safe",
"robotic",
"autonomous",
"robot",
"forgetting",
"navigation",
"robots",
"robotics",
"execute"
],
[
"transient",
"chemical",
"flow",
"particle",
"optical",
"remote",
"physics",
"spectrum",
"inside",
"phases"
],
[
"sgd",
"convex",
"stochastic",
"convergence",
"descent",
"accelerated",
"mathbb",
"gradient",
"proximal",
"lipschitz"
],
[
"agent",
"reinforcement",
"rl",
"agents",
"policy",
"control",
"rewards",
"trajectory",
"reward",
"imitation"
],
[
"networks",
"neural",
"deep",
"network",
"performance",
"model",
"training",
"on",
"with",
"however"
],
[
"transfer",
"target",
"domain",
"source",
"generation",
"transferring",
"scratch",
"alignment",
"shot",
"tune"
],
[
"robustness",
"against",
"adversarial",
"attacks",
"attack",
"perturbations",
"gans",
"security",
"defense",
"attacker"
],
[
"temporal",
"traffic",
"series",
"monitoring",
"service",
"sensors",
"iot",
"vehicle",
"sensor",
"weather"
],
[
"subspaces",
"spectral",
"dictionary",
"tensor",
"subspace",
"clusters",
"rank",
"clustering",
"matrix",
"principal"
],
[
"causal",
"counterfactual",
"players",
"practices",
"principles",
"targeting",
"behaviour",
"projects",
"consequences",
"credit"
],
[
"graphs",
"graph",
"representations",
"representation",
"structure",
"nodes",
"embedding",
"embeddings",
"protein",
"node"
],
[
"inference",
"variational",
"uncertainty",
"bayesian",
"distributions",
"probabilistic",
"divergence",
"likelihood",
"posterior",
"monte"
],
[
"brain",
"imaging",
"segmentation",
"image",
"images",
"medical",
"disease",
"cancer",
"cell",
"covid"
],
[
"molecular",
"gnn",
"topological",
"quantum",
"passing",
"directed",
"paths",
"inner",
"generalizing",
"trick"
],
[
"laplacian",
"characterization",
"beta",
"square",
"gps",
"cycles",
"analytically",
"omega",
"theorem",
"infinite"
],
[
"feature",
"selection",
"test",
"best",
"classification",
"was",
"developed",
"techniques",
"features",
"detection"
],
[
"supervised",
"labels",
"label",
"classification",
"class",
"datasets",
"unlabeled",
"classifier",
"labeled",
"loss"
],
[
"distinguish",
"91",
"scenes",
"scene",
"architectural",
"pixels",
"visual",
"captured",
"invariance",
"color"
],
[
"wearable",
"media",
"diseases",
"health",
"names",
"countries",
"databases",
"reports",
"intrusion",
"authors"
],
[
"how",
"machine",
"use",
"ml",
"these",
"systems",
"human",
"research",
"models",
"their"
],
[
"bounds",
"sample",
"complexity",
"algorithms",
"algorithm",
"theoretical",
"number",
"case",
"problem",
"optimal"
]
] | 3,312.861683 | all-MiniLM-L6-v2 | 0.996667 | -0.151122 | 0.151985 | 0.887027 |
ArXiv ML Papers | FASTopic | 43 | 40 | [
[
"bayesian",
"divergence",
"distributions",
"variational",
"posterior",
"mutual",
"covariates",
"mcmc",
"nonparametric",
"inducing"
],
[
"softmax",
"transferring",
"backbone",
"architectural",
"composed",
"reconstructing",
"decoding",
"perceptual",
"scratch",
"purely"
],
[
"outlier",
"bias",
"analyses",
"imbalance",
"imbalanced",
"dropout",
"statistically",
"accuracies",
"interpolation",
"proposing"
],
[
"integer",
"width",
"solvers",
"hilbert",
"quantum",
"unconstrained",
"communications",
"hessian",
"satisfying",
"newton"
],
[
"machine",
"models",
"prediction",
"ml",
"predictive",
"their",
"predictions",
"research",
"how",
"dataset"
],
[
"algorithms",
"optimal",
"algorithm",
"sample",
"bounds",
"under",
"problem",
"case",
"where",
"theoretical"
],
[
"systems",
"forecasting",
"modeling",
"time",
"temporal",
"causal",
"series",
"process",
"inference",
"complex"
],
[
"cloud",
"service",
"air",
"compliance",
"things",
"smart",
"services",
"market",
"price",
"trend"
],
[
"autoencoders",
"gan",
"latent",
"source",
"target",
"generative",
"generation",
"fidelity",
"domain",
"synthetic"
],
[
"graph",
"nodes",
"representation",
"graphs",
"embedding",
"pruning",
"node",
"gnns",
"structure",
"scaling"
],
[
"explanation",
"explanations",
"interpretability",
"explainable",
"explaining",
"symbolic",
"programs",
"grounded",
"effort",
"abstract"
],
[
"gnn",
"fitting",
"sign",
"convolutions",
"tensors",
"pruned",
"facilitates",
"activations",
"neuron",
"denoising"
],
[
"user",
"item",
"media",
"items",
"recommender",
"sentiment",
"users",
"recommendation",
"preferences",
"names"
],
[
"scene",
"visual",
"images",
"video",
"segmentation",
"image",
"object",
"gans",
"videos",
"pixel"
],
[
"convolutional",
"neural",
"networks",
"deep",
"architectures",
"layer",
"network",
"architecture",
"layers",
"nas"
],
[
"regret",
"stochastic",
"gradient",
"convex",
"convergence",
"descent",
"optimization",
"bandit",
"bandits",
"proximal"
],
[
"the",
"and",
"to",
"for",
"of",
"in",
"is",
"we",
"with",
"as"
],
[
"day",
"were",
"signals",
"eeg",
"was",
"frequency",
"test",
"signal",
"measurements",
"mri"
],
[
"turns",
"decades",
"thresholding",
"separation",
"coefficient",
"assisted",
"phases",
"plus",
"penalty",
"blind"
],
[
"overview",
"forgetting",
"incremental",
"proposals",
"updating",
"list",
"stream",
"searching",
"streams",
"learner"
],
[
"reward",
"rl",
"agent",
"agents",
"game",
"policy",
"reinforcement",
"policies",
"rewards",
"imitation"
],
[
"curve",
"diagnostic",
"94",
"surface",
"95",
"83",
"96",
"85",
"28",
"ct"
],
[
"distance",
"metric",
"nearest",
"neighbor",
"clustering",
"selection",
"cluster",
"distances",
"search",
"clusters"
],
[
"estimation",
"interval",
"probabilistic",
"uncertainty",
"calibration",
"imputation",
"belief",
"estimated",
"uncertain",
"former"
],
[
"hardware",
"implementations",
"library",
"gpu",
"accelerators",
"1d",
"engine",
"configuration",
"gpus",
"stacked"
],
[
"audio",
"language",
"speech",
"words",
"word",
"style",
"text",
"rnn",
"bert",
"languages"
],
[
"labels",
"label",
"unlabeled",
"ensemble",
"supervised",
"class",
"feature",
"detection",
"classifier",
"classification"
],
[
"modes",
"cell",
"capturing",
"cognitive",
"biology",
"biological",
"protein",
"entities",
"molecular",
"cells"
],
[
"pac",
"fair",
"frac",
"private",
"delta",
"fairness",
"differentially",
"arms",
"possibly",
"varepsilon"
],
[
"topics",
"projections",
"documents",
"nonnegative",
"separated",
"document",
"manifold",
"orthogonal",
"discrimination",
"topological"
],
[
"interventions",
"outcome",
"observational",
"counterfactual",
"decision",
"biases",
"trees",
"outcomes",
"treatment",
"conditioning"
],
[
"matrix",
"tensor",
"rank",
"sparse",
"regression",
"kernel",
"norm",
"mathbb",
"subspace",
"dimension"
],
[
"tasks",
"task",
"learning",
"multi",
"framework",
"propose",
"state",
"reasoning",
"learn",
"knowledge"
],
[
"disease",
"software",
"health",
"cancer",
"clinical",
"patient",
"reports",
"healthcare",
"patients",
"practices"
],
[
"city",
"events",
"transportation",
"weather",
"trends",
"credit",
"year",
"periods",
"urban",
"projects"
],
[
"navigation",
"purposes",
"vehicles",
"behavioral",
"driving",
"human",
"wearable",
"tracking",
"force",
"car"
],
[
"memory",
"privacy",
"distributed",
"server",
"cost",
"efficiency",
"energy",
"devices",
"student",
"teacher"
],
[
"equations",
"dynamical",
"chemical",
"inverse",
"informed",
"physics",
"controller",
"particle",
"analytically",
"periodic"
],
[
"attack",
"adversarial",
"attacks",
"robustness",
"perturbations",
"attacker",
"defense",
"perturbation",
"against",
"malware"
],
[
"neighborhood",
"vertex",
"message",
"vertices",
"laplacian",
"np",
"symmetric",
"links",
"directed",
"walk"
]
] | 3,499.81487 | all-MiniLM-L6-v2 | 1 | -0.182938 | 0.148416 | 0.894687 |
ArXiv ML Papers | FASTopic | 44 | 40 | [
[
"words",
"bert",
"word",
"languages",
"text",
"language",
"style",
"nlp",
"answering",
"sentences"
],
[
"logic",
"scientific",
"molecular",
"priori",
"temperature",
"chemical",
"solver",
"integration",
"biology",
"motivation"
],
[
"bayesian",
"distributions",
"gaussian",
"estimation",
"uncertainty",
"variational",
"inference",
"distribution",
"likelihood",
"probabilistic"
],
[
"private",
"fairness",
"federated",
"privacy",
"sensitive",
"fair",
"server",
"concerns",
"centralized",
"differentially"
],
[
"dynamics",
"equations",
"physics",
"particle",
"analytically",
"informed",
"inverse",
"dynamical",
"derivatives",
"optimisation"
],
[
"magnetic",
"mri",
"localization",
"spectrum",
"resonance",
"reconstructed",
"optical",
"thorough",
"spatially",
"reconstructing"
],
[
"adversarial",
"perturbations",
"malware",
"robustness",
"attacks",
"attack",
"security",
"defense",
"against",
"perturbation"
],
[
"label",
"speech",
"labels",
"speaker",
"unlabeled",
"audio",
"labeling",
"acoustic",
"supervised",
"imbalanced"
],
[
"targets",
"200",
"meta",
"tune",
"coverage",
"learnt",
"distillation",
"incremental",
"forgetting",
"boost"
],
[
"rank",
"matrix",
"sparse",
"tensor",
"subspace",
"recovery",
"covariance",
"entries",
"principal",
"nonnegative"
],
[
"convolutional",
"image",
"segmentation",
"pixel",
"images",
"pooling",
"gans",
"imaging",
"gan",
"resolution"
],
[
"clinical",
"health",
"patients",
"disease",
"patient",
"people",
"covid",
"age",
"healthcare",
"diseases"
],
[
"optimizer",
"compression",
"batch",
"parallel",
"compressed",
"python",
"mini",
"accelerate",
"proposals",
"surrogate"
],
[
"tilde",
"epsilon",
"sqrt",
"sgd",
"varepsilon",
"mathcal",
"max",
"minimax",
"delta",
"frac"
],
[
"user",
"their",
"research",
"models",
"human",
"systems",
"machine",
"how",
"ml",
"they"
],
[
"allocation",
"delay",
"quantum",
"controller",
"solvers",
"reaching",
"communications",
"distributed",
"transitions",
"communication"
],
[
"width",
"vanishing",
"decay",
"combinatorial",
"bounds",
"cycles",
"greedy",
"optimum",
"approximates",
"approximating"
],
[
"transfer",
"domain",
"target",
"task",
"tasks",
"source",
"self",
"objects",
"representations",
"object"
],
[
"imputation",
"bias",
"determining",
"outlier",
"uncertainties",
"calibration",
"measuring",
"calibrated",
"causes",
"know"
],
[
"manipulation",
"interactive",
"robotic",
"explanation",
"reasoning",
"robots",
"grounded",
"easier",
"reason",
"boxes"
],
[
"autoencoders",
"impressive",
"vae",
"latent",
"autoencoder",
"patches",
"autoregressive",
"dropout",
"normalization",
"wasserstein"
],
[
"urban",
"activity",
"frame",
"frames",
"cognitive",
"event",
"transportation",
"city",
"scenes",
"association"
],
[
"training",
"neural",
"network",
"networks",
"trained",
"loss",
"train",
"input",
"deep",
"student"
],
[
"clustering",
"representation",
"feature",
"features",
"communities",
"structure",
"clusters",
"relational",
"protein",
"group"
],
[
"in",
"of",
"to",
"this",
"and",
"the",
"that",
"on",
"is",
"as"
],
[
"reward",
"rl",
"agent",
"agents",
"imitation",
"policy",
"reinforcement",
"rewards",
"bandits",
"game"
],
[
"explainable",
"software",
"industry",
"material",
"developers",
"intrusion",
"authors",
"workflow",
"project",
"practices"
],
[
"gradient",
"optimization",
"algorithm",
"linear",
"stochastic",
"problems",
"convergence",
"function",
"optimal",
"non"
],
[
"dirichlet",
"kernel",
"multivariate",
"curse",
"gps",
"overlapping",
"nearest",
"neighbor",
"kernels",
"regression"
],
[
"graphs",
"nodes",
"graph",
"node",
"link",
"gnn",
"laplacian",
"gnns",
"neighborhood",
"topology"
],
[
"recommendation",
"items",
"ranking",
"taxonomy",
"item",
"summarize",
"preference",
"recommender",
"preferences",
"collaborative"
],
[
"hardware",
"architectures",
"nas",
"pruned",
"pruning",
"gpu",
"activations",
"neurons",
"backbone",
"accelerators"
],
[
"systems",
"time",
"forecasts",
"forecast",
"series",
"prediction",
"power",
"demand",
"service",
"forecasting"
],
[
"behaviour",
"principles",
"add",
"experiences",
"compliance",
"span",
"look",
"consequences",
"behavioral",
"players"
],
[
"classification",
"classifier",
"was",
"accuracy",
"detection",
"were",
"svm",
"recognition",
"test",
"ensemble"
],
[
"wireless",
"weather",
"mobile",
"sensors",
"transmission",
"smart",
"things",
"remote",
"air",
"trend"
],
[
"theoretic",
"arms",
"pac",
"interventions",
"ucb",
"bounds",
"expectation",
"observational",
"armed",
"revisit"
],
[
"sequence",
"attention",
"recurrent",
"rnn",
"bidirectional",
"decoding",
"rnns",
"express",
"character",
"decoder"
],
[
"anomalies",
"anomaly",
"xgboost",
"ct",
"curve",
"detector",
"diagnostic",
"96",
"inspection",
"classified"
],
[
"metric",
"sample",
"samples",
"given",
"case",
"distance",
"number",
"when",
"error",
"set"
]
] | 3,558.686721 | all-MiniLM-L6-v2 | 0.995 | -0.176121 | 0.149609 | 0.891071 |
ArXiv ML Papers | FASTopic | 45 | 40 | [
[
"architecture",
"convolutional",
"network",
"neural",
"networks",
"layer",
"input",
"layers",
"deep",
"pooling"
],
[
"theoretic",
"asymptotic",
"characterization",
"frac",
"oracle",
"decay",
"theorem",
"estimator",
"omega",
"imply"
],
[
"visual",
"3d",
"images",
"segmentation",
"image",
"object",
"video",
"shape",
"imaging",
"scene"
],
[
"outlier",
"modify",
"anomalies",
"facial",
"svhn",
"imbalanced",
"invariance",
"imbalance",
"labeling",
"categorical"
],
[
"the",
"of",
"in",
"and",
"to",
"as",
"we",
"is",
"that",
"for"
],
[
"attacks",
"privacy",
"robustness",
"attack",
"adversarial",
"against",
"perturbations",
"security",
"traffic",
"driving"
],
[
"projections",
"dictionary",
"principal",
"orthogonal",
"sketch",
"algebra",
"tensors",
"nonnegative",
"fourier",
"spectrum"
],
[
"domain",
"transfer",
"source",
"target",
"language",
"knowledge",
"adaptation",
"pre",
"transformer",
"shot"
],
[
"bound",
"algorithm",
"bounds",
"algorithms",
"case",
"sample",
"error",
"under",
"number",
"optimal"
],
[
"matrix",
"sparse",
"dimensional",
"kernel",
"clustering",
"rank",
"tensor",
"manifold",
"dimensionality",
"kernels"
],
[
"paths",
"impressive",
"facilitates",
"backbone",
"architectural",
"color",
"depth",
"convolutions",
"vanilla",
"expand"
],
[
"interpolation",
"spatially",
"provable",
"thresholding",
"overlapping",
"ell_2",
"penalty",
"satisfy",
"regime",
"optimizer"
],
[
"supervised",
"data",
"datasets",
"training",
"classification",
"label",
"labels",
"class",
"samples",
"large"
],
[
"wearable",
"signals",
"sound",
"frequencies",
"separation",
"signal",
"purposes",
"sensing",
"frequency",
"recognize"
],
[
"grounded",
"programs",
"compositional",
"scratch",
"rarely",
"serves",
"answer",
"reason",
"guided",
"transferring"
],
[
"compliance",
"services",
"service",
"platforms",
"individuals",
"intended",
"behavioral",
"things",
"consequences",
"accessible"
],
[
"software",
"engineering",
"reports",
"healthcare",
"project",
"material",
"care",
"records",
"databases",
"developers"
],
[
"quantum",
"regret",
"bandit",
"exploitation",
"horizon",
"exchange",
"bandits",
"armed",
"transitions",
"arms"
],
[
"fair",
"learners",
"losses",
"adversarially",
"discrimination",
"learnt",
"differentially",
"private",
"decentralized",
"perturbed"
],
[
"recognition",
"were",
"was",
"detection",
"feature",
"features",
"sensors",
"svm",
"eeg",
"forest"
],
[
"prediction",
"clinical",
"series",
"protein",
"patients",
"forecast",
"disease",
"patient",
"treatment",
"forecasting"
],
[
"generative",
"latent",
"gan",
"generation",
"gans",
"decoder",
"encoder",
"vae",
"autoencoders",
"discriminator"
],
[
"convex",
"stochastic",
"descent",
"convergence",
"gradient",
"optimization",
"sgd",
"proximal",
"sqrt",
"varepsilon"
],
[
"exhaustive",
"proposals",
"catastrophic",
"forgetting",
"incremental",
"search",
"adapting",
"updated",
"engine",
"streams"
],
[
"learn",
"space",
"problems",
"multi",
"constraints",
"state",
"learned",
"problem",
"process",
"framework"
],
[
"chemical",
"neurons",
"forward",
"backpropagation",
"physics",
"landscape",
"reconstructing",
"equations",
"inverse",
"informed"
],
[
"policy",
"agent",
"rl",
"reinforcement",
"agents",
"reward",
"action",
"actions",
"robot",
"policies"
],
[
"corpus",
"nlp",
"audio",
"text",
"sentence",
"sentences",
"speech",
"speaker",
"languages",
"english"
],
[
"19",
"diagnosis",
"curve",
"temperature",
"ct",
"covid",
"mri",
"cell",
"resonance",
"diagnostic"
],
[
"calibration",
"uncertainty",
"outcome",
"calibrated",
"interventions",
"observational",
"severely",
"counterfactual",
"insufficient",
"know"
],
[
"communication",
"systems",
"energy",
"cost",
"federated",
"distributed",
"server",
"control",
"demand",
"dynamics"
],
[
"recommendation",
"item",
"users",
"recommender",
"ranking",
"user",
"language",
"natural",
"topic",
"items"
],
[
"how",
"research",
"machine",
"ml",
"bias",
"human",
"trust",
"biases",
"fairness",
"their"
],
[
"sequence",
"lstm",
"recurrent",
"rnn",
"transportation",
"temporal",
"modes",
"urban",
"rnns",
"cognitive"
],
[
"selection",
"regression",
"nearest",
"neighbor",
"logistic",
"bayes",
"screening",
"predictors",
"partitioning",
"price"
],
[
"pruning",
"hardware",
"quantization",
"nas",
"gpu",
"activations",
"bandwidth",
"resnet",
"bit",
"pruned"
],
[
"bayesian",
"posterior",
"monte",
"distributions",
"carlo",
"density",
"variational",
"likelihood",
"mcmc",
"priors"
],
[
"explainable",
"explanations",
"entities",
"expression",
"passing",
"message",
"explanation",
"relational",
"explaining",
"taxonomy"
],
[
"hyperparameter",
"speedups",
"par",
"pytorch",
"200",
"python",
"asynchronous",
"implementing",
"implementations",
"indicating"
],
[
"graph",
"graphs",
"nodes",
"cluster",
"node",
"link",
"gnn",
"clusters",
"gnns",
"neighborhood"
]
] | 3,540.229822 | all-MiniLM-L6-v2 | 0.9975 | -0.174638 | 0.145059 | 0.88748 |
ArXiv ML Papers | FASTopic | 46 | 40 | [
[
"programming",
"grid",
"solves",
"heuristic",
"forgetting",
"turns",
"cycles",
"incremental",
"runs",
"optimizer"
],
[
"normalized",
"truncated",
"analytically",
"analyse",
"decompose",
"graphical",
"multivariate",
"motivation",
"insufficient",
"described"
],
[
"distributed",
"communication",
"decentralized",
"private",
"centralized",
"fairness",
"server",
"federated",
"privacy",
"aggregated"
],
[
"equations",
"diffusion",
"flows",
"flow",
"inverse",
"physics",
"dynamical",
"operator",
"chemical",
"normalizing"
],
[
"year",
"industrial",
"developers",
"day",
"engineering",
"intrusion",
"material",
"monitoring",
"materials",
"trends"
],
[
"sqrt",
"varepsilon",
"sgd",
"tilde",
"proximal",
"frac",
"lasso",
"newton",
"provable",
"nonconvex"
],
[
"gradient",
"optimization",
"algorithm",
"optimal",
"algorithms",
"stochastic",
"convergence",
"problems",
"problem",
"complexity"
],
[
"graphs",
"nodes",
"graph",
"node",
"link",
"gnn",
"relational",
"gnns",
"vertex",
"teacher"
],
[
"dimensional",
"rank",
"sparse",
"dictionary",
"low",
"tensor",
"clustering",
"matrix",
"completion",
"subspace"
],
[
"cnn",
"normalization",
"enhancement",
"pruned",
"nas",
"convolutional",
"pooling",
"pruning",
"cnns",
"layers"
],
[
"reasoning",
"answering",
"counterfactual",
"programs",
"abstraction",
"contexts",
"answers",
"grounded",
"symbolic",
"answer"
],
[
"inference",
"distribution",
"estimation",
"bayesian",
"variational",
"uncertainty",
"variables",
"sampling",
"parameters",
"distributions"
],
[
"modeling",
"time",
"prediction",
"series",
"systems",
"forecasts",
"predictions",
"weather",
"forecast",
"demand"
],
[
"gaussian",
"linear",
"error",
"regression",
"kernel",
"functions",
"generalization",
"nonlinear",
"kernels",
"relu"
],
[
"social",
"media",
"services",
"activities",
"city",
"twitter",
"spread",
"credit",
"daily",
"granularity"
],
[
"explanations",
"driving",
"explanation",
"transportation",
"inform",
"urban",
"cognitive",
"explainable",
"explaining",
"vehicle"
],
[
"quantum",
"energy",
"hardware",
"device",
"consumption",
"iot",
"execution",
"latency",
"bit",
"bandwidth"
],
[
"bias",
"interventions",
"prevalence",
"assessing",
"measuring",
"claim",
"biases",
"gender",
"identity",
"outcomes"
],
[
"neuron",
"width",
"activation",
"implicitly",
"nets",
"backpropagation",
"neurons",
"landscape",
"cifar10",
"dropout"
],
[
"as",
"and",
"in",
"to",
"this",
"the",
"on",
"of",
"by",
"are"
],
[
"accuracy",
"ml",
"system",
"classification",
"classifier",
"machine",
"techniques",
"outlier",
"detection",
"traffic"
],
[
"disease",
"patients",
"health",
"protein",
"clinical",
"cancer",
"patient",
"assessment",
"19",
"cell"
],
[
"algebra",
"nearest",
"principal",
"neighbor",
"defines",
"distances",
"similarities",
"mining",
"expression",
"curse"
],
[
"word",
"language",
"text",
"words",
"languages",
"corpus",
"nlp",
"sentences",
"sentiment",
"sentence"
],
[
"tree",
"boosting",
"losses",
"trees",
"imputation",
"bayes",
"know",
"streams",
"uncertainties",
"calibrated"
],
[
"2015",
"subsets",
"consequence",
"assumed",
"selection",
"attractive",
"interpolation",
"constructing",
"pseudo",
"screening"
],
[
"supervised",
"speech",
"label",
"source",
"domain",
"labels",
"transfer",
"pre",
"representations",
"task"
],
[
"tensors",
"recommender",
"item",
"items",
"collaborative",
"user",
"personalized",
"recommendation",
"products",
"preferences"
],
[
"end",
"deep",
"neural",
"parameters",
"networks",
"training",
"network",
"input",
"layer",
"architecture"
],
[
"object",
"mri",
"images",
"generative",
"image",
"shape",
"segmentation",
"gans",
"3d",
"gan"
],
[
"agents",
"policy",
"games",
"agent",
"reward",
"policies",
"rl",
"reinforcement",
"bandit",
"game"
],
[
"adversarial",
"attack",
"attacks",
"against",
"robustness",
"defense",
"vulnerability",
"attacker",
"perturbation",
"vulnerable"
],
[
"par",
"imbalance",
"asr",
"conversion",
"acoustic",
"28",
"voice",
"imbalanced",
"accuracies",
"sound"
],
[
"metric",
"information",
"they",
"their",
"features",
"find",
"use",
"how",
"identify",
"research"
],
[
"modify",
"accelerators",
"anomalies",
"pipelines",
"svhn",
"release",
"library",
"code",
"creating",
"detector"
],
[
"motion",
"human",
"robot",
"videos",
"robotic",
"interactive",
"music",
"manipulation",
"participants",
"navigation"
],
[
"attention",
"sequence",
"vae",
"rnn",
"decoder",
"autoencoders",
"recurrent",
"rnns",
"encoder",
"lstm"
],
[
"trick",
"subsequently",
"reconstructing",
"conclusion",
"color",
"invariance",
"pixels",
"composed",
"spatially",
"constrain"
],
[
"signals",
"transform",
"denoising",
"separable",
"topological",
"localization",
"spectrum",
"separation",
"fourier",
"distortion"
],
[
"asymptotic",
"estimators",
"mcmc",
"em",
"nonparametric",
"covariates",
"uniform",
"lipschitz",
"relaxation",
"beta"
]
] | 3,523.672514 | all-MiniLM-L6-v2 | 0.9975 | -0.181406 | 0.153885 | 0.889343 |
ArXiv ML Papers | FASTopic | 43 | 50 | [
[
"of",
"to",
"the",
"in",
"and",
"are",
"that",
"this",
"as",
"by"
],
[
"svhn",
"ensembles",
"biases",
"assessing",
"bias",
"equal",
"invariance",
"dropout",
"statistically",
"verification"
],
[
"summary",
"compliance",
"credit",
"genetic",
"counterfactual",
"protein",
"questions",
"observational",
"answers",
"causal"
],
[
"transfer",
"domain",
"object",
"target",
"representations",
"attention",
"pre",
"feature",
"source",
"supervision"
],
[
"gradient",
"convergence",
"proximal",
"descent",
"smoothness",
"accelerated",
"sgd",
"infty",
"hessian",
"differentially"
],
[
"text",
"language",
"word",
"nlp",
"sentences",
"words",
"textual",
"bert",
"languages",
"sentiment"
],
[
"energy",
"iot",
"memory",
"cloud",
"power",
"cost",
"hardware",
"design",
"efficiency",
"computational"
],
[
"hyper",
"amenable",
"implicitly",
"optimizer",
"encourages",
"unconstrained",
"adapts",
"stable",
"prominent",
"hyperparameter"
],
[
"varepsilon",
"armed",
"sqrt",
"bandit",
"exchange",
"bandits",
"mathcal",
"regret",
"ucb",
"tilde"
],
[
"reports",
"media",
"software",
"developers",
"industry",
"website",
"project",
"web",
"workflow",
"projects"
],
[
"decoder",
"audio",
"acoustic",
"channel",
"encoder",
"vae",
"speaker",
"decoding",
"speech",
"character"
],
[
"forgetting",
"critic",
"imitation",
"atari",
"control",
"reaching",
"controller",
"robotics",
"catastrophic",
"safe"
],
[
"reinforcement",
"agents",
"agent",
"policy",
"rl",
"reward",
"rewards",
"game",
"policies",
"games"
],
[
"estimation",
"regression",
"uncertainty",
"noise",
"estimate",
"prediction",
"selection",
"distribution",
"robust",
"error"
],
[
"random",
"bounds",
"generalization",
"sample",
"bound",
"theoretical",
"complexity",
"entropy",
"class",
"case"
],
[
"width",
"double",
"rigorously",
"square",
"corresponds",
"decay",
"quantum",
"hilbert",
"integer",
"exponentially"
],
[
"monte",
"distributions",
"posterior",
"bayesian",
"variational",
"carlo",
"nonparametric",
"mcmc",
"priors",
"mutual"
],
[
"generative",
"latent",
"inference",
"how",
"modeling",
"complex",
"discrete",
"dynamics",
"models",
"systems"
],
[
"allocation",
"price",
"streams",
"keeping",
"updating",
"correction",
"parallel",
"budget",
"exhaustive",
"proposals"
],
[
"flows",
"informed",
"optimisation",
"differentiation",
"equations",
"normalizing",
"particle",
"physics",
"inverse",
"derivatives"
],
[
"care",
"xgboost",
"material",
"market",
"wearable",
"countries",
"healthy",
"city",
"days",
"heart"
],
[
"objects",
"reasoning",
"robot",
"symbolic",
"interactive",
"cognitive",
"visual",
"execute",
"observing",
"pose"
],
[
"trends",
"facilitating",
"spatiotemporal",
"spatio",
"optical",
"strengths",
"weather",
"temperature",
"96",
"surface"
],
[
"discrimination",
"proposing",
"weak",
"curse",
"interpolation",
"brings",
"accuracies",
"margin",
"meanwhile",
"tail"
],
[
"node",
"relational",
"gnn",
"gnns",
"graph",
"graphs",
"vertex",
"link",
"nodes",
"teacher"
],
[
"look",
"cooperative",
"constrain",
"solves",
"compatible",
"intuition",
"deviations",
"offs",
"firstly",
"secondly"
],
[
"dictionary",
"subspace",
"subspaces",
"sparse",
"recovery",
"projections",
"principal",
"tensor",
"thresholding",
"nonnegative"
],
[
"sound",
"today",
"suggested",
"indicating",
"preprocessing",
"28",
"mainly",
"taxonomy",
"investigation",
"characteristic"
],
[
"problems",
"problem",
"algorithm",
"optimization",
"algorithms",
"function",
"order",
"non",
"convex",
"optimal"
],
[
"temporal",
"series",
"event",
"events",
"detection",
"forecasting",
"time",
"monitoring",
"anomaly",
"outlier"
],
[
"neighboring",
"euclidean",
"topological",
"geometric",
"molecular",
"walk",
"geometry",
"passing",
"links",
"directed"
],
[
"images",
"segmentation",
"image",
"medical",
"imaging",
"clinical",
"patients",
"cancer",
"disease",
"cell"
],
[
"eeg",
"plus",
"signals",
"frequency",
"frequencies",
"orthogonal",
"denoising",
"spectrum",
"signal",
"spatially"
],
[
"clustering",
"distance",
"metric",
"matrix",
"nearest",
"manifold",
"neighbor",
"rank",
"kernel",
"mining"
],
[
"quantized",
"convolution",
"frame",
"backbone",
"nas",
"gpus",
"pruned",
"pruning",
"convolutions",
"resnet"
],
[
"items",
"users",
"trust",
"user",
"item",
"recommender",
"recommendation",
"social",
"service",
"recommendations"
],
[
"used",
"ml",
"machine",
"research",
"features",
"study",
"analysis",
"techniques",
"been",
"system"
],
[
"summarize",
"granularity",
"pixels",
"fed",
"categorical",
"track",
"views",
"guidance",
"patch",
"annotation"
],
[
"privacy",
"distributed",
"communication",
"fair",
"fairness",
"server",
"sensitive",
"decentralized",
"private",
"federated"
],
[
"acquisition",
"degrade",
"imputation",
"processes",
"consequences",
"uncertain",
"uncertainties",
"interval",
"failure",
"know"
],
[
"preferences",
"partitioning",
"communities",
"overlapping",
"alternatives",
"cycles",
"tree",
"pairwise",
"partition",
"trees"
],
[
"gender",
"camera",
"human",
"age",
"head",
"eye",
"explainable",
"facial",
"explanations",
"person"
],
[
"scratch",
"neurons",
"expand",
"inductive",
"gradually",
"sequence",
"compositional",
"express",
"modules",
"feed"
],
[
"attacks",
"adversarial",
"attack",
"gan",
"perturbations",
"robustness",
"gans",
"attacker",
"defense",
"malware"
],
[
"art",
"dataset",
"models",
"state",
"training",
"large",
"datasets",
"propose",
"trained",
"performance"
],
[
"network",
"deep",
"networks",
"layer",
"convolutional",
"layers",
"neural",
"architecture",
"activation",
"pooling"
],
[
"smart",
"99",
"95",
"svm",
"97",
"intrusion",
"remote",
"internet",
"engine",
"throughput"
],
[
"conclusion",
"broader",
"ultimately",
"coefficient",
"describes",
"defines",
"subsequently",
"membership",
"transparent",
"proving"
],
[
"car",
"sensors",
"sensor",
"wireless",
"driving",
"vehicles",
"tracking",
"physical",
"vehicle",
"traffic"
],
[
"classifiers",
"imbalance",
"class",
"classes",
"label",
"classification",
"labeling",
"classifier",
"labels",
"supervised"
]
] | 3,440.081508 | all-MiniLM-L6-v2 | 0.996 | -0.193447 | 0.143969 | 0.894682 |
ArXiv ML Papers | FASTopic | 44 | 50 | [
[
"of",
"the",
"and",
"in",
"to",
"that",
"is",
"for",
"we",
"this"
],
[
"matrix",
"tensor",
"clustering",
"sparse",
"rank",
"entries",
"factorization",
"completion",
"subspace",
"dictionary"
],
[
"bounds",
"sample",
"bound",
"error",
"quantum",
"estimator",
"generalization",
"complexity",
"lower",
"under"
],
[
"periods",
"signals",
"signal",
"material",
"day",
"spectrum",
"channels",
"remote",
"separation",
"frequency"
],
[
"prediction",
"were",
"19",
"protein",
"was",
"covid",
"patients",
"test",
"study",
"used"
],
[
"bandit",
"sqrt",
"regret",
"tilde",
"mathcal",
"armed",
"arms",
"exchange",
"varepsilon",
"bandits"
],
[
"order",
"algorithms",
"algorithm",
"convergence",
"optimal",
"optimization",
"problems",
"problem",
"efficient",
"complexity"
],
[
"margin",
"par",
"accuracies",
"split",
"branch",
"themselves",
"tail",
"extracts",
"appealing",
"proposals"
],
[
"hundreds",
"reference",
"mining",
"list",
"overview",
"want",
"documents",
"taxonomy",
"authors",
"detail"
],
[
"descent",
"gradient",
"stochastic",
"momentum",
"proximal",
"accelerated",
"convergence",
"sgd",
"quadratic",
"hessian"
],
[
"uncertainty",
"forecasting",
"forecast",
"equations",
"dynamical",
"calibrated",
"forecasts",
"uncertainties",
"weather",
"calibration"
],
[
"classifier",
"class",
"supervised",
"label",
"feature",
"labels",
"classification",
"features",
"classifiers",
"classes"
],
[
"constrain",
"enabled",
"logic",
"robots",
"robotic",
"robotics",
"catastrophic",
"communicate",
"forgetting",
"softmax"
],
[
"textual",
"item",
"answering",
"reasoning",
"programs",
"linguistic",
"answers",
"reading",
"names",
"texts"
],
[
"gans",
"source",
"target",
"gan",
"transfer",
"generation",
"generative",
"domain",
"speaker",
"discriminator"
],
[
"detection",
"anomalies",
"99",
"anomaly",
"detected",
"detecting",
"intrusion",
"detect",
"correctly",
"inspection"
],
[
"boosting",
"criteria",
"predictors",
"failure",
"learners",
"imputation",
"examining",
"imbalanced",
"missing",
"bayes"
],
[
"paths",
"graphical",
"visualization",
"representing",
"molecular",
"inferring",
"discovering",
"biology",
"message",
"passing"
],
[
"explanations",
"gender",
"explanation",
"interpretability",
"explainable",
"explaining",
"internal",
"interpretable",
"distinguish",
"counterfactual"
],
[
"players",
"outcome",
"preferences",
"treatment",
"observational",
"outcomes",
"alternatives",
"competition",
"compliance",
"interventions"
],
[
"nearest",
"neighbor",
"kernel",
"covariates",
"overlapping",
"regression",
"discrimination",
"consequence",
"partitioning",
"distances"
],
[
"images",
"video",
"camera",
"pixel",
"segmentation",
"visual",
"object",
"image",
"scene",
"3d"
],
[
"recommendation",
"server",
"collaborative",
"knowledge",
"multi",
"federated",
"user",
"meta",
"recommender",
"each"
],
[
"penalty",
"decreases",
"landscape",
"interpolation",
"ell_2",
"thorough",
"note",
"spatially",
"regularization",
"lipschitz"
],
[
"exhaustive",
"implementations",
"targeted",
"engine",
"centralized",
"intensive",
"adjust",
"configuration",
"cloud",
"execution"
],
[
"concerns",
"trust",
"heterogeneity",
"svhn",
"remove",
"differentially",
"compromising",
"sacrificing",
"preserving",
"locally"
],
[
"control",
"rl",
"reward",
"reinforcement",
"policy",
"agents",
"agent",
"imitation",
"robot",
"actions"
],
[
"link",
"vertex",
"gnn",
"graphs",
"gnns",
"graph",
"node",
"nodes",
"embedding",
"structure"
],
[
"invariance",
"motion",
"perception",
"perceptual",
"videos",
"color",
"scenes",
"frame",
"frames",
"moving"
],
[
"dimensional",
"metric",
"space",
"distance",
"noise",
"linear",
"function",
"vector",
"sampling",
"structure"
],
[
"bit",
"quantization",
"pruning",
"bandwidth",
"activations",
"pruned",
"accelerators",
"quantized",
"compress",
"pytorch"
],
[
"memory",
"attention",
"architecture",
"convolutional",
"nas",
"convolution",
"architectures",
"recognition",
"lstm",
"layers"
],
[
"individual",
"human",
"fair",
"fairness",
"bias",
"how",
"they",
"decision",
"biases",
"their"
],
[
"auc",
"cell",
"detector",
"diagnosis",
"curve",
"diagnostic",
"facial",
"classified",
"85",
"93"
],
[
"distributions",
"priors",
"wasserstein",
"likelihood",
"bayesian",
"mcmc",
"posterior",
"variational",
"inference",
"normalizing"
],
[
"directed",
"laplacian",
"cycles",
"vertices",
"omega",
"walk",
"communities",
"localized",
"inequality",
"count"
],
[
"controller",
"game",
"safe",
"cooperative",
"atari",
"critic",
"games",
"offline",
"solver",
"inefficient"
],
[
"inverse",
"notions",
"nonlinear",
"width",
"vanishing",
"approximates",
"integer",
"operator",
"approximated",
"exponentially"
],
[
"decoder",
"vae",
"autoencoders",
"rnn",
"encoder",
"acoustic",
"speech",
"translation",
"autoregressive",
"asr"
],
[
"sensors",
"iot",
"traffic",
"energy",
"power",
"mobile",
"design",
"sensor",
"consumption",
"devices"
],
[
"shifts",
"broader",
"prevalence",
"affecting",
"intended",
"intuitive",
"impressive",
"looking",
"conclusion",
"opportunities"
],
[
"modeling",
"series",
"dynamics",
"temporal",
"process",
"causal",
"field",
"complex",
"time",
"processes"
],
[
"attack",
"attacks",
"adversarial",
"privacy",
"perturbations",
"robustness",
"against",
"security",
"defense",
"perturbation"
],
[
"health",
"software",
"reports",
"eeg",
"activity",
"healthcare",
"subjects",
"project",
"care",
"medicine"
],
[
"media",
"platforms",
"transportation",
"topics",
"services",
"service",
"twitter",
"evolving",
"analytics",
"spread"
],
[
"denoising",
"patch",
"fusion",
"patches",
"resonance",
"mri",
"optical",
"enhancement",
"architectural",
"imaging"
],
[
"physics",
"neurons",
"chemical",
"informed",
"particle",
"backpropagation",
"biological",
"neuron",
"forward",
"diffusion"
],
[
"equally",
"trick",
"draw",
"analytically",
"controlling",
"priori",
"theorem",
"principled",
"imply",
"rigorous"
],
[
"bert",
"languages",
"words",
"language",
"text",
"representations",
"word",
"natural",
"semantic",
"task"
],
[
"neural",
"networks",
"deep",
"network",
"loss",
"training",
"trained",
"input",
"dnn",
"layer"
]
] | 3,374.699858 | all-MiniLM-L6-v2 | 0.994 | -0.196233 | 0.151798 | 0.898911 |
ArXiv ML Papers | FASTopic | 45 | 50 | [
[
"smoothness",
"lipschitz",
"mcmc",
"nonparametric",
"bernoulli",
"densities",
"expectation",
"em",
"divergence",
"carlo"
],
[
"specify",
"degrade",
"acquisition",
"tree",
"add",
"appear",
"trees",
"sequential",
"opposed",
"broadly"
],
[
"generative",
"gans",
"gan",
"image",
"segmentation",
"images",
"3d",
"shape",
"object",
"reconstruction"
],
[
"bidirectional",
"taxonomy",
"labeling",
"answers",
"summarize",
"grounded",
"annotated",
"guidance",
"answering",
"annotations"
],
[
"knowledge",
"graph",
"representation",
"representations",
"graphs",
"structure",
"embedding",
"nodes",
"node",
"embeddings"
],
[
"detection",
"accuracy",
"features",
"feature",
"classification",
"recognition",
"proposed",
"outlier",
"compared",
"technique"
],
[
"mutual",
"constructing",
"deviation",
"square",
"interval",
"decreases",
"recovered",
"occur",
"dependence",
"uncertainties"
],
[
"teacher",
"learnt",
"bias",
"student",
"imbalance",
"overfitting",
"distillation",
"imbalanced",
"svhn",
"gender"
],
[
"discriminator",
"supervision",
"generation",
"transformations",
"generator",
"augment",
"pose",
"patches",
"patch",
"variations"
],
[
"estimator",
"bound",
"sample",
"regression",
"under",
"generalization",
"error",
"bounds",
"gaussian",
"probability"
],
[
"distance",
"nearest",
"subspace",
"kernel",
"kernels",
"manifold",
"distances",
"subspaces",
"neighbor",
"metric"
],
[
"adversarial",
"robustness",
"attacks",
"attack",
"against",
"defense",
"perturbations",
"security",
"perturbation",
"attacker"
],
[
"cluster",
"clustering",
"gnns",
"vertex",
"euclidean",
"links",
"edges",
"spectrum",
"walk",
"topological"
],
[
"software",
"creating",
"materials",
"project",
"reports",
"developers",
"growth",
"workflow",
"industry",
"actively"
],
[
"accelerators",
"gpu",
"hardware",
"throughput",
"30",
"implementations",
"python",
"parallel",
"latency",
"execution"
],
[
"lstm",
"demand",
"traffic",
"service",
"vehicle",
"modes",
"smart",
"vehicles",
"transportation",
"urban"
],
[
"architecture",
"network",
"neural",
"input",
"layer",
"deep",
"networks",
"loss",
"recurrent",
"dnn"
],
[
"accelerated",
"momentum",
"quadratic",
"optimum",
"solvers",
"derivatives",
"optimisation",
"hessian",
"newton",
"optimizer"
],
[
"explanation",
"ensembles",
"explain",
"interpretability",
"explaining",
"proposing",
"counterfactual",
"explanations",
"evidence",
"grained"
],
[
"models",
"their",
"data",
"model",
"these",
"from",
"and",
"such",
"have",
"different"
],
[
"factorization",
"tensor",
"dictionary",
"tensors",
"decomposition",
"algebra",
"principal",
"matrices",
"nonnegative",
"completion"
],
[
"problem",
"rank",
"sparse",
"algorithms",
"algorithm",
"matrix",
"number",
"problems",
"efficient",
"each"
],
[
"mri",
"localization",
"whole",
"resonance",
"brain",
"imaging",
"ct",
"reconstructed",
"diagnostic",
"diagnosis"
],
[
"video",
"visual",
"videos",
"camera",
"head",
"scenes",
"frame",
"facial",
"frames",
"pixels"
],
[
"formally",
"fair",
"pac",
"notion",
"losses",
"learner",
"choosing",
"probabilities",
"entropy",
"distributional"
],
[
"predictions",
"prediction",
"forecasting",
"disease",
"validation",
"forecast",
"patient",
"patients",
"clinical",
"covid"
],
[
"convergence",
"gradient",
"optimization",
"quantum",
"private",
"stochastic",
"convex",
"descent",
"sgd",
"minimization"
],
[
"preference",
"preferences",
"alternatives",
"overlapping",
"recommender",
"communities",
"pairwise",
"ad",
"ranking",
"impossible"
],
[
"spread",
"xgboost",
"healthcare",
"care",
"market",
"wearable",
"healthy",
"countries",
"days",
"records"
],
[
"languages",
"text",
"language",
"words",
"word",
"bert",
"nlp",
"corpus",
"sentences",
"sentence"
],
[
"series",
"systems",
"physical",
"physics",
"motion",
"time",
"sensor",
"complex",
"dynamics",
"system"
],
[
"policy",
"reinforcement",
"agent",
"rl",
"agents",
"reward",
"games",
"robot",
"game",
"policies"
],
[
"sqrt",
"tilde",
"regret",
"bandits",
"frac",
"epsilon",
"bandit",
"mathbb",
"lasso",
"armed"
],
[
"convolutional",
"pruning",
"activation",
"relu",
"pooling",
"cnns",
"convolutions",
"activations",
"convolution",
"nas"
],
[
"perception",
"biological",
"possibilities",
"dynamical",
"identifies",
"span",
"inside",
"overlap",
"inform",
"itself"
],
[
"thorough",
"normalizing",
"diffusion",
"flows",
"inverse",
"equations",
"reconstruct",
"informed",
"forward",
"boundary"
],
[
"self",
"transfer",
"domain",
"unlabeled",
"supervised",
"adaptation",
"pre",
"task",
"target",
"source"
],
[
"atari",
"trajectory",
"safe",
"imitation",
"forgetting",
"trajectories",
"critic",
"robots",
"controller",
"offline"
],
[
"techniques",
"fairness",
"machine",
"ml",
"classification",
"research",
"some",
"classifier",
"classifiers",
"label"
],
[
"explainable",
"user",
"ai",
"compliance",
"personal",
"players",
"human",
"participants",
"behavioral",
"humans"
],
[
"observational",
"causal",
"topics",
"imputation",
"twitter",
"treatment",
"topic",
"scientific",
"credit",
"documents"
],
[
"scale",
"efficiency",
"efficient",
"cost",
"memory",
"accuracy",
"devices",
"computational",
"design",
"up"
],
[
"server",
"streaming",
"privacy",
"communication",
"distributed",
"federated",
"decentralized",
"centralized",
"differentially",
"speedups"
],
[
"audio",
"vae",
"decoding",
"speaker",
"coding",
"speech",
"channel",
"acoustic",
"asr",
"distortion"
],
[
"compromising",
"targeted",
"streams",
"adjust",
"exploited",
"popularity",
"trains",
"exposure",
"predictor",
"hyperparameters"
],
[
"eeg",
"material",
"svm",
"intrusion",
"95",
"classify",
"detecting",
"classifying",
"detected",
"99"
],
[
"allocation",
"wireless",
"backpropagation",
"facilitates",
"device",
"compress",
"employing",
"fitting",
"coordinates",
"capacity"
],
[
"be",
"is",
"we",
"show",
"of",
"the",
"an",
"that",
"which",
"methods"
],
[
"inference",
"latent",
"variational",
"bayesian",
"estimation",
"distribution",
"generative",
"uncertainty",
"distributions",
"likelihood"
],
[
"chemical",
"compositional",
"passing",
"molecular",
"recursive",
"reasoning",
"symbolic",
"abstraction",
"entities",
"heuristic"
]
] | 3,337.448148 | all-MiniLM-L6-v2 | 0.992 | -0.176289 | 0.146356 | 0.88898 |
ArXiv ML Papers | FASTopic | 46 | 50 | [
[
"explanation",
"human",
"explanations",
"questions",
"answer",
"gender",
"explaining",
"explainable",
"answering",
"answers"
],
[
"recursive",
"identifies",
"inferring",
"deviations",
"static",
"generalizing",
"possibilities",
"feed",
"phenomena",
"dependency"
],
[
"equation",
"hessian",
"vanishing",
"differentially",
"differential",
"solvers",
"momentum",
"newton",
"nonconvex",
"satisfying"
],
[
"lasso",
"relaxation",
"subspaces",
"dictionary",
"recovery",
"sparse",
"entries",
"thresholding",
"provable",
"ell_2"
],
[
"asynchronous",
"engine",
"speedups",
"speedup",
"acceleration",
"computations",
"stored",
"running",
"communications",
"retaining"
],
[
"event",
"long",
"missing",
"prediction",
"series",
"forecast",
"temporal",
"time",
"demand",
"forecasting"
],
[
"split",
"imbalance",
"meanwhile",
"hyper",
"imbalanced",
"proposals",
"appealing",
"coverage",
"addresses",
"stream"
],
[
"topological",
"orthogonal",
"coefficients",
"transform",
"euclidean",
"geometric",
"spatially",
"principal",
"manifold",
"nonnegative"
],
[
"asr",
"modalities",
"rnn",
"acoustic",
"audio",
"enhancement",
"speech",
"recurrent",
"speaker",
"rnns"
],
[
"regularization",
"generalization",
"interpolation",
"perturbed",
"dropout",
"overfitting",
"optimizer",
"mini",
"double",
"learnt"
],
[
"domain",
"tasks",
"art",
"transfer",
"training",
"target",
"task",
"state",
"self",
"source"
],
[
"molecular",
"package",
"chemical",
"described",
"material",
"period",
"materials",
"biology",
"scientific",
"science"
],
[
"gpu",
"quantized",
"quantization",
"pruning",
"backbone",
"pruned",
"resnet",
"accelerators",
"nas",
"gpus"
],
[
"recommendation",
"ranking",
"users",
"user",
"trust",
"student",
"item",
"collaborative",
"knowledge",
"recommender"
],
[
"creates",
"software",
"projects",
"surface",
"classified",
"strengths",
"developers",
"project",
"creating",
"engineering"
],
[
"proposed",
"used",
"most",
"system",
"algorithms",
"search",
"multi",
"compared",
"been",
"best"
],
[
"clinical",
"patients",
"medical",
"health",
"disease",
"patient",
"cancer",
"diagnosis",
"19",
"covid"
],
[
"image",
"object",
"images",
"segmentation",
"gan",
"video",
"3d",
"scene",
"reconstruction",
"driving"
],
[
"transportation",
"market",
"weather",
"day",
"urban",
"year",
"wearable",
"modes",
"daily",
"city"
],
[
"matrix",
"distance",
"kernel",
"regression",
"rank",
"tensor",
"quantum",
"neighbor",
"low",
"dimensional"
],
[
"impacts",
"compliance",
"intrusion",
"detect",
"detected",
"smart",
"things",
"affecting",
"inspection",
"security"
],
[
"regret",
"bandits",
"bandit",
"horizon",
"arms",
"armed",
"exploitation",
"ucb",
"arm",
"iterations"
],
[
"consensus",
"alternatives",
"price",
"passing",
"message",
"pairwise",
"impossible",
"percentage",
"directed",
"analyse"
],
[
"models",
"model",
"inference",
"modeling",
"generative",
"trained",
"autoencoders",
"samples",
"use",
"parameters"
],
[
"distributions",
"inverse",
"variational",
"carlo",
"flows",
"monte",
"mcmc",
"normalizing",
"inducing",
"gps"
],
[
"symbolic",
"sequences",
"sequence",
"programs",
"forgetting",
"formalize",
"compositional",
"vae",
"softmax",
"sequential"
],
[
"recognition",
"detection",
"ensemble",
"labels",
"class",
"classifier",
"label",
"classification",
"supervised",
"classes"
],
[
"bound",
"sample",
"complexity",
"sampling",
"distribution",
"case",
"bounds",
"error",
"random",
"algorithm"
],
[
"delta",
"em",
"epsilon",
"varepsilon",
"minimax",
"frac",
"tight",
"greedy",
"oracle",
"ascent"
],
[
"beta",
"mutual",
"bayes",
"bayesian",
"expectation",
"nonparametric",
"assumed",
"bernoulli",
"pac",
"kl"
],
[
"iot",
"consumption",
"memory",
"mobile",
"traffic",
"device",
"devices",
"hardware",
"execution",
"channel"
],
[
"representation",
"feature",
"representations",
"features",
"structure",
"embedding",
"information",
"protein",
"transformation",
"attention"
],
[
"magnetic",
"signals",
"eeg",
"separation",
"resonance",
"optical",
"spectrum",
"cognitive",
"remote",
"association"
],
[
"policy",
"reinforcement",
"agents",
"agent",
"rl",
"game",
"cooperative",
"actions",
"players",
"controller"
],
[
"adversarial",
"attack",
"attacks",
"robustness",
"defense",
"malware",
"malicious",
"against",
"attacker",
"perturbations"
],
[
"the",
"this",
"of",
"and",
"for",
"to",
"in",
"is",
"as",
"are"
],
[
"graphs",
"nodes",
"graph",
"node",
"clustering",
"gnns",
"clusters",
"cluster",
"gnn",
"edges"
],
[
"neural",
"networks",
"network",
"deep",
"convolutional",
"architecture",
"layer",
"architectures",
"activation",
"layers"
],
[
"interventions",
"fairness",
"fair",
"preferences",
"outcome",
"treatment",
"outcomes",
"observational",
"bias",
"causal"
],
[
"constraints",
"active",
"function",
"problems",
"control",
"systems",
"learn",
"design",
"space",
"optimization"
],
[
"robot",
"imitation",
"robots",
"car",
"perception",
"constrain",
"robotic",
"robotics",
"navigation",
"motion"
],
[
"languages",
"nlp",
"bert",
"text",
"corpora",
"english",
"word",
"sentence",
"sentences",
"language"
],
[
"coding",
"modules",
"convolutions",
"architectural",
"pytorch",
"1d",
"digital",
"transmission",
"pipelines",
"codes"
],
[
"patches",
"supervision",
"frame",
"pixels",
"patch",
"labeling",
"segment",
"augmentation",
"transformations",
"svhn"
],
[
"stochastic",
"distributed",
"optimization",
"convergence",
"convex",
"gradient",
"descent",
"server",
"sgd",
"federated"
],
[
"work",
"methods",
"our",
"show",
"not",
"its",
"propose",
"only",
"their",
"also"
],
[
"distances",
"remove",
"curse",
"distinguishing",
"discrimination",
"overlapping",
"storing",
"analyses",
"identity",
"families"
],
[
"mining",
"trees",
"boosting",
"attractive",
"describes",
"add",
"tree",
"partitioning",
"programming",
"learners"
],
[
"names",
"website",
"sentiment",
"documents",
"texts",
"document",
"media",
"twitter",
"topics",
"personal"
],
[
"uncertainty",
"physics",
"uncertainties",
"uncertain",
"analytically",
"particle",
"optimisation",
"informed",
"specify",
"calibration"
]
] | 3,392.000896 | all-MiniLM-L6-v2 | 0.998 | -0.17476 | 0.152955 | 0.888539 |
ArXiv ML Papers | S³ | 43 | 10 | [
[
"outperforming",
"aiming",
"screening",
"estimator",
"optimizer",
"calibration",
"accuracy",
"calibrated",
"performance",
"tuning"
],
[
"bandwidth",
"streaming",
"speedup",
"throughput",
"cloud",
"benchmarks",
"faster",
"streams",
"runtime",
"bottleneck"
],
[
"bandits",
"regularization",
"sparse",
"minimax",
"lasso",
"regularized",
"leveraging",
"minimize",
"optimal",
"optimality"
],
[
"forecasts",
"predicts",
"forecasting",
"softmax",
"prediction",
"backpropagation",
"predicting",
"forecast",
"rnn",
"autoregressive"
],
[
"lasso",
"discover",
"discovering",
"discovery",
"sensing",
"convex",
"classification",
"classifier",
"clustering",
"shape"
],
[
"selecting",
"iterative",
"unstructured",
"guide",
"iteratively",
"completion",
"rapid",
"retrieval",
"planning",
"steps"
],
[
"denoising",
"compression",
"clustering",
"diagnostic",
"prevalence",
"datasets",
"transformer",
"fitting",
"outliers",
"outlier"
],
[
"encountered",
"statistical",
"cluster",
"clustering",
"criteria",
"prevalence",
"analytics",
"statistics",
"methodology",
"popularity"
],
[
"annotations",
"annotation",
"predictive",
"automated",
"annotated",
"dataset",
"supervised",
"datasets",
"categories",
"attributes"
],
[
"generative",
"variational",
"simulations",
"priors",
"bayesian",
"carlo",
"likelihood",
"inference",
"modeling",
"monte"
]
] | 295.721126 | all-MiniLM-L6-v2 | 0.95 | -0.334627 | 0.212657 | 0.892302 |
ArXiv ML Papers | S³ | 44 | 10 | [
[
"generative",
"forecasting",
"simulations",
"forecasts",
"modelling",
"models",
"autoregressive",
"forecast",
"likelihood",
"variational"
],
[
"occur",
"planning",
"generation",
"automated",
"events",
"scenarios",
"interactive",
"observing",
"participants",
"behaviors"
],
[
"classify",
"classification",
"classifier",
"classifying",
"learns",
"predicting",
"classifiers",
"predictive",
"prediction",
"predicts"
],
[
"textual",
"speech",
"corpus",
"linguistic",
"texts",
"languages",
"corpora",
"speaker",
"language",
"text"
],
[
"control",
"behavioral",
"strategy",
"rewards",
"strategies",
"exploration",
"planning",
"reinforcement",
"reward",
"robotics"
],
[
"statistical",
"estimating",
"clustering",
"algorithm",
"ranking",
"oracle",
"cluster",
"averaging",
"averaged",
"prevalence"
],
[
"bias",
"supervised",
"labeled",
"retrieval",
"leveraging",
"labeling",
"bandits",
"broadly",
"discriminative",
"regularized"
],
[
"security",
"adversarially",
"adversary",
"attacks",
"adversarial",
"adversaries",
"vulnerability",
"attacker",
"exploits",
"privacy"
],
[
"cloud",
"speedup",
"streaming",
"throughput",
"bandwidth",
"faster",
"benchmarks",
"runtime",
"streams",
"scalable"
],
[
"graphs",
"networks",
"nodes",
"graph",
"vertex",
"node",
"vertices",
"communities",
"relational",
"neighbors"
]
] | 251.923385 | all-MiniLM-L6-v2 | 0.99 | -0.259835 | 0.2023 | 0.889624 |
ArXiv ML Papers | S³ | 45 | 10 | [
[
"strategy",
"reinforcement",
"rnn",
"learnt",
"learns",
"future",
"neuron",
"atari",
"backpropagation",
"ai"
],
[
"classifier",
"predictors",
"classify",
"classifiers",
"supervised",
"classifying",
"classification",
"predicting",
"predictive",
"prediction"
],
[
"workflow",
"streaming",
"iot",
"cloud",
"speedup",
"xgboost",
"streams",
"cluster",
"database",
"platform"
],
[
"clustering",
"methodology",
"estimating",
"analytics",
"cluster",
"algorithm",
"criteria",
"selecting",
"prevalence",
"ranking"
],
[
"robotics",
"recognition",
"classify",
"dynamics",
"objects",
"motion",
"robotic",
"perception",
"controller",
"imaging"
],
[
"annotation",
"distinguishing",
"labeled",
"reward",
"reinforcement",
"annotations",
"labeling",
"demonstrate",
"discriminate",
"rewards"
],
[
"symmetric",
"detecting",
"thresholding",
"recognition",
"detection",
"discriminator",
"security",
"detector",
"algorithm",
"norm"
],
[
"cloud",
"3d",
"simulations",
"traffic",
"covid",
"believe",
"pixel",
"infrastructure",
"intrusion",
"observing"
],
[
"attacks",
"interpretability",
"security",
"vulnerability",
"attacker",
"vulnerable",
"adversaries",
"intrusion",
"noise",
"adversarially"
],
[
"nodes",
"graphs",
"graph",
"networks",
"vertex",
"node",
"vertices",
"communities",
"neighbors",
"relational"
]
] | 316.460816 | all-MiniLM-L6-v2 | 0.92 | -0.331667 | 0.173737 | 0.869225 |
ArXiv ML Papers | S³ | 46 | 10 | [
[
"clustering",
"statistical",
"criteria",
"ranking",
"prevalence",
"cluster",
"encountered",
"statistics",
"analytics",
"popularity"
],
[
"sensing",
"discovering",
"discover",
"3d",
"classifier",
"classification",
"discovery",
"shape",
"calibration",
"clustering"
],
[
"retrieval",
"selecting",
"iteratively",
"unstructured",
"steps",
"iterative",
"planning",
"completion",
"guide",
"rapid"
],
[
"cloud",
"streaming",
"faster",
"benchmarks",
"throughput",
"speedup",
"bandwidth",
"streams",
"bottleneck",
"runtime"
],
[
"learns",
"strategy",
"reinforcement",
"reward",
"rewards",
"critic",
"exploration",
"strategies",
"behavioral",
"control"
],
[
"generative",
"simulations",
"likelihood",
"modelling",
"bayesian",
"variational",
"modeling",
"priors",
"models",
"forecasts"
],
[
"datasets",
"labeling",
"guided",
"labeled",
"annotated",
"annotation",
"annotations",
"predictive",
"supervised",
"leveraging"
],
[
"behaviors",
"modelling",
"temporal",
"dynamics",
"simulation",
"simulations",
"workflow",
"monitoring",
"development",
"methodology"
],
[
"vision",
"capturing",
"3d",
"capture",
"observing",
"scenes",
"interactive",
"spatiotemporal",
"camera",
"pixels"
],
[
"graphs",
"graph",
"node",
"nodes",
"vertices",
"networks",
"vertex",
"communities",
"relational",
"neighbors"
]
] | 336.242681 | all-MiniLM-L6-v2 | 0.96 | -0.315028 | 0.184422 | 0.889205 |
ArXiv ML Papers | S³ | 43 | 20 | [
[
"statistically",
"asymptotically",
"outliers",
"outlier",
"asymptotic",
"deviations",
"classifiers",
"fitting",
"corpora",
"overfitting"
],
[
"learns",
"ensembles",
"resonance",
"optimizes",
"distillation",
"tuning",
"generates",
"generating",
"ensemble",
"generator"
],
[
"exploits",
"attacks",
"attacker",
"intrusion",
"adversarial",
"adversary",
"exploit",
"vulnerability",
"security",
"adversarially"
],
[
"classifier",
"clustering",
"clusters",
"cluster",
"svm",
"randomized",
"segmentation",
"algorithm",
"classification",
"mixture"
],
[
"accelerators",
"speedup",
"benchmarks",
"throughput",
"builds",
"faster",
"optimized",
"build",
"runtime",
"implementations"
],
[
"conditioning",
"channels",
"seek",
"probability",
"greedy",
"expectation",
"bayes",
"bandits",
"hypothesis",
"bandit"
],
[
"classifiers",
"classifier",
"recognition",
"minimizing",
"memory",
"perceptual",
"optimizer",
"decoder",
"vision",
"denoising"
],
[
"collected",
"attention",
"learns",
"learned",
"forgetting",
"recall",
"collect",
"annotated",
"leveraging",
"retaining"
],
[
"records",
"metrics",
"lstm",
"epsilon",
"length",
"logistic",
"log",
"healthcare",
"privacy",
"measure"
],
[
"densities",
"bayes",
"stochastic",
"likelihood",
"probabilistic",
"bayesian",
"variational",
"distributions",
"sampling",
"markov"
],
[
"automated",
"encountered",
"clustering",
"methodology",
"preprocessing",
"implementations",
"oracle",
"modification",
"correspond",
"interactive"
],
[
"gradients",
"hessian",
"minimization",
"convexity",
"convex",
"gradient",
"minimax",
"optimization",
"convergence",
"smoothness"
],
[
"highest",
"hyperparameter",
"hyperparameters",
"bayes",
"minimum",
"nearest",
"bayesian",
"outperformed",
"maximize",
"likelihood"
],
[
"potential",
"length",
"attempts",
"coordinates",
"parameters",
"comparisons",
"parametric",
"magnitude",
"coordinate",
"distances"
],
[
"linguistic",
"texts",
"textual",
"nlp",
"corpora",
"corpus",
"semantic",
"sentences",
"text",
"language"
],
[
"labeled",
"discriminator",
"quantization",
"distinguishing",
"labeling",
"recognizing",
"recognition",
"supervised",
"structured",
"embeddings"
],
[
"clinical",
"medical",
"diagnostic",
"diagnosis",
"disease",
"cancer",
"diseases",
"medicine",
"healthcare",
"mri"
],
[
"matrix",
"factorization",
"matrices",
"tensor",
"sparse",
"factors",
"factor",
"tensors",
"subspaces",
"subspace"
],
[
"appealing",
"fairness",
"preference",
"profile",
"profiles",
"gender",
"discrimination",
"personalized",
"bias",
"criteria"
],
[
"intuitive",
"discovering",
"generalizing",
"views",
"generalization",
"granularity",
"intuition",
"hierarchical",
"reasoning",
"phenomenon"
]
] | 275.911884 | all-MiniLM-L6-v2 | 0.935 | -0.320491 | 0.19318 | 0.912932 |
ArXiv ML Papers | S³ | 44 | 20 | [
[
"sensing",
"selected",
"selection",
"ensemble",
"chosen",
"dimensionality",
"bandwidth",
"choosing",
"frequencies",
"trained"
],
[
"camera",
"vision",
"traffic",
"capturing",
"tracking",
"gps",
"ranging",
"3d",
"vehicle",
"driving"
],
[
"coordinates",
"attempts",
"comparisons",
"parameters",
"parametric",
"potential",
"length",
"magnitude",
"distances",
"coordinate"
],
[
"formulated",
"3d",
"nonlinear",
"differentiable",
"modeling",
"modelling",
"differential",
"manifold",
"expressive",
"autoencoders"
],
[
"length",
"metrics",
"log",
"records",
"epsilon",
"logistic",
"lstm",
"healthcare",
"privacy",
"measure"
],
[
"methodology",
"preprocessing",
"encountered",
"automated",
"implementations",
"interactive",
"oracle",
"correspond",
"analyse",
"procedures"
],
[
"statistically",
"fitting",
"deviations",
"classifiers",
"asymptotically",
"outlier",
"overfitting",
"outliers",
"asymptotic",
"granularity"
],
[
"accuracy",
"classes",
"deterministic",
"classifiers",
"classifier",
"adaptation",
"evolving",
"trees",
"transitions",
"classification"
],
[
"exploits",
"adversarial",
"attacks",
"attacker",
"vulnerability",
"adversary",
"intrusion",
"exploit",
"security",
"adversarially"
],
[
"fairness",
"appealing",
"profiles",
"discrimination",
"preference",
"gender",
"profile",
"bias",
"personalized",
"criteria"
],
[
"autoregressive",
"forecast",
"weather",
"predicting",
"forecasts",
"forecasting",
"trend",
"regression",
"prediction",
"predicted"
],
[
"graph",
"vertices",
"vertex",
"graphs",
"nodes",
"edges",
"networks",
"communities",
"node",
"edge"
],
[
"dimensionality",
"approximating",
"approximations",
"embeddings",
"dimensional",
"invariance",
"approximates",
"embedding",
"abstraction",
"approximation"
],
[
"assumptions",
"capability",
"guarantees",
"projections",
"estimates",
"uncertainties",
"estimation",
"confidence",
"limitations",
"capabilities"
],
[
"gradient",
"convex",
"convexity",
"convergence",
"optimization",
"hessian",
"minimax",
"gradients",
"perturbation",
"minimization"
],
[
"symmetric",
"naive",
"superiority",
"linear",
"matrices",
"vectors",
"linearly",
"determine",
"matrix",
"cooperative"
],
[
"bayes",
"hyperparameter",
"nearest",
"highest",
"outperformed",
"bayesian",
"minimum",
"hyperparameters",
"algorithm",
"minimize"
],
[
"recall",
"annotated",
"learns",
"attention",
"forgetting",
"leveraging",
"collect",
"learned",
"retaining",
"collected"
],
[
"speedup",
"builds",
"throughput",
"benchmarks",
"implementations",
"accelerators",
"build",
"faster",
"runtime",
"optimized"
],
[
"frequencies",
"speaker",
"recorded",
"audio",
"acoustic",
"voice",
"sound",
"frequency",
"speech",
"spectrum"
]
] | 326.008333 | all-MiniLM-L6-v2 | 0.97 | -0.327587 | 0.19234 | 0.903546 |
ArXiv ML Papers | S³ | 45 | 20 | [
[
"sensing",
"selected",
"selection",
"chosen",
"ensemble",
"dimensionality",
"frequencies",
"bandwidth",
"settings",
"choosing"
],
[
"gps",
"tracking",
"vision",
"camera",
"capturing",
"driving",
"3d",
"ranging",
"traffic",
"spatiotemporal"
],
[
"outlier",
"outliers",
"asymptotically",
"asymptotic",
"fitting",
"statistically",
"deviations",
"corpora",
"classifiers",
"overfitting"
],
[
"estimates",
"insights",
"guarantees",
"capability",
"confidence",
"imputation",
"limitations",
"overhead",
"uncertainties",
"capabilities"
],
[
"recognizing",
"distinguishing",
"quantization",
"discriminator",
"labeling",
"recognition",
"labeled",
"supervised",
"structured",
"retrieval"
],
[
"seek",
"bandits",
"expectation",
"hypothesis",
"greedy",
"conditioning",
"bandit",
"bayes",
"probability",
"choosing"
],
[
"relations",
"relational",
"inference",
"inferring",
"bayes",
"bayesian",
"logic",
"knowledge",
"probabilities",
"selects"
],
[
"parametric",
"comparisons",
"attempts",
"potential",
"parameters",
"magnitude",
"coordinates",
"length",
"coordinate",
"selection"
],
[
"lstm",
"epsilon",
"records",
"metrics",
"log",
"healthcare",
"logistic",
"activity",
"length",
"minutes"
],
[
"neurons",
"gaussian",
"neural",
"backpropagation",
"neuron",
"uncertainty",
"networks",
"probabilistic",
"uncertainties",
"projections"
],
[
"factor",
"matrix",
"factorization",
"matrices",
"factors",
"sparse",
"tensor",
"tensors",
"subspaces",
"subspace"
],
[
"classifier",
"optimizer",
"perceptual",
"recognition",
"minimizing",
"classifiers",
"decoder",
"classification",
"denoising",
"vision"
],
[
"hierarchical",
"generalizing",
"intuitive",
"generalization",
"views",
"reasoning",
"intuition",
"discovering",
"systematically",
"phenomenon"
],
[
"behavioral",
"behaviors",
"dynamics",
"cognitive",
"behaviour",
"behavior",
"phenomena",
"phenomenon",
"brain",
"neuron"
],
[
"bayes",
"bayesian",
"likelihood",
"distributions",
"stochastic",
"probabilistic",
"densities",
"variational",
"markov",
"posterior"
],
[
"appealing",
"profiles",
"fairness",
"discrimination",
"preference",
"personalized",
"profile",
"bias",
"gender",
"biases"
],
[
"intrusion",
"security",
"attacks",
"attacker",
"adversarial",
"vulnerability",
"adversarially",
"exploit",
"adversary",
"exploits"
],
[
"dimensionality",
"approximating",
"dimensional",
"invariance",
"approximates",
"approximations",
"imitation",
"embeddings",
"learnable",
"deterministic"
],
[
"algorithm",
"computational",
"implementations",
"implementation",
"computation",
"procedures",
"implementing",
"computationally",
"computations",
"methodology"
],
[
"optimized",
"builds",
"architecture",
"throughput",
"faster",
"bandwidth",
"accelerators",
"architectural",
"benchmarks",
"hardware"
]
] | 328.684549 | all-MiniLM-L6-v2 | 0.925 | -0.347456 | 0.193481 | 0.907652 |
ArXiv ML Papers | S³ | 46 | 20 | [
[
"guarantees",
"estimates",
"uncertainties",
"capability",
"estimation",
"projections",
"assumptions",
"confidence",
"limitations",
"interpretability"
],
[
"profile",
"preference",
"profiles",
"fairness",
"gender",
"appealing",
"discrimination",
"personalized",
"bias",
"criteria"
],
[
"generates",
"distillation",
"tuning",
"resonance",
"learns",
"ensembles",
"optimizes",
"generating",
"ensemble",
"generator"
],
[
"matrices",
"sparse",
"factors",
"factor",
"matrix",
"factorization",
"tensor",
"subspace",
"subspaces",
"tensors"
],
[
"learned",
"collected",
"recall",
"attention",
"forgetting",
"annotated",
"retaining",
"leveraging",
"collect",
"learns"
],
[
"graphs",
"vertex",
"graph",
"vertices",
"edge",
"node",
"edges",
"networks",
"nodes",
"communities"
],
[
"cognitive",
"interpolation",
"labeling",
"marginal",
"planning",
"completion",
"consist",
"partitioning",
"intermediate",
"combines"
],
[
"distributed",
"shared",
"federated",
"privacy",
"private",
"decentralized",
"sharing",
"independently",
"cloud",
"cooperative"
],
[
"ensemble",
"selection",
"chosen",
"selected",
"sensing",
"dimensionality",
"bandwidth",
"frequencies",
"choosing",
"trained"
],
[
"convexity",
"gradient",
"hessian",
"convex",
"gradients",
"minimax",
"convergence",
"optimization",
"smoothness",
"perturbation"
],
[
"neural",
"neuron",
"neurons",
"cnn",
"imagenet",
"backpropagation",
"cnns",
"softmax",
"networks",
"network"
],
[
"implementations",
"benchmarks",
"throughput",
"speedup",
"accelerators",
"faster",
"builds",
"build",
"optimized",
"runtime"
],
[
"seek",
"bandits",
"greedy",
"expectation",
"bandit",
"probability",
"hypothesis",
"conditioning",
"bayes",
"choosing"
],
[
"multimodal",
"adaptation",
"generative",
"synthesis",
"generating",
"generator",
"generates",
"gan",
"gans",
"generate"
],
[
"behavioral",
"behaviors",
"phenomenon",
"cognitive",
"behavior",
"behaviour",
"brain",
"dynamics",
"phenomena",
"eeg"
],
[
"densities",
"distributions",
"bayes",
"stochastic",
"variational",
"likelihood",
"probabilistic",
"bayesian",
"markov",
"sampling"
],
[
"controlling",
"reinforcement",
"controller",
"interactive",
"exploration",
"behavioral",
"planning",
"control",
"controlled",
"interaction"
],
[
"healthcare",
"medicine",
"medical",
"cancer",
"diseases",
"clinical",
"diagnosis",
"disease",
"diagnostic",
"screening"
],
[
"quantization",
"recognizing",
"recognition",
"distinguishing",
"discriminator",
"labeling",
"structured",
"labeled",
"supervised",
"embeddings"
],
[
"voice",
"acoustic",
"audio",
"speaker",
"frequencies",
"recorded",
"sound",
"frequency",
"speech",
"spectrum"
]
] | 252.77266 | all-MiniLM-L6-v2 | 0.935 | -0.280087 | 0.208209 | 0.907094 |
ArXiv ML Papers | S³ | 43 | 30 | [
[
"dynamics",
"diffusion",
"landscape",
"researchers",
"perturbation",
"variance",
"perturbations",
"effect",
"effects",
"efficacy"
],
[
"records",
"temporal",
"autoregressive",
"sequence",
"periodic",
"spatiotemporal",
"sequences",
"recurrent",
"rnns",
"lstm"
],
[
"measure",
"estimators",
"estimation",
"estimate",
"measuring",
"measurements",
"measurement",
"estimator",
"measured",
"task"
],
[
"tend",
"convexity",
"convex",
"phenomenon",
"trend",
"intuition",
"viewed",
"intuitive",
"emerging",
"increasingly"
],
[
"appealing",
"biased",
"discriminate",
"bias",
"discrimination",
"fairness",
"biases",
"heterogeneity",
"equally",
"equal"
],
[
"movement",
"neuron",
"rnn",
"networks",
"evolving",
"transition",
"proportional",
"neurons",
"neural",
"chain"
],
[
"ensemble",
"pipelines",
"pac",
"forecasts",
"forecast",
"ensembles",
"deep",
"pipeline",
"allocation",
"aggregation"
],
[
"deviation",
"outlier",
"asymptotically",
"statistically",
"asymptotic",
"accuracy",
"deviations",
"fitting",
"outliers",
"overfitting"
],
[
"learns",
"supervised",
"samples",
"machine",
"sample",
"leads",
"learners",
"batch",
"dataset",
"datasets"
],
[
"classifier",
"recognition",
"images",
"image",
"segmentation",
"classifying",
"classification",
"art",
"convolutional",
"imagenet"
],
[
"bayes",
"likelihood",
"bayesian",
"densities",
"posterior",
"gaussian",
"probabilistic",
"distributions",
"distributional",
"priors"
],
[
"quantum",
"quantized",
"carlo",
"particle",
"randomly",
"accuracy",
"experiments",
"compressed",
"probability",
"probabilities"
],
[
"diseases",
"mri",
"patients",
"medical",
"clinical",
"disease",
"medicine",
"healthcare",
"diagnosis",
"diagnostic"
],
[
"effects",
"effect",
"attention",
"influence",
"screening",
"trend",
"themselves",
"observing",
"targeted",
"follow"
],
[
"uncertainties",
"robustness",
"explanations",
"accuracy",
"reliability",
"reliable",
"interpretability",
"interpretable",
"uncertainty",
"interpretation"
],
[
"labeled",
"classifier",
"supervised",
"classification",
"classifiers",
"classifying",
"labeling",
"labels",
"covariance",
"classify"
],
[
"weather",
"autoregressive",
"forecasting",
"forecast",
"traffic",
"forecasts",
"gps",
"prediction",
"predicting",
"transportation"
],
[
"tensors",
"role",
"multivariate",
"respectively",
"tensor",
"multimodal",
"transformations",
"rank",
"representations",
"vectors"
],
[
"private",
"decentralized",
"public",
"publicly",
"independently",
"federated",
"sharing",
"privacy",
"shared",
"accessible"
],
[
"predictors",
"hand",
"regret",
"https",
"predictor",
"information",
"learn",
"details",
"have",
"fidelity"
],
[
"centralized",
"collaborative",
"stream",
"combines",
"fusion",
"distributed",
"learners",
"learns",
"learner",
"federated"
],
[
"expectation",
"sampling",
"bandits",
"bandit",
"retrieval",
"seek",
"searching",
"randomly",
"greedy",
"sample"
],
[
"optical",
"interpolation",
"labeling",
"labels",
"sampled",
"collaborative",
"resolution",
"perception",
"marginal",
"consist"
],
[
"notions",
"phase",
"distributional",
"signals",
"phases",
"theory",
"corpus",
"classes",
"speech",
"distributions"
],
[
"generated",
"generates",
"generating",
"generate",
"generator",
"gans",
"generative",
"gan",
"autoencoder",
"generation"
],
[
"expressive",
"analyse",
"interactive",
"process",
"research",
"manipulation",
"visualization",
"preprocessing",
"art",
"developments"
],
[
"track",
"music",
"prediction",
"sensing",
"ensembles",
"ensemble",
"predicting",
"tracking",
"perception",
"audio"
],
[
"minimizing",
"field",
"combining",
"fusion",
"ensemble",
"ensembles",
"densities",
"mixture",
"allocation",
"optimality"
],
[
"minimize",
"convex",
"convexity",
"optimization",
"minimization",
"hessian",
"gradient",
"gradients",
"sgd",
"minimax"
],
[
"gps",
"highlights",
"ranging",
"tracking",
"coordinates",
"track",
"capturing",
"camera",
"poses",
"3d"
]
] | 298.363925 | all-MiniLM-L6-v2 | 0.873333 | -0.295973 | 0.189832 | 0.884518 |
ArXiv ML Papers | S³ | 44 | 30 | [
[
"protein",
"encoder",
"decoding",
"domains",
"encoded",
"molecular",
"encodes",
"encoding",
"encode",
"autoencoder"
],
[
"hand",
"https",
"fidelity",
"information",
"longer",
"dependent",
"via",
"predictor",
"cross",
"future"
],
[
"approximates",
"threshold",
"bounds",
"approximation",
"sequences",
"approximations",
"thresholding",
"approximating",
"approximated",
"optimality"
],
[
"minimizing",
"decoder",
"ensemble",
"ensembles",
"mixture",
"optimizer",
"combining",
"field",
"fusion",
"optimality"
],
[
"tracking",
"vision",
"observing",
"poses",
"ranging",
"perspective",
"pose",
"3d",
"camera",
"coordinates"
],
[
"tend",
"transferred",
"proportional",
"aggregating",
"target",
"centralized",
"towards",
"sequential",
"transfer",
"subsequent"
],
[
"independently",
"decentralized",
"public",
"privacy",
"publicly",
"private",
"federated",
"accessible",
"shared",
"sharing"
],
[
"sampling",
"dimensionality",
"quantum",
"frequencies",
"discovery",
"carlo",
"experiments",
"discovering",
"characteristic",
"selects"
],
[
"constraints",
"relations",
"logic",
"inference",
"conditional",
"reasoning",
"relational",
"probabilistic",
"constrained",
"constraint"
],
[
"disease",
"patients",
"clinical",
"diagnosis",
"mri",
"diseases",
"healthcare",
"medical",
"medicine",
"diagnostic"
],
[
"tuning",
"music",
"sensing",
"filters",
"ensemble",
"track",
"ensembles",
"prediction",
"tracking",
"sampled"
],
[
"outlier",
"convolutions",
"asymptotic",
"asymptotically",
"fitting",
"outliers",
"convolution",
"overfitting",
"throughput",
"bottleneck"
],
[
"images",
"recognition",
"svm",
"classification",
"classifier",
"algorithm",
"tensor",
"classifiers",
"classifying",
"classify"
],
[
"classifier",
"interpretability",
"accuracy",
"tree",
"forest",
"reliable",
"classifiers",
"outperforming",
"outperform",
"ensemble"
],
[
"gan",
"generating",
"generates",
"generated",
"generate",
"generative",
"generator",
"gans",
"generation",
"autoencoder"
],
[
"hessian",
"calibration",
"differentiation",
"nonlinear",
"gradient",
"differentiable",
"derivatives",
"backpropagation",
"perturbed",
"differential"
],
[
"area",
"traffic",
"planning",
"bandwidth",
"urban",
"gps",
"distance",
"iot",
"transportation",
"consumption"
],
[
"wearable",
"behavioral",
"behaviors",
"devices",
"signal",
"device",
"activity",
"eeg",
"signals",
"demonstrate"
],
[
"bayesian",
"likelihood",
"probabilistic",
"distributions",
"posterior",
"densities",
"bayes",
"distributional",
"stochastic",
"distribution"
],
[
"attacker",
"exploits",
"security",
"adversarial",
"vulnerability",
"exploit",
"attacks",
"intrusion",
"adversary",
"exploiting"
],
[
"labeled",
"supervised",
"labeling",
"labels",
"classifiers",
"label",
"classifier",
"classification",
"classifying",
"unlabeled"
],
[
"interactive",
"process",
"manipulation",
"analyse",
"visualization",
"research",
"developments",
"expressive",
"preprocessing",
"art"
],
[
"similarity",
"distances",
"metrics",
"metric",
"distance",
"corresponds",
"dimensionality",
"dimension",
"dimensional",
"correspond"
],
[
"records",
"rnns",
"occurrence",
"sequences",
"temporal",
"recurrent",
"record",
"events",
"spatiotemporal",
"periodic"
],
[
"convex",
"convexity",
"trend",
"viewed",
"tend",
"emerging",
"subsequently",
"phenomenon",
"increasingly",
"similar"
],
[
"reconstructing",
"navigation",
"autoencoder",
"forgetting",
"iteration",
"iteratively",
"iterations",
"iterative",
"discover",
"exploration"
],
[
"forecast",
"autoregressive",
"weather",
"predict",
"predicted",
"predicting",
"forecasts",
"forecasting",
"predicts",
"predictions"
],
[
"abstraction",
"cognitive",
"neuron",
"neural",
"brain",
"perception",
"interpretability",
"explanations",
"intuition",
"intuitive"
],
[
"supervised",
"generalizes",
"learns",
"algorithms",
"generalization",
"datasets",
"dataset",
"ai",
"algorithmic",
"generalize"
],
[
"phases",
"phase",
"classes",
"distributions",
"distributional",
"class",
"inducing",
"corpus",
"transitions",
"signals"
]
] | 356.388772 | all-MiniLM-L6-v2 | 0.923333 | -0.288788 | 0.205288 | 0.898208 |
ArXiv ML Papers | S³ | 45 | 30 | [
[
"labeling",
"experiments",
"posterior",
"labeled",
"unlabeled",
"structured",
"mechanisms",
"quantization",
"imitation",
"experimental"
],
[
"rnn",
"sequences",
"sequence",
"attention",
"sequential",
"lstm",
"recurrent",
"rnns",
"encode",
"private"
],
[
"retrieval",
"expectation",
"coverage",
"seek",
"searching",
"sampling",
"bandits",
"bandit",
"greedy",
"hypothesis"
],
[
"fairness",
"biases",
"bias",
"discriminate",
"discrimination",
"heterogeneity",
"biased",
"appealing",
"demand",
"representative"
],
[
"classifying",
"classify",
"classification",
"supervised",
"classifier",
"classifiers",
"labeled",
"labeling",
"backpropagation",
"regularization"
],
[
"recall",
"forest",
"boost",
"emerging",
"boosting",
"covid",
"armed",
"xgboost",
"bit",
"positive"
],
[
"across",
"combines",
"stream",
"transfer",
"transformation",
"adaptation",
"target",
"transferring",
"transferred",
"incorporating"
],
[
"traffic",
"distance",
"urban",
"autoencoder",
"city",
"bandwidth",
"bounded",
"planning",
"iot",
"area"
],
[
"tend",
"convexity",
"viewed",
"subsequently",
"convex",
"neurons",
"trend",
"emerging",
"similar",
"increasingly"
],
[
"generator",
"distillation",
"energy",
"generates",
"generating",
"optimizes",
"compressed",
"resonance",
"generative",
"imbalanced"
],
[
"events",
"record",
"periodic",
"occurrence",
"measure",
"temporal",
"recurrent",
"records",
"measures",
"outliers"
],
[
"ensembles",
"mixture",
"ensemble",
"decoder",
"fusion",
"minimizing",
"classifier",
"field",
"combining",
"allocation"
],
[
"dimensionality",
"similarity",
"distances",
"metrics",
"dimensional",
"metric",
"distance",
"embeddings",
"dimension",
"dimensions"
],
[
"shared",
"centralized",
"federated",
"distributed",
"privacy",
"private",
"public",
"independently",
"decentralized",
"sharing"
],
[
"learns",
"samples",
"supervised",
"learners",
"sample",
"dataset",
"datasets",
"machine",
"tasks",
"machines"
],
[
"generate",
"generating",
"generates",
"generated",
"generative",
"generator",
"gans",
"gan",
"generation",
"autoencoder"
],
[
"training",
"assessment",
"reliability",
"trained",
"augment",
"grounded",
"cnn",
"learned",
"gnn",
"learns"
],
[
"partitioning",
"clustering",
"segmentation",
"aggregated",
"clusters",
"aggregation",
"cluster",
"partition",
"ensemble",
"ensembles"
],
[
"invariant",
"covariance",
"signals",
"symmetric",
"variational",
"notions",
"invariance",
"quantization",
"arbitrary",
"abstraction"
],
[
"classifier",
"evolving",
"ensemble",
"probabilistic",
"classify",
"transition",
"branch",
"classifiers",
"mcmc",
"stochastic"
],
[
"outliers",
"deviations",
"statistically",
"approximations",
"lipschitz",
"asymptotically",
"accuracy",
"asymptotic",
"outlier",
"fitting"
],
[
"svm",
"images",
"art",
"classification",
"convolution",
"convolutional",
"gaussian",
"image",
"recognition",
"tensor"
],
[
"dense",
"forecasts",
"forecast",
"deep",
"pipeline",
"pipelines",
"pac",
"ensembles",
"allocation",
"ensemble"
],
[
"sampling",
"carlo",
"quantum",
"frequencies",
"bayesian",
"probabilities",
"bayes",
"experiments",
"dimensionality",
"discovery"
],
[
"interpolation",
"optical",
"sampled",
"collaborative",
"completion",
"consist",
"labeling",
"intermediate",
"labels",
"coherent"
],
[
"converge",
"convergence",
"simulations",
"researchers",
"perturbations",
"perturbation",
"converges",
"diffusion",
"variance",
"experimentally"
],
[
"neural",
"intuitive",
"generalizing",
"intuition",
"abstraction",
"brain",
"cognitive",
"perception",
"explanations",
"complexities"
],
[
"distributions",
"likelihood",
"densities",
"bayesian",
"bayes",
"posterior",
"variational",
"gaussian",
"distributional",
"distribution"
],
[
"distributional",
"phase",
"classes",
"class",
"phases",
"distributions",
"corpus",
"demonstrating",
"inducing",
"transient"
],
[
"tuning",
"selection",
"boosting",
"prediction",
"ensembles",
"ensemble",
"predicting",
"optimizing",
"track",
"sensing"
]
] | 323.824499 | all-MiniLM-L6-v2 | 0.863333 | -0.314059 | 0.175188 | 0.884025 |
ArXiv ML Papers | S³ | 46 | 30 | [
[
"incremental",
"iteration",
"posterior",
"forgetting",
"iteratively",
"iterative",
"iterations",
"perturbation",
"learns",
"diagnostic"
],
[
"configuration",
"enhance",
"deployment",
"enhancing",
"enable",
"superiority",
"configurations",
"calibration",
"feedback",
"preferences"
],
[
"converge",
"perturbations",
"convergence",
"converges",
"simulations",
"perturbation",
"diffusion",
"researchers",
"variance",
"experimentally"
],
[
"play",
"reinforcement",
"controller",
"behavioral",
"exploration",
"reward",
"game",
"agents",
"strategy",
"rewards"
],
[
"text",
"linguistic",
"texts",
"sentences",
"corpora",
"textual",
"literature",
"corpus",
"nlp",
"semantic"
],
[
"measurement",
"optimum",
"parameter",
"parameters",
"parametric",
"neuron",
"curve",
"coordinates",
"neurons",
"distances"
],
[
"tensor",
"images",
"image",
"recognition",
"art",
"convolution",
"convolutional",
"imagenet",
"svm",
"classification"
],
[
"resolution",
"sparse",
"denoising",
"imaging",
"subspace",
"matrices",
"cover",
"subspaces",
"compressed",
"solves"
],
[
"classification",
"supervised",
"classifiers",
"classifying",
"classifier",
"labeling",
"classify",
"labeled",
"backpropagation",
"regularization"
],
[
"gans",
"gan",
"generative",
"generating",
"generator",
"generate",
"generates",
"generated",
"autoencoder",
"generation"
],
[
"graphs",
"graph",
"vertex",
"networks",
"edges",
"vertices",
"nodes",
"node",
"edge",
"trees"
],
[
"optical",
"interpolation",
"consist",
"completion",
"intermediate",
"labeling",
"labels",
"collaborative",
"sampled",
"accuracies"
],
[
"predictor",
"hand",
"https",
"information",
"longer",
"policy",
"estimates",
"predictors",
"publicly",
"fidelity"
],
[
"monitoring",
"fail",
"reliability",
"uncertainties",
"uncertainty",
"reliable",
"validity",
"qualitative",
"uncertain",
"measuring"
],
[
"supervised",
"datasets",
"learners",
"learns",
"machine",
"samples",
"sample",
"batch",
"dataset",
"leads"
],
[
"densities",
"bayesian",
"likelihood",
"posterior",
"gaussian",
"distributions",
"distribution",
"variational",
"distributional",
"bayes"
],
[
"wearable",
"behavioral",
"signals",
"eeg",
"activity",
"behaviors",
"signal",
"monitoring",
"devices",
"device"
],
[
"sound",
"audio",
"music",
"voice",
"acoustic",
"speaker",
"frequencies",
"speech",
"noise",
"spectrum"
],
[
"gradient",
"hessian",
"convex",
"minimization",
"convexity",
"optimization",
"optimisation",
"minimize",
"gradients",
"optimizer"
],
[
"metric",
"metrics",
"distances",
"dimensional",
"dimensionality",
"similarity",
"distance",
"embeddings",
"dimension",
"dimensions"
],
[
"cognitive",
"intuition",
"explanations",
"brain",
"perception",
"abstraction",
"intuitive",
"generalizing",
"complexities",
"neural"
],
[
"distributional",
"phases",
"phase",
"corpus",
"class",
"classes",
"distributions",
"inducing",
"transient",
"demonstrating"
],
[
"public",
"centralized",
"private",
"shared",
"distributed",
"federated",
"decentralized",
"independently",
"privacy",
"sharing"
],
[
"ranging",
"gps",
"robotics",
"vision",
"3d",
"tracking",
"coordinates",
"track",
"motion",
"spatially"
],
[
"forecasts",
"forecast",
"forecasting",
"predicted",
"autoregressive",
"predicting",
"weather",
"predict",
"predicts",
"predictions"
],
[
"approximating",
"interpretability",
"approximations",
"interpretable",
"approximates",
"quantized",
"thresholding",
"threshold",
"approximation",
"quantum"
],
[
"clinical",
"medical",
"patients",
"diseases",
"healthcare",
"disease",
"mri",
"diagnosis",
"medicine",
"cancer"
],
[
"retrieval",
"seek",
"sampling",
"bandits",
"expectation",
"coverage",
"searching",
"greedy",
"selection",
"bandit"
],
[
"iot",
"urban",
"bandwidth",
"planning",
"distance",
"traffic",
"area",
"transportation",
"city",
"gps"
],
[
"algorithms",
"algorithmic",
"deterministic",
"computationally",
"synthesis",
"computational",
"algorithm",
"manipulation",
"expressive",
"preprocessing"
]
] | 358.120531 | all-MiniLM-L6-v2 | 0.95 | -0.261242 | 0.215086 | 0.902691 |
ArXiv ML Papers | S³ | 43 | 40 | [
[
"student",
"studying",
"learner",
"learners",
"teacher",
"revisit",
"learns",
"study",
"consistency",
"learnt"
],
[
"adversarial",
"exploit",
"adversary",
"attack",
"attacker",
"exploits",
"security",
"attacks",
"intrusion",
"threat"
],
[
"weaker",
"enhance",
"enhanced",
"enhances",
"superiority",
"sensitivity",
"exploring",
"susceptible",
"enhancement",
"spatially"
],
[
"accuracy",
"classifiers",
"classifier",
"randomized",
"encoder",
"similarity",
"algorithm",
"ranking",
"encode",
"encoded"
],
[
"convexity",
"convex",
"neurons",
"tend",
"represents",
"viewed",
"similar",
"grid",
"fields",
"attribute"
],
[
"dynamically",
"build",
"discover",
"iterations",
"iteration",
"inform",
"generate",
"posterior",
"iteratively",
"without"
],
[
"textual",
"texts",
"literature",
"corpus",
"linguistic",
"nlp",
"text",
"semantic",
"document",
"dictionary"
],
[
"robots",
"ai",
"machines",
"tasks",
"task",
"robot",
"robotic",
"machine",
"heuristics",
"automated"
],
[
"covid",
"represented",
"representing",
"interpret",
"represent",
"represents",
"interpretable",
"dynamics",
"environments",
"observations"
],
[
"medicine",
"clinical",
"disease",
"diseases",
"medical",
"patients",
"health",
"patient",
"healthcare",
"covid"
],
[
"public",
"influence",
"publicly",
"significant",
"principle",
"https",
"follow",
"consensus",
"principles",
"advances"
],
[
"rigorous",
"patterns",
"smoothness",
"smoothing",
"smooth",
"sequences",
"regularized",
"approximating",
"transitions",
"trajectories"
],
[
"bandit",
"exploration",
"retrieval",
"choosing",
"seek",
"optimality",
"greedy",
"searching",
"market",
"bandits"
],
[
"matrix",
"reconstructing",
"decoding",
"denoising",
"reconstructed",
"reconstruction",
"imputation",
"resolution",
"sparse",
"compressed"
],
[
"acceleration",
"fail",
"inequality",
"vectors",
"magnitude",
"efforts",
"vector",
"imbalance",
"impacts",
"attempts"
],
[
"convolutions",
"imagenet",
"recognition",
"probabilistic",
"representations",
"cnn",
"poses",
"unsupervised",
"convolution",
"bernoulli"
],
[
"reliability",
"recall",
"probabilistic",
"guarantees",
"risk",
"markov",
"discriminator",
"bayes",
"guarantee",
"safety"
],
[
"contrast",
"recognize",
"recognition",
"face",
"facial",
"recognizing",
"faces",
"present",
"identifies",
"contrastive"
],
[
"gan",
"gans",
"generative",
"generator",
"generates",
"generate",
"generating",
"generation",
"producing",
"generated"
],
[
"cluster",
"clusters",
"clustering",
"unsupervised",
"partitioning",
"communities",
"phases",
"segmentation",
"regions",
"neighborhood"
],
[
"correlated",
"parallel",
"tensors",
"vectors",
"vector",
"respectively",
"multivariate",
"tensor",
"dimensionality",
"cross"
],
[
"devices",
"eeg",
"tested",
"physical",
"running",
"wearable",
"device",
"arm",
"activity",
"monitoring"
],
[
"cognitive",
"brain",
"intuition",
"neuron",
"intelligence",
"neurons",
"neural",
"eeg",
"explanations",
"perception"
],
[
"insights",
"deep",
"discover",
"benchmarks",
"benchmark",
"tests",
"coverage",
"explored",
"challenging",
"metrics"
],
[
"combines",
"combine",
"split",
"ensemble",
"combination",
"combining",
"ensembles",
"hybrid",
"fusion",
"simultaneous"
],
[
"labels",
"labeling",
"labeled",
"classifiers",
"classifier",
"supervised",
"classifying",
"label",
"hessian",
"classification"
],
[
"uncertainty",
"empirically",
"digital",
"designing",
"flows",
"qualitative",
"rigorous",
"empirical",
"formalism",
"naive"
],
[
"mri",
"extraction",
"segmentation",
"segment",
"stages",
"imaging",
"screening",
"region",
"regions",
"diagnosis"
],
[
"sampling",
"dimensional",
"characteristic",
"dimensionality",
"characteristics",
"dimensions",
"dimension",
"consist",
"combinatorial",
"depth"
],
[
"forecasts",
"predicted",
"forecast",
"forecasting",
"weather",
"predictions",
"autoregressive",
"predict",
"predicts",
"predicting"
],
[
"deterministic",
"corpus",
"corpora",
"distributional",
"theoretical",
"speech",
"theory",
"hypothesis",
"structures",
"class"
],
[
"quantization",
"quantum",
"quantized",
"particle",
"observable",
"resonance",
"physics",
"accelerators",
"fourier",
"subspace"
],
[
"vertex",
"edges",
"graphs",
"nodes",
"graph",
"edge",
"vertices",
"node",
"spectral",
"networks"
],
[
"predicts",
"predict",
"predictive",
"calibration",
"confidence",
"predicted",
"trust",
"accuracy",
"model",
"calibrated"
],
[
"vehicle",
"transportation",
"gps",
"traffic",
"navigation",
"driving",
"transport",
"drive",
"distance",
"vehicles"
],
[
"98",
"asymptotically",
"throughput",
"latency",
"estimation",
"measuring",
"likelihood",
"fitting",
"asymptotic",
"correlations"
],
[
"symbolic",
"solvers",
"constraint",
"logic",
"diagnostic",
"relational",
"semantic",
"solver",
"constraints",
"abstraction"
],
[
"constrain",
"constraint",
"constraints",
"euclidean",
"linearly",
"constrained",
"dual",
"flexibility",
"symmetric",
"optimality"
],
[
"minimal",
"trees",
"interpretability",
"tree",
"noise",
"minimum",
"softmax",
"thresholding",
"threshold",
"approximations"
],
[
"boosting",
"boost",
"samples",
"classification",
"sample",
"actively",
"popular",
"classifier",
"sampling",
"collect"
]
] | 258.919685 | all-MiniLM-L6-v2 | 0.9275 | -0.307787 | 0.213316 | 0.907612 |
ArXiv ML Papers | S³ | 44 | 40 | [
[
"positive",
"risks",
"safety",
"reliability",
"guarantees",
"covariance",
"risk",
"probabilistic",
"markov",
"bayes"
],
[
"tend",
"distribution",
"distributions",
"regularized",
"regular",
"standard",
"preferences",
"distributional",
"parametric",
"parameters"
],
[
"fusion",
"combining",
"combination",
"dual",
"combine",
"simultaneously",
"combines",
"multi",
"combined",
"combinations"
],
[
"test",
"tests",
"testing",
"inspection",
"developers",
"patches",
"development",
"experimental",
"tested",
"screening"
],
[
"imbalance",
"vector",
"fail",
"causes",
"vectors",
"accelerate",
"decreases",
"magnitude",
"acceleration",
"attempts"
],
[
"edges",
"node",
"graphs",
"vertex",
"nodes",
"graph",
"vertices",
"edge",
"spectral",
"networks"
],
[
"manner",
"tensors",
"vertices",
"tensor",
"robotic",
"accelerators",
"scales",
"leads",
"robots",
"oriented"
],
[
"deterministic",
"theory",
"parallel",
"distributional",
"paradigm",
"hypothesis",
"theoretical",
"corpus",
"corpora",
"complexity"
],
[
"eeg",
"brain",
"cognitive",
"neuron",
"neurons",
"intuition",
"paths",
"intuitive",
"neural",
"mri"
],
[
"algorithm",
"predicts",
"predict",
"predictions",
"optimal",
"predicting",
"optimizing",
"optimizer",
"optimization",
"randomized"
],
[
"characteristic",
"factors",
"fidelity",
"transferring",
"ensures",
"characteristics",
"curse",
"transferred",
"character",
"transfer"
],
[
"trajectory",
"drive",
"traffic",
"gps",
"transport",
"vehicle",
"driving",
"transportation",
"navigation",
"vehicles"
],
[
"features",
"generalized",
"feature",
"lasso",
"bounded",
"sparsity",
"alleviate",
"benefits",
"extent",
"dual"
],
[
"autoencoder",
"autoencoders",
"encoder",
"encoders",
"encode",
"modelling",
"encodes",
"trainable",
"model",
"models"
],
[
"gaussian",
"likelihood",
"distributions",
"diffusion",
"densities",
"distribution",
"distributional",
"laplacian",
"covariance",
"processes"
],
[
"interpolation",
"jointly",
"maintaining",
"collaborative",
"resolution",
"marginal",
"structured",
"consist",
"outcomes",
"consistent"
],
[
"stability",
"perturbations",
"perturbation",
"perturbed",
"catastrophic",
"variance",
"dynamics",
"stable",
"correlated",
"coupled"
],
[
"gan",
"generative",
"generating",
"generator",
"generates",
"generate",
"gans",
"generation",
"producing",
"generated"
],
[
"millions",
"experimental",
"confidence",
"tested",
"researchers",
"experimentally",
"gan",
"accuracy",
"findings",
"research"
],
[
"tensors",
"embeddings",
"unstructured",
"compressed",
"storing",
"stored",
"consist",
"generate",
"truncated",
"rigorous"
],
[
"models",
"autoencoders",
"autoencoder",
"personalized",
"algorithm",
"algorithmic",
"algorithms",
"modeled",
"model",
"automated"
],
[
"resonance",
"fourier",
"particle",
"quantum",
"quantized",
"physics",
"quantization",
"correlated",
"accelerators",
"observable"
],
[
"biases",
"bias",
"discrimination",
"fairness",
"biased",
"equal",
"unbiased",
"discriminate",
"appealing",
"heterogeneity"
],
[
"seek",
"searching",
"bandits",
"bandit",
"greedy",
"market",
"strategy",
"exploration",
"combinatorial",
"heuristics"
],
[
"neurons",
"views",
"viewed",
"convex",
"convexity",
"tend",
"fields",
"grid",
"represents",
"mri"
],
[
"perceptual",
"signals",
"imitation",
"abstraction",
"covariance",
"notions",
"behavioral",
"behaviors",
"invariance",
"dependencies"
],
[
"mri",
"basis",
"decomposition",
"denoising",
"tensor",
"matrices",
"matrix",
"imputation",
"tensors",
"factorization"
],
[
"dramatically",
"variational",
"substantially",
"sparse",
"severe",
"minimal",
"insufficient",
"compressed",
"drastically",
"distinct"
],
[
"dimensionality",
"characteristic",
"characteristics",
"representation",
"representations",
"dimension",
"transformers",
"sampling",
"dimensional",
"dimensions"
],
[
"priors",
"classifiers",
"classify",
"hessian",
"bayesian",
"conventional",
"softmax",
"classifying",
"classifier",
"autoregressive"
],
[
"forecast",
"forecasts",
"weather",
"predictions",
"prediction",
"predicting",
"predicted",
"predict",
"forecasting",
"predicts"
],
[
"visualization",
"classification",
"video",
"convolutions",
"decisions",
"quantitative",
"videos",
"audio",
"music",
"features"
],
[
"transforms",
"dependent",
"cross",
"uncertainties",
"information",
"visual",
"authors",
"rather",
"blind",
"uncertainty"
],
[
"spread",
"downstream",
"dense",
"decentralized",
"pipelines",
"resnet",
"acquire",
"pipeline",
"cnn",
"widespread"
],
[
"convex",
"width",
"covariates",
"jointly",
"multivariate",
"multi",
"vectors",
"network",
"networks",
"centralized"
],
[
"behavioral",
"policy",
"enforce",
"reinforcement",
"behaviors",
"action",
"interventions",
"actions",
"rewards",
"policies"
],
[
"encode",
"spectral",
"recognition",
"normalized",
"accuracy",
"measure",
"randomized",
"gaussian",
"classifier",
"mri"
],
[
"similarity",
"distance",
"distances",
"comparable",
"ranking",
"nearest",
"metrics",
"compares",
"similarities",
"metric"
],
[
"task",
"uncertain",
"uncertainties",
"uncertainty",
"processes",
"reliability",
"measurement",
"machines",
"tasks",
"robots"
],
[
"sequential",
"forecasts",
"rnns",
"autoregressive",
"lstm",
"sequence",
"sequences",
"temporal",
"recurrent",
"forecasting"
]
] | 331.399202 | all-MiniLM-L6-v2 | 0.8575 | -0.326221 | 0.210523 | 0.896985 |
ArXiv ML Papers | S³ | 45 | 40 | [
[
"equal",
"discrimination",
"biased",
"bias",
"unbiased",
"fairness",
"discriminate",
"biases",
"assess",
"representative"
],
[
"tensors",
"tensor",
"vectors",
"matrices",
"factorization",
"matrix",
"vector",
"multivariate",
"decomposition",
"3d"
],
[
"reliability",
"consistent",
"uncertainties",
"accuracy",
"uncertainty",
"reliable",
"consistency",
"consistently",
"precision",
"uncertain"
],
[
"attacks",
"exploits",
"exploit",
"attacker",
"adversarial",
"attack",
"vulnerability",
"adversary",
"intrusion",
"security"
],
[
"samples",
"sample",
"tested",
"screening",
"finds",
"datasets",
"benchmark",
"dataset",
"batch",
"discover"
],
[
"lstm",
"recurrent",
"sequences",
"rnn",
"sequential",
"autoregressive",
"sequence",
"forecasting",
"temporal",
"rnns"
],
[
"flexibility",
"potential",
"scalability",
"marginal",
"scalable",
"shape",
"multivariate",
"jointly",
"covariates",
"distributional"
],
[
"dimensionality",
"spectral",
"perturbation",
"perturbations",
"laplacian",
"covariance",
"landscape",
"estimates",
"gaussian",
"diffusion"
],
[
"weather",
"forecasting",
"forecasts",
"forecast",
"predicted",
"predicting",
"predicts",
"predict",
"predictions",
"prediction"
],
[
"predict",
"encode",
"encodes",
"algorithm",
"predicts",
"predictive",
"randomized",
"predictions",
"predicting",
"prediction"
],
[
"embedding",
"embeddings",
"estimator",
"monitoring",
"estimation",
"description",
"solvers",
"estimates",
"estimate",
"estimators"
],
[
"covariance",
"perceptual",
"invariance",
"signals",
"abstraction",
"notions",
"behavioral",
"perception",
"imitation",
"dependencies"
],
[
"capabilities",
"objectives",
"multivariate",
"projections",
"projection",
"goals",
"fidelity",
"mri",
"incorporate",
"objective"
],
[
"recommendation",
"personalized",
"factorization",
"appealing",
"rank",
"popularity",
"ranking",
"recommendations",
"recommender",
"preferences"
],
[
"convolutions",
"imagenet",
"representations",
"mnist",
"poses",
"cnn",
"unsupervised",
"pose",
"threat",
"bernoulli"
],
[
"probabilistic",
"markov",
"monte",
"carlo",
"simulation",
"bayesian",
"mcmc",
"posterior",
"stochastic",
"simulations"
],
[
"smoothing",
"dynamical",
"dynamics",
"modeled",
"differentially",
"nonlinear",
"smoothness",
"modeling",
"rigorous",
"modelling"
],
[
"acoustic",
"audio",
"voice",
"sound",
"speaker",
"speech",
"interplay",
"communicate",
"music",
"frequencies"
],
[
"normalization",
"regularization",
"normalizing",
"normalized",
"backpropagation",
"regularized",
"transformation",
"factorization",
"transformations",
"invariant"
],
[
"vertex",
"graph",
"graphs",
"vertices",
"nodes",
"networks",
"edges",
"edge",
"spectral",
"node"
],
[
"considerable",
"substantially",
"sufficiently",
"insufficient",
"approximates",
"dramatically",
"gnn",
"threshold",
"minimal",
"metrics"
],
[
"stages",
"extraction",
"edges",
"period",
"periods",
"segment",
"segmentation",
"stage",
"inspection",
"labeling"
],
[
"fail",
"vectors",
"acceleration",
"magnitude",
"vector",
"attempts",
"imbalance",
"accelerate",
"causes",
"decreases"
],
[
"explanations",
"interpretability",
"interpretable",
"interpret",
"explaining",
"visual",
"describes",
"interpretation",
"intuition",
"attention"
],
[
"resnet",
"decentralized",
"pipelines",
"pipeline",
"filters",
"denoising",
"spread",
"downstream",
"filtering",
"flows"
],
[
"quantum",
"quantization",
"particle",
"resonance",
"quantized",
"fourier",
"observable",
"physics",
"accelerators",
"correlated"
],
[
"diversity",
"diverse",
"individually",
"combines",
"broadly",
"ensembles",
"ensemble",
"weakly",
"experimental",
"weak"
],
[
"digital",
"visualization",
"decisions",
"videos",
"denoising",
"audio",
"music",
"probabilistic",
"classification",
"video"
],
[
"ai",
"robot",
"robotic",
"robots",
"machines",
"tasks",
"autonomous",
"machine",
"automated",
"robotics"
],
[
"transform",
"conventional",
"dependent",
"blind",
"information",
"visual",
"transformation",
"derive",
"transforms",
"encoded"
],
[
"process",
"decentralized",
"randomized",
"machines",
"decomposition",
"generator",
"weights",
"decompose",
"processes",
"variability"
],
[
"trainable",
"rank",
"ranging",
"attention",
"rnn",
"ranking",
"boosting",
"selection",
"vertex",
"models"
],
[
"scores",
"parametric",
"measurement",
"measured",
"precision",
"reach",
"examine",
"offs",
"measuring",
"assess"
],
[
"sensors",
"detect",
"ucb",
"basic",
"system",
"languages",
"framework",
"sparse",
"systems",
"detector"
],
[
"softmax",
"gradients",
"gradient",
"nonparametric",
"hessian",
"conventional",
"derivatives",
"contrastive",
"differentiation",
"priors"
],
[
"asymptotically",
"latency",
"fitting",
"likelihood",
"asymptotic",
"throughput",
"bottleneck",
"98",
"measuring",
"phenomenon"
],
[
"characteristic",
"representations",
"dimension",
"dimensional",
"dimensionality",
"dimensions",
"sampling",
"consist",
"characteristics",
"representation"
],
[
"measure",
"softmax",
"ranking",
"encode",
"randomized",
"classifier",
"algorithm",
"encoder",
"spectral",
"classifiers"
],
[
"intuition",
"neurons",
"eeg",
"cognitive",
"densities",
"brain",
"mri",
"neural",
"neuron",
"analytically"
],
[
"calibration",
"networks",
"models",
"modeled",
"modelling",
"model",
"modeling",
"simulation",
"suited",
"interpolation"
]
] | 321.9191 | all-MiniLM-L6-v2 | 0.8725 | -0.321882 | 0.22006 | 0.902597 |
ArXiv ML Papers | S³ | 46 | 40 | [
[
"beneficial",
"mainly",
"boosting",
"giving",
"attention",
"boost",
"extracts",
"popular",
"interest",
"traffic"
],
[
"visual",
"cross",
"passing",
"via",
"information",
"transforms",
"blind",
"derive",
"rather",
"dependent"
],
[
"extraction",
"mri",
"segmentation",
"segment",
"imaging",
"stages",
"regions",
"screening",
"region",
"diagnosis"
],
[
"generalizes",
"multimodal",
"generalization",
"generalizing",
"incremental",
"generalize",
"granularity",
"parameterized",
"automated",
"procedures"
],
[
"visualization",
"digital",
"videos",
"probabilistic",
"audio",
"music",
"decisions",
"classifier",
"classification",
"video"
],
[
"contexts",
"pixels",
"grounded",
"contextual",
"spectral",
"context",
"convolutions",
"scenes",
"notions",
"distinguish"
],
[
"reconstructed",
"dynamics",
"equations",
"nonlinear",
"analyze",
"denoising",
"analytical",
"simulations",
"solvers",
"analyse"
],
[
"resnet",
"tail",
"energy",
"bounded",
"bandwidth",
"yielding",
"consumption",
"evolution",
"evolving",
"continuously"
],
[
"vertex",
"rnn",
"rank",
"gps",
"attention",
"autoencoder",
"ranging",
"trainable",
"rnns",
"autoencoders"
],
[
"cnn",
"mnist",
"convolutions",
"probabilistic",
"unbiased",
"backpropagation",
"imagenet",
"cnns",
"off",
"networks"
],
[
"averaging",
"outcomes",
"meaningful",
"marginal",
"maintaining",
"consist",
"consistent",
"collaborative",
"aggregation",
"resolution"
],
[
"gan",
"gans",
"generating",
"generative",
"generator",
"producing",
"generates",
"generate",
"generation",
"generated"
],
[
"diverse",
"individually",
"diversity",
"ensemble",
"ensembles",
"multi",
"mixed",
"combines",
"multiple",
"broadly"
],
[
"sequences",
"recurrent",
"temporal",
"sequence",
"autoregressive",
"forecasting",
"sequential",
"lstm",
"rnns",
"forecasts"
],
[
"softmax",
"backpropagation",
"hessian",
"gradient",
"gradients",
"classifiers",
"classifying",
"contrastive",
"regression",
"labeling"
],
[
"modeled",
"predict",
"model",
"predictive",
"models",
"predicts",
"modeling",
"modelling",
"predictors",
"predictor"
],
[
"convexity",
"tend",
"neurons",
"convex",
"represents",
"fields",
"similar",
"views",
"grid",
"viewed"
],
[
"nonlinear",
"invariant",
"dynamics",
"dynamical",
"stability",
"notions",
"embeddings",
"flows",
"stationary",
"geometric"
],
[
"bandit",
"greedy",
"seek",
"bandits",
"searching",
"market",
"combinatorial",
"choosing",
"strategy",
"optimality"
],
[
"combines",
"separately",
"ensembles",
"additive",
"mixture",
"combining",
"formulation",
"combine",
"theoretic",
"concepts"
],
[
"gaussian",
"parametric",
"parameters",
"distribution",
"distributions",
"standard",
"posterior",
"activation",
"parameter",
"criterion"
],
[
"gaussian",
"landscape",
"estimates",
"laplacian",
"asymptotically",
"estimators",
"estimate",
"variance",
"expectation",
"perturbation"
],
[
"decentralized",
"weights",
"decomposition",
"process",
"core",
"generator",
"processes",
"variability",
"decompose",
"softmax"
],
[
"lasso",
"regularization",
"setting",
"sparse",
"sparsity",
"constrained",
"flexibility",
"constrain",
"trivial",
"regularized"
],
[
"optimizing",
"optimization",
"optimisation",
"optimize",
"optimizer",
"maximizing",
"maximization",
"constraints",
"constrained",
"optimized"
],
[
"enforce",
"behaviors",
"actions",
"reinforcement",
"behavioral",
"policy",
"controlled",
"policies",
"behavior",
"action"
],
[
"transfer",
"complementary",
"transforms",
"transferred",
"domains",
"adaptation",
"transferring",
"transformation",
"across",
"pretrained"
],
[
"convexity",
"convex",
"gradients",
"backpropagation",
"perturbation",
"sgd",
"accelerate",
"gradient",
"accelerated",
"hessian"
],
[
"forecasting",
"forecast",
"forecasts",
"weather",
"predictions",
"predicted",
"predict",
"predicting",
"predicts",
"autoregressive"
],
[
"stochastic",
"latent",
"variational",
"autoencoder",
"bayesian",
"likelihood",
"autoencoders",
"modelling",
"modeling",
"modeled"
],
[
"eeg",
"tested",
"generator",
"physical",
"device",
"running",
"devices",
"wearable",
"activity",
"testing"
],
[
"observation",
"sampled",
"sampling",
"estimator",
"observations",
"population",
"observed",
"estimators",
"sample",
"noisy"
],
[
"fairness",
"bias",
"biases",
"discrimination",
"biased",
"discriminate",
"equal",
"heterogeneity",
"representative",
"unbiased"
],
[
"eeg",
"perception",
"neuron",
"cognitive",
"intelligence",
"attention",
"brain",
"perceptual",
"imaging",
"focusing"
],
[
"combines",
"multi",
"iot",
"simultaneously",
"task",
"deployment",
"distributed",
"communicate",
"tasks",
"availability"
],
[
"tensors",
"robots",
"machine",
"tasks",
"machines",
"tensor",
"robotic",
"leads",
"perform",
"processes"
],
[
"recommender",
"factorization",
"recommendations",
"recommendation",
"appealing",
"personalized",
"ranking",
"rank",
"preferences",
"popularity"
],
[
"boundaries",
"volume",
"paths",
"densities",
"boundary",
"density",
"divergence",
"partitioning",
"pipelines",
"dense"
],
[
"causal",
"covariates",
"interventions",
"interactions",
"factors",
"impacts",
"effects",
"facilitate",
"involve",
"interaction"
],
[
"causes",
"attempts",
"magnitude",
"fail",
"vectors",
"vector",
"acceleration",
"efforts",
"imbalance",
"inequality"
]
] | 268.612561 | all-MiniLM-L6-v2 | 0.8925 | -0.325224 | 0.19703 | 0.909345 |
ArXiv ML Papers | S³ | 43 | 50 | [
[
"flows",
"downstream",
"pipelines",
"pipeline",
"directed",
"flow",
"superior",
"centralized",
"causal",
"predictive"
],
[
"representation",
"grid",
"patient",
"embedding",
"embeddings",
"matrix",
"subspaces",
"random",
"subspace",
"matrices"
],
[
"exploring",
"sensitivity",
"integrating",
"intelligent",
"uncertainty",
"explored",
"superiority",
"exploration",
"susceptible",
"enhance"
],
[
"contrastive",
"utility",
"frequencies",
"convolutional",
"importance",
"hypothesis",
"complexities",
"convolution",
"distributions",
"efforts"
],
[
"biases",
"biased",
"discrimination",
"discriminate",
"outcomes",
"fair",
"fairness",
"bias",
"representative",
"equal"
],
[
"interpretation",
"financial",
"interpretability",
"interpretable",
"industry",
"credit",
"products",
"production",
"interpret",
"product"
],
[
"kernels",
"gaussian",
"kernel",
"covariance",
"multivariate",
"convolution",
"processes",
"convolutions",
"convolutional",
"covariates"
],
[
"projections",
"neuron",
"layer",
"layers",
"neurons",
"network",
"neural",
"networks",
"backpropagation",
"net"
],
[
"counterfactual",
"interventions",
"aid",
"causal",
"involve",
"impacts",
"imputation",
"interactions",
"associated",
"consequence"
],
[
"impacts",
"popularity",
"perturbations",
"effect",
"perturbation",
"scale",
"scaling",
"magnitude",
"diffusion",
"scales"
],
[
"transforms",
"adaptation",
"transferred",
"transfer",
"transferring",
"transform",
"transformation",
"transformations",
"adapting",
"adapt"
],
[
"critic",
"interventions",
"policy",
"behaviors",
"behavioral",
"enforce",
"policies",
"reinforcement",
"textit",
"actions"
],
[
"computation",
"throughput",
"implementations",
"accelerators",
"computing",
"performance",
"computations",
"processing",
"latency",
"hardware"
],
[
"tensors",
"sensors",
"ucb",
"basic",
"languages",
"tensor",
"molecular",
"basis",
"detect",
"calibration"
],
[
"nearest",
"metrics",
"comparable",
"compares",
"distances",
"similarity",
"distance",
"metric",
"quantify",
"comparison"
],
[
"extending",
"boundary",
"acoustic",
"sampling",
"constructing",
"extends",
"trees",
"arbitrary",
"extensive",
"construction"
],
[
"convex",
"convexity",
"frequently",
"represents",
"increasingly",
"pattern",
"naive",
"fields",
"image",
"grid"
],
[
"transforms",
"dynamics",
"transitions",
"flexible",
"dynamical",
"invariant",
"dynamic",
"preserving",
"transformations",
"flows"
],
[
"bounded",
"demand",
"embedding",
"notions",
"imagenet",
"energy",
"half",
"threat",
"representations",
"camera"
],
[
"optimization",
"constraint",
"constraints",
"minimum",
"constrained",
"sensitivity",
"proposes",
"allocation",
"optimisation",
"range"
],
[
"attacks",
"attacker",
"adversaries",
"adversarially",
"adversarial",
"exploits",
"security",
"adversary",
"exploit",
"threat"
],
[
"character",
"limitation",
"limitations",
"allowed",
"factors",
"restricted",
"allow",
"able",
"allowing",
"prevalence"
],
[
"labels",
"labeling",
"label",
"labeled",
"classify",
"categories",
"unlabeled",
"taxonomy",
"classification",
"classified"
],
[
"pose",
"robotics",
"objects",
"robotic",
"poses",
"robot",
"robots",
"joint",
"hand",
"flexibility"
],
[
"assess",
"phases",
"extraction",
"verify",
"spectral",
"distinguish",
"partial",
"confirm",
"intrinsic",
"evaluate"
],
[
"market",
"seek",
"bandits",
"exploitation",
"bandit",
"acquisition",
"searching",
"acquiring",
"exploration",
"acquire"
],
[
"protein",
"molecular",
"chain",
"generate",
"structural",
"discover",
"synthesis",
"genetic",
"generating",
"mechanisms"
],
[
"acceleration",
"wireless",
"reflect",
"descent",
"targets",
"impacts",
"signal",
"fail",
"completion",
"down"
],
[
"markov",
"stochastic",
"monte",
"simulations",
"simulation",
"randomly",
"mcmc",
"carlo",
"sampling",
"random"
],
[
"multimodal",
"empirical",
"spatio",
"corpus",
"contextual",
"granularity",
"spatiotemporal",
"software",
"vanilla",
"tend"
],
[
"risk",
"vary",
"marginal",
"dependence",
"covariates",
"divergence",
"width",
"risks",
"potential",
"wide"
],
[
"dimensionality",
"capacity",
"dimension",
"dimensional",
"allocation",
"dataset",
"datasets",
"selects",
"representations",
"aggregation"
],
[
"reconstructed",
"compression",
"channel",
"reconstruct",
"encoded",
"decoder",
"decoding",
"compress",
"decompose",
"denoising"
],
[
"corpora",
"nlp",
"sentences",
"semantic",
"text",
"linguistic",
"corpus",
"textual",
"texts",
"literature"
],
[
"minimal",
"distinct",
"compressed",
"compress",
"reduces",
"reduced",
"pruning",
"pruned",
"compression",
"trivial"
],
[
"diffusion",
"processes",
"bottleneck",
"scenes",
"process",
"flow",
"scene",
"content",
"topic",
"throughput"
],
[
"dependent",
"individual",
"attributes",
"outcomes",
"lightweight",
"authors",
"attribute",
"independent",
"passing",
"last"
],
[
"continuous",
"ensemble",
"combines",
"ensembles",
"streams",
"stream",
"combine",
"collect",
"flows",
"entries"
],
[
"autoencoder",
"autoencoders",
"encodes",
"encoders",
"encode",
"encoder",
"variational",
"code",
"encoded",
"encoding"
],
[
"producing",
"poses",
"production",
"core",
"embeddings",
"energy",
"embedded",
"embedding",
"relu",
"soft"
],
[
"federated",
"distributed",
"cooperative",
"communicate",
"communications",
"shared",
"protocol",
"centralized",
"collaborative",
"communication"
],
[
"edges",
"node",
"graph",
"vertices",
"vertex",
"nodes",
"graphs",
"edge",
"graphical",
"networks"
],
[
"patients",
"clinical",
"mri",
"imaging",
"diagnostic",
"diagnosis",
"medical",
"medicine",
"cancer",
"healthcare"
],
[
"defines",
"dynamically",
"termed",
"characterized",
"recurrent",
"describes",
"monte",
"initialization",
"defined",
"continuously"
],
[
"style",
"imitation",
"manipulation",
"proposes",
"manner",
"patterns",
"formalism",
"designs",
"traditional",
"fashion"
],
[
"forest",
"classifiers",
"classifier",
"boosting",
"tree",
"boost",
"ensemble",
"trees",
"branch",
"classify"
],
[
"recommender",
"preferences",
"ranking",
"recommendations",
"recommendation",
"factorization",
"personalized",
"suggests",
"collaborative",
"factor"
],
[
"outlier",
"event",
"deviation",
"anomaly",
"anomalies",
"iot",
"outliers",
"particle",
"intrusion",
"distributions"
],
[
"quality",
"parameters",
"estimates",
"thousands",
"hundreds",
"full",
"demand",
"patches",
"uniformly",
"round"
],
[
"differentiation",
"derivatives",
"nonlinear",
"hessian",
"dynamics",
"equations",
"differentiable",
"differential",
"differentially",
"implicit"
]
] | 315.762408 | all-MiniLM-L6-v2 | 0.908 | -0.330432 | 0.200003 | 0.908556 |
ArXiv ML Papers | S³ | 44 | 50 | [
[
"recurrent",
"forecasting",
"demand",
"forecast",
"lstm",
"forecasts",
"sequential",
"records",
"sequences",
"temporal"
],
[
"protein",
"molecular",
"structural",
"chain",
"discover",
"genetic",
"return",
"generate",
"module",
"domains"
],
[
"publicly",
"privacy",
"private",
"federated",
"independent",
"independently",
"public",
"conventional",
"differentially",
"standard"
],
[
"video",
"social",
"principle",
"lead",
"twitter",
"https",
"post",
"public",
"influence",
"publicly"
],
[
"transformations",
"transforms",
"invariant",
"preserving",
"aggregation",
"transform",
"transformation",
"dynamical",
"dynamics",
"balance"
],
[
"euclidean",
"norm",
"lipschitz",
"preferences",
"bound",
"enables",
"criterion",
"regularization",
"lasso",
"bounded"
],
[
"boost",
"collect",
"feature",
"features",
"boosting",
"character",
"collected",
"collection",
"speedups",
"driving"
],
[
"gan",
"generative",
"gans",
"generator",
"producing",
"generates",
"generate",
"generating",
"generation",
"hilbert"
],
[
"mixture",
"partial",
"mini",
"composed",
"aspect",
"aspects",
"minimax",
"ensembles",
"localized",
"theoretic"
],
[
"density",
"embeddings",
"capacity",
"incorporate",
"inequality",
"embedding",
"densities",
"embedded",
"defend",
"divergence"
],
[
"bayesian",
"priors",
"posterior",
"prior",
"bayes",
"likelihood",
"distributions",
"distributional",
"priori",
"densities"
],
[
"exploration",
"inform",
"modes",
"iteration",
"iterations",
"expansion",
"massive",
"incremental",
"perturbations",
"perturbation"
],
[
"sampling",
"empirically",
"sample",
"experiments",
"empirical",
"observations",
"ensembles",
"samples",
"researchers",
"systematically"
],
[
"modeled",
"models",
"modeling",
"structure",
"computationally",
"model",
"modelling",
"imitation",
"formulate",
"structures"
],
[
"gradient",
"combines",
"fusion",
"fidelity",
"coupled",
"mutual",
"share",
"mismatch",
"complementary",
"backpropagation"
],
[
"factor",
"tensor",
"tensors",
"latent",
"matrices",
"vector",
"factorization",
"pytorch",
"factors",
"vectors"
],
[
"audio",
"detecting",
"recommender",
"detect",
"ensemble",
"videos",
"predictors",
"classifier",
"prediction",
"probabilistic"
],
[
"recommender",
"factorization",
"recommendation",
"suggests",
"recommendations",
"collaborative",
"ranking",
"profiles",
"suggest",
"personalized"
],
[
"industry",
"interpretable",
"interpretability",
"interpretation",
"development",
"production",
"describes",
"developers",
"service",
"industrial"
],
[
"production",
"embedding",
"poses",
"soft",
"embeddings",
"relu",
"embedded",
"core",
"crafted",
"github"
],
[
"adaptively",
"individuals",
"person",
"predictor",
"deterministic",
"controller",
"adaptation",
"known",
"similarity",
"inefficient"
],
[
"brain",
"cognitive",
"neural",
"neuron",
"eeg",
"neurons",
"mri",
"intractable",
"coherent",
"intelligence"
],
[
"measure",
"outcomes",
"aggregate",
"averaging",
"nonnegative",
"measures",
"marginal",
"score",
"quantify",
"averaged"
],
[
"dependent",
"attributes",
"outcomes",
"independent",
"get",
"attribute",
"expected",
"gives",
"passing",
"lightweight"
],
[
"resonance",
"hilbert",
"quantization",
"particle",
"quantized",
"quantum",
"fourier",
"spectrum",
"signal",
"discovery"
],
[
"nearest",
"estimated",
"distance",
"sensitivity",
"proposes",
"estimation",
"minimum",
"near",
"approximation",
"estimate"
],
[
"camera",
"representations",
"machines",
"demand",
"energy",
"sensing",
"recognition",
"imagenet",
"notions",
"convolutions"
],
[
"threshold",
"characterized",
"dramatically",
"severe",
"restricted",
"periods",
"tractable",
"sparse",
"compressed",
"period"
],
[
"gan",
"scoring",
"scores",
"assessment",
"rank",
"assess",
"highest",
"assessing",
"ranking",
"surrogate"
],
[
"bayes",
"detection",
"probabilistic",
"markov",
"ensuring",
"expectation",
"observing",
"security",
"guarantees",
"outcomes"
],
[
"values",
"filtering",
"numbers",
"filters",
"numerically",
"sampling",
"numerical",
"filter",
"parameters",
"statistics"
],
[
"optimizing",
"optimisation",
"objectives",
"optimizes",
"simultaneously",
"multivariate",
"optimize",
"multi",
"optimization",
"dual"
],
[
"decomposition",
"maintaining",
"normalizing",
"normalized",
"normalization",
"factorization",
"regularization",
"transformations",
"invariant",
"transform"
],
[
"spectrum",
"dependency",
"frequency",
"frequencies",
"contrast",
"contrastive",
"spectral",
"prevalence",
"utility",
"sign"
],
[
"randomized",
"boosting",
"classifiers",
"classifier",
"transformer",
"recognition",
"ranking",
"algorithm",
"ensemble",
"algorithms"
],
[
"weather",
"forecast",
"predictions",
"forecasting",
"forecasts",
"predicted",
"predict",
"reflect",
"predicts",
"predicting"
],
[
"convex",
"convexity",
"trend",
"fields",
"tend",
"represents",
"image",
"naive",
"attribute",
"grid"
],
[
"model",
"models",
"latency",
"fitting",
"candidate",
"availability",
"feasibility",
"joint",
"phenomenon",
"having"
],
[
"auc",
"reliability",
"infinite",
"dnn",
"dnns",
"imbalance",
"finite",
"ratio",
"imbalanced",
"magnitude"
],
[
"computing",
"hardware",
"accelerators",
"processing",
"performance",
"computation",
"latency",
"throughput",
"cloud",
"implementations"
],
[
"validity",
"validate",
"counterfactual",
"profiles",
"methodology",
"identifies",
"reliable",
"validation",
"determined",
"automated"
],
[
"flows",
"pipelines",
"downstream",
"flow",
"pipeline",
"superior",
"centralized",
"decentralized",
"causal",
"catastrophic"
],
[
"vision",
"attention",
"focus",
"focusing",
"observing",
"perceptual",
"summary",
"eye",
"leading",
"focuses"
],
[
"diffusion",
"traffic",
"bandwidth",
"formulation",
"mixture",
"terms",
"flow",
"granularity",
"compression",
"downstream"
],
[
"clinical",
"healthcare",
"patient",
"disease",
"medicine",
"patients",
"medical",
"diseases",
"health",
"practitioners"
],
[
"hessian",
"gradient",
"sgd",
"convex",
"gradients",
"descent",
"convexity",
"backpropagation",
"ascent",
"accelerated"
],
[
"program",
"detect",
"detecting",
"recursive",
"algorithms",
"algorithmic",
"algorithm",
"solvers",
"implementations",
"programming"
],
[
"coefficients",
"selected",
"detector",
"include",
"estimator",
"82",
"estimators",
"themselves",
"83",
"coefficient"
],
[
"evaluations",
"analyse",
"contrastive",
"wasserstein",
"assessing",
"contrast",
"hessian",
"imaging",
"assessment",
"diagnostic"
],
[
"dependence",
"covariates",
"potential",
"stable",
"vary",
"specialized",
"wide",
"scalability",
"oriented",
"width"
]
] | 347.488215 | all-MiniLM-L6-v2 | 0.884 | -0.332021 | 0.182943 | 0.910486 |
ArXiv ML Papers | S³ | 45 | 50 | [
[
"collect",
"ensembles",
"consist",
"ensemble",
"defined",
"generator",
"numerically",
"continuous",
"determined",
"collected"
],
[
"effects",
"magnitude",
"scale",
"popularity",
"perturbations",
"perturbation",
"scaling",
"impacts",
"impact",
"effect"
],
[
"energy",
"imagenet",
"representations",
"convolution",
"recognition",
"convolutions",
"machine",
"robots",
"bernoulli",
"robotic"
],
[
"sparse",
"severe",
"threshold",
"characterized",
"tractable",
"dramatically",
"compressed",
"periods",
"gender",
"dense"
],
[
"clusters",
"clustering",
"segment",
"segmentation",
"unsupervised",
"cluster",
"regions",
"partitioning",
"separation",
"overlap"
],
[
"flow",
"scene",
"bottleneck",
"scenes",
"lipschitz",
"process",
"throughput",
"processes",
"diffusion",
"differentiation"
],
[
"layer",
"backpropagation",
"neurons",
"layers",
"networks",
"projections",
"neural",
"projection",
"neuron",
"network"
],
[
"transformations",
"regularization",
"normalized",
"characterize",
"invariant",
"characterizing",
"decomposition",
"normalizing",
"normalization",
"maintaining"
],
[
"projections",
"task",
"projection",
"procedures",
"measurement",
"decade",
"technologies",
"measurements",
"tasks",
"reliability"
],
[
"mini",
"partial",
"mixture",
"aspects",
"minimax",
"ensembles",
"aspect",
"composed",
"phase",
"phases"
],
[
"monte",
"defined",
"characterized",
"dynamically",
"describes",
"defines",
"continuously",
"designed",
"termed",
"compact"
],
[
"expect",
"calibration",
"predicted",
"accuracy",
"validate",
"validated",
"validation",
"calibrated",
"scores",
"predictions"
],
[
"patient",
"diagnosis",
"disease",
"diseases",
"clinical",
"healthcare",
"medical",
"medicine",
"mri",
"patients"
],
[
"hessian",
"gradient",
"convex",
"sgd",
"convexity",
"backpropagation",
"gradients",
"descent",
"convergence",
"accelerated"
],
[
"tail",
"scoring",
"https",
"leading",
"goals",
"goal",
"lead",
"public",
"curve",
"fast"
],
[
"matrix",
"matrices",
"linear",
"subspace",
"patient",
"representation",
"subspaces",
"grid",
"representations",
"vectors"
],
[
"dynamic",
"constrained",
"transformations",
"dynamics",
"preserving",
"dynamical",
"transforms",
"flows",
"invariant",
"flexible"
],
[
"learner",
"learns",
"student",
"learners",
"study",
"studying",
"teacher",
"classes",
"studied",
"class"
],
[
"protein",
"genetic",
"structural",
"mechanisms",
"molecular",
"chain",
"mechanism",
"module",
"biology",
"biological"
],
[
"embeddings",
"energy",
"soft",
"core",
"embedding",
"poses",
"embedded",
"relu",
"crafted",
"weakly"
],
[
"production",
"development",
"industrial",
"industry",
"processes",
"service",
"product",
"process",
"products",
"developers"
],
[
"recurrent",
"sequences",
"lstm",
"sequence",
"temporal",
"rnns",
"sequential",
"rnn",
"forecasts",
"forecasting"
],
[
"discrimination",
"bias",
"discriminate",
"biases",
"fairness",
"biased",
"outcomes",
"representative",
"equal",
"fair"
],
[
"softmax",
"generalization",
"characteristic",
"generalize",
"randomly",
"generalizing",
"summarize",
"ensemble",
"distinguishing",
"random"
],
[
"modes",
"behavioral",
"action",
"interventions",
"reinforcement",
"actions",
"policy",
"policies",
"behaviors",
"enforce"
],
[
"spectrum",
"quantization",
"particle",
"signal",
"observable",
"quantized",
"quantum",
"hilbert",
"resonance",
"fourier"
],
[
"collect",
"82",
"incorporates",
"themselves",
"coefficients",
"detector",
"demand",
"estimator",
"popular",
"screening"
],
[
"complementary",
"combines",
"mixed",
"coupled",
"mutual",
"fusion",
"fidelity",
"mismatch",
"gradient",
"share"
],
[
"heuristics",
"optimization",
"optimize",
"optimisation",
"algorithm",
"optimized",
"heuristic",
"algorithms",
"optimizing",
"transformer"
],
[
"lipschitz",
"norm",
"lasso",
"euclidean",
"regularization",
"enables",
"preferences",
"generalized",
"operators",
"linearly"
],
[
"restricted",
"allowed",
"encoded",
"risks",
"codes",
"deal",
"risk",
"curse",
"encoding",
"allow"
],
[
"completion",
"down",
"fail",
"descent",
"reflect",
"impacts",
"efforts",
"targets",
"acceleration",
"poisoning"
],
[
"detecting",
"audio",
"videos",
"perception",
"detect",
"sensing",
"monitoring",
"ensemble",
"recommender",
"predictors"
],
[
"differential",
"hessian",
"implicit",
"differentiation",
"nonlinear",
"derivatives",
"differentially",
"equations",
"dynamics",
"modelling"
],
[
"ucb",
"tensors",
"sensors",
"basic",
"calibration",
"languages",
"tensor",
"basis",
"detect",
"molecular"
],
[
"quality",
"calibration",
"benchmark",
"highlights",
"respect",
"remains",
"patches",
"imagenet",
"uniform",
"treat"
],
[
"imitation",
"expect",
"amenable",
"promise",
"proposes",
"sketch",
"than",
"adopt",
"adopted",
"adopting"
],
[
"robot",
"poses",
"joint",
"robust",
"robots",
"robotics",
"robotic",
"pose",
"hand",
"capability"
],
[
"federated",
"distributed",
"centralized",
"cooperative",
"shared",
"communicate",
"collaborative",
"communications",
"communication",
"protocol"
],
[
"intelligence",
"methodology",
"discriminative",
"solver",
"ai",
"ranking",
"algorithmic",
"controlling",
"controller",
"adaptively"
],
[
"kernel",
"kernels",
"covariates",
"parametric",
"multivariate",
"processes",
"manifold",
"gaussian",
"orthogonal",
"spatial"
],
[
"assessing",
"assessment",
"analyse",
"contrastive",
"mri",
"examining",
"evaluations",
"examine",
"analyze",
"insight"
],
[
"risk",
"covariates",
"vary",
"flexibility",
"risks",
"potential",
"width",
"scalability",
"dependence",
"specialized"
],
[
"attacks",
"attacker",
"adversarial",
"exploits",
"adversaries",
"adversarially",
"security",
"adversary",
"exploit",
"exploiting"
],
[
"compression",
"prediction",
"predicting",
"compress",
"predicts",
"embedding",
"projected",
"predict",
"predictions",
"predicted"
],
[
"fields",
"grid",
"represents",
"convexity",
"image",
"convex",
"naive",
"pattern",
"subsequently",
"trend"
],
[
"modeled",
"structure",
"models",
"modelling",
"modeling",
"model",
"structures",
"computationally",
"formulate",
"profiles"
],
[
"activity",
"sensing",
"sensor",
"sensors",
"wearable",
"running",
"monitoring",
"physical",
"device",
"devices"
],
[
"designs",
"manner",
"manipulation",
"proposes",
"imitation",
"designing",
"art",
"fashion",
"style",
"formalism"
],
[
"theory",
"facial",
"practice",
"faces",
"shifts",
"phase",
"learnt",
"training",
"testing",
"tests"
]
] | 347.27739 | all-MiniLM-L6-v2 | 0.898 | -0.338495 | 0.200578 | 0.902082 |
ArXiv ML Papers | S³ | 46 | 50 | [
[
"exploits",
"attacker",
"attacks",
"adversarial",
"adversaries",
"adversarially",
"security",
"exploit",
"adversary",
"exploiting"
],
[
"hyperparameter",
"hyperparameters",
"optimisation",
"optimizes",
"projected",
"parameterized",
"hyper",
"optimizer",
"scores",
"projections"
],
[
"mixture",
"partial",
"mini",
"minimax",
"ensembles",
"aspects",
"composed",
"aspect",
"combining",
"topics"
],
[
"labeling",
"label",
"labels",
"labeled",
"taxonomy",
"unlabeled",
"classify",
"classification",
"categories",
"classified"
],
[
"flows",
"superior",
"downstream",
"pipelines",
"centralized",
"flow",
"pipeline",
"diffusion",
"causal",
"filtering"
],
[
"estimating",
"rank",
"population",
"ranking",
"approximating",
"adaptive",
"similarity",
"adaptively",
"ai",
"solver"
],
[
"effect",
"variance",
"perturbation",
"popularity",
"sparsity",
"perturbations",
"perturbed",
"effects",
"distortion",
"impacts"
],
[
"clusters",
"segmentation",
"cluster",
"clustering",
"segment",
"separation",
"counterparts",
"regions",
"partitioning",
"unsupervised"
],
[
"fail",
"down",
"completion",
"descent",
"reflect",
"efforts",
"targets",
"acceleration",
"impacts",
"poisoning"
],
[
"anomalies",
"outlier",
"iot",
"anomaly",
"outliers",
"deviation",
"event",
"particle",
"sensor",
"intrusion"
],
[
"medical",
"patients",
"diseases",
"clinical",
"healthcare",
"medicine",
"patient",
"diagnosis",
"practitioners",
"disease"
],
[
"interpretable",
"interpretation",
"interpret",
"interpreted",
"explaining",
"explanations",
"credit",
"interpretability",
"meaningful",
"understand"
],
[
"benefits",
"embedding",
"separately",
"embeddings",
"densities",
"estimator",
"incorporate",
"query",
"capacity",
"multivariate"
],
[
"reduced",
"minimal",
"pruned",
"distinct",
"pruning",
"trivial",
"redundant",
"randomized",
"reduction",
"compressed"
],
[
"constrained",
"risks",
"risk",
"constraint",
"penalty",
"constraints",
"constrain",
"ensures",
"allowed",
"restricted"
],
[
"student",
"learns",
"studying",
"study",
"classes",
"learner",
"teacher",
"learners",
"studied",
"class"
],
[
"tests",
"testing",
"patches",
"test",
"experimental",
"screening",
"patch",
"tested",
"experiments",
"development"
],
[
"compressed",
"tractable",
"threshold",
"severe",
"restricted",
"sparse",
"periods",
"dramatically",
"during",
"minimal"
],
[
"feature",
"aggregated",
"up",
"boosting",
"collected",
"collect",
"ensemble",
"boost",
"ensembles",
"aggregating"
],
[
"personalized",
"recommendations",
"factorization",
"suggests",
"recommender",
"recommendation",
"collaborative",
"ubiquitous",
"preferences",
"ranking"
],
[
"uncertainties",
"reliable",
"reliability",
"uncertain",
"uncertainty",
"validation",
"validate",
"accurately",
"accuracy",
"accuracies"
],
[
"filtering",
"numerical",
"values",
"numerically",
"numbers",
"filter",
"statistics",
"filters",
"accelerate",
"parameters"
],
[
"bandit",
"exploitation",
"seek",
"bandits",
"exploration",
"market",
"price",
"acquisition",
"acquire",
"searching"
],
[
"focusing",
"smoothing",
"smoothness",
"robustness",
"robust",
"noisy",
"focus",
"focused",
"smooth",
"consistent"
],
[
"invariant",
"stability",
"systems",
"transformations",
"preserving",
"stable",
"invariance",
"balance",
"transforms",
"mathematical"
],
[
"guarantees",
"markov",
"observing",
"outcomes",
"bayes",
"randomized",
"probabilistic",
"coverage",
"ensuring",
"reliability"
],
[
"forecasting",
"weather",
"landscape",
"forecast",
"forecasts",
"reflect",
"spectrum",
"forest",
"nature",
"spectral"
],
[
"subspaces",
"matrix",
"matrices",
"linear",
"subspace",
"grid",
"describes",
"randomized",
"representation",
"basis"
],
[
"continuously",
"recurrent",
"characterized",
"termed",
"dynamically",
"defines",
"defined",
"monte",
"initialization",
"compact"
],
[
"difficulty",
"dnn",
"challenging",
"forgetting",
"precision",
"reliability",
"difficulties",
"challenges",
"dnns",
"variability"
],
[
"designs",
"designing",
"variation",
"manipulation",
"variations",
"expressive",
"coding",
"spatiotemporal",
"effects",
"fashion"
],
[
"texts",
"textual",
"nlp",
"linguistic",
"corpora",
"corpus",
"sentences",
"text",
"semantic",
"literature"
],
[
"twitter",
"goals",
"goal",
"video",
"https",
"social",
"principle",
"published",
"publicly",
"influence"
],
[
"models",
"capture",
"quality",
"rate",
"respect",
"capturing",
"remains",
"treat",
"calibration",
"highlights"
],
[
"granularity",
"similarity",
"distance",
"distances",
"molecular",
"metrics",
"metric",
"meaningful",
"comparable",
"space"
],
[
"dimensional",
"dimensionality",
"dimension",
"representations",
"allocation",
"representation",
"capacity",
"subspaces",
"interactive",
"selects"
],
[
"discrimination",
"fairness",
"bias",
"biases",
"discriminate",
"representative",
"biased",
"outcomes",
"fair",
"equal"
],
[
"involve",
"causal",
"interventions",
"counterfactual",
"consequence",
"impacts",
"aid",
"effects",
"dependence",
"interactions"
],
[
"robot",
"robotics",
"robots",
"robotic",
"pose",
"poses",
"machines",
"robust",
"joint",
"objects"
],
[
"software",
"multimodal",
"contextual",
"tend",
"spatio",
"corpus",
"vanilla",
"raw",
"spatiotemporal",
"empirical"
],
[
"bounded",
"sampling",
"limitations",
"estimate",
"asymptotically",
"complexity",
"arbitrarily",
"bounds",
"bound",
"extent"
],
[
"model",
"candidate",
"latency",
"fitting",
"models",
"correlated",
"joint",
"jointly",
"availability",
"having"
],
[
"classifying",
"generalizing",
"generalization",
"distinguishing",
"softmax",
"randomly",
"classifiers",
"ensemble",
"ensembles",
"generalize"
],
[
"width",
"oriented",
"vary",
"wide",
"dependence",
"scalability",
"covariates",
"specialized",
"variances",
"potential"
],
[
"voice",
"acoustic",
"speaker",
"audio",
"sound",
"speech",
"music",
"noise",
"frequencies",
"tune"
],
[
"producing",
"generates",
"generate",
"gans",
"gan",
"generating",
"generation",
"generative",
"generator",
"generated"
],
[
"markov",
"simulations",
"simulation",
"stochastic",
"sampling",
"monte",
"randomly",
"mcmc",
"carlo",
"probabilistic"
],
[
"rnns",
"backpropagation",
"neighboring",
"rnn",
"nearest",
"predict",
"directly",
"tilde",
"behind",
"predicts"
],
[
"driving",
"transportation",
"traffic",
"navigation",
"gps",
"transport",
"vehicles",
"vehicle",
"trajectories",
"paths"
],
[
"recurrent",
"forecasting",
"forecast",
"forecasts",
"sequences",
"temporal",
"sequential",
"autoregressive",
"demand",
"shifts"
]
] | 357.870167 | all-MiniLM-L6-v2 | 0.922 | -0.335476 | 0.191429 | 0.913816 |
ArXiv ML Papers | S³_angular | 43 | 10 | [
[
"overfitting",
"outperforms",
"outperform",
"outperformed",
"outperforming",
"accuracy",
"pretrained",
"tuning",
"worst",
"optimizer"
],
[
"throughput",
"bottleneck",
"speedup",
"crucial",
"inefficient",
"benchmarks",
"faster",
"scalable",
"streaming",
"pipeline"
],
[
"minimax",
"bandits",
"regularization",
"lasso",
"leveraging",
"optimal",
"minimize",
"regularized",
"sparse",
"softmax"
],
[
"softmax",
"forecasts",
"forecasting",
"forecast",
"prediction",
"predicting",
"predicts",
"predicted",
"predict",
"predictions"
],
[
"convex",
"discover",
"discovery",
"datasets",
"lasso",
"dataset",
"discovering",
"shape",
"optimization",
"hyperparameter"
],
[
"completion",
"iterative",
"iteratively",
"guide",
"includes",
"rapid",
"descent",
"pretrained",
"selecting",
"recommendation"
],
[
"outliers",
"outlier",
"overfitting",
"dataset",
"fitting",
"datasets",
"prevalence",
"compression",
"normalized",
"diagnostic"
],
[
"recommendations",
"suggest",
"recommendation",
"known",
"commonly",
"widely",
"recommender",
"suggesting",
"nearly",
"estimating"
],
[
"annotated",
"datasets",
"predicted",
"annotations",
"dataset",
"guidance",
"predict",
"helpful",
"predictive",
"annotation"
],
[
"extensively",
"inference",
"generative",
"gans",
"densities",
"likelihood",
"mcmc",
"prior",
"parameterized",
"priors"
]
] | 357.042876 | all-MiniLM-L6-v2 | 0.89 | -0.345606 | 0.210934 | 0.880243 |
ArXiv ML Papers | S³_angular | 44 | 10 | [
[
"parameterized",
"forecast",
"forecasting",
"generative",
"forecasts",
"models",
"rnns",
"likelihood",
"autoregressive",
"modelling"
],
[
"directed",
"helpful",
"several",
"answers",
"tools",
"reports",
"many",
"far",
"scenarios",
"guided"
],
[
"learns",
"classifiers",
"prediction",
"predicting",
"predicted",
"predict",
"classifier",
"predicts",
"predictive",
"learnt"
],
[
"texts",
"textual",
"corpus",
"sentences",
"languages",
"speech",
"linguistic",
"corpora",
"nlp",
"sentence"
],
[
"bandit",
"learnt",
"strategy",
"rewards",
"learns",
"strategies",
"reinforcement",
"learned",
"reward",
"learn"
],
[
"widely",
"hyperparameters",
"bandits",
"known",
"commonly",
"recommender",
"estimating",
"recommendations",
"hyperparameter",
"recent"
],
[
"leveraging",
"labeled",
"broadly",
"labeling",
"retrieval",
"deep",
"regularized",
"supervised",
"bias",
"bandits"
],
[
"attacker",
"adversaries",
"adversarial",
"adversarially",
"adversary",
"malicious",
"vulnerability",
"exploits",
"attacks",
"mitigate"
],
[
"speedup",
"faster",
"inefficient",
"benchmarks",
"crucial",
"throughput",
"bottleneck",
"scalable",
"pipeline",
"streaming"
],
[
"graphs",
"networks",
"vertex",
"graph",
"vertices",
"node",
"nodes",
"communities",
"links",
"neighbors"
]
] | 273.656424 | all-MiniLM-L6-v2 | 0.97 | -0.283362 | 0.189928 | 0.872499 |
ArXiv ML Papers | S³_angular | 45 | 10 | [
[
"rnn",
"trained",
"atari",
"learns",
"learnable",
"learnt",
"rnns",
"learned",
"strategy",
"future"
],
[
"classify",
"predictors",
"classifiers",
"classifier",
"supervised",
"classifying",
"classification",
"predicting",
"prediction",
"predictive"
],
[
"databases",
"feed",
"boost",
"cifar10",
"core",
"xgboost",
"pipeline",
"vast",
"database",
"report"
],
[
"selecting",
"known",
"recommendations",
"estimating",
"recommendation",
"recommender",
"hyperparameter",
"hyperparameters",
"selects",
"ensembles"
],
[
"biology",
"physics",
"manifold",
"robotics",
"object",
"fundamental",
"objects",
"shape",
"robotic",
"biological"
],
[
"learned",
"trained",
"annotations",
"annotation",
"learners",
"encourages",
"annotated",
"labeled",
"superiority",
"reward"
],
[
"thresholding",
"convexity",
"convex",
"pruning",
"norm",
"regularized",
"boosting",
"overfitting",
"regularization",
"threshold"
],
[
"pipelines",
"gan",
"covid",
"believe",
"coverage",
"tool",
"agnostic",
"gans",
"pipeline",
"tools"
],
[
"types",
"errors",
"empirically",
"interpretable",
"weak",
"interpretability",
"noisy",
"kinds",
"authors",
"noise"
],
[
"nodes",
"graphs",
"vertex",
"networks",
"graph",
"communities",
"neighbors",
"node",
"vertices",
"links"
]
] | 323.570477 | all-MiniLM-L6-v2 | 0.97 | -0.334025 | 0.15643 | 0.886956 |
ArXiv ML Papers | S³_angular | 46 | 10 | [
[
"bandits",
"commonly",
"widely",
"known",
"hyperparameters",
"recommendations",
"ranking",
"recommender",
"nearly",
"estimating"
],
[
"shape",
"supervised",
"lasso",
"discover",
"datasets",
"dataset",
"discovering",
"discovery",
"hyperparameter",
"convex"
],
[
"iterative",
"includes",
"guide",
"completion",
"rapid",
"iteratively",
"descent",
"selecting",
"pretrained",
"recommendation"
],
[
"speedup",
"throughput",
"streaming",
"pipeline",
"scalable",
"inefficient",
"benchmarks",
"faster",
"bottleneck",
"crucial"
],
[
"reinforcement",
"bandit",
"learnt",
"learns",
"learned",
"learn",
"reward",
"critic",
"strategies",
"strategy"
],
[
"generative",
"priors",
"likelihood",
"prior",
"parameterized",
"inference",
"models",
"bayesian",
"mcmc",
"modelling"
],
[
"annotated",
"annotations",
"leveraging",
"annotation",
"guided",
"helpful",
"datasets",
"labeled",
"supervised",
"guidance"
],
[
"machines",
"machine",
"developing",
"future",
"integrate",
"integrating",
"development",
"integration",
"integrates",
"expert"
],
[
"scene",
"videos",
"landscape",
"proposals",
"depth",
"scenes",
"captures",
"viewed",
"capturing",
"capture"
],
[
"graphs",
"nodes",
"vertex",
"networks",
"graph",
"communities",
"vertices",
"links",
"node",
"neighbors"
]
] | 331.378568 | all-MiniLM-L6-v2 | 0.98 | -0.324628 | 0.162727 | 0.855496 |
ArXiv ML Papers | S³_angular | 43 | 20 | [
[
"overfitting",
"outlier",
"outliers",
"fitting",
"cnns",
"asymptotically",
"nonparametric",
"statistically",
"smaller",
"quantify"
],
[
"distillation",
"tuning",
"learned",
"ensembles",
"learns",
"ensemble",
"resonance",
"learnt",
"leveraging",
"learning"
],
[
"adversary",
"adversarial",
"attacks",
"attacker",
"vulnerability",
"exploits",
"threat",
"adversaries",
"adversarially",
"malicious"
],
[
"clustering",
"cluster",
"svm",
"boosting",
"randomized",
"mining",
"batch",
"dirichlet",
"boost",
"clusters"
],
[
"builds",
"gpu",
"throughput",
"inefficient",
"faster",
"accelerators",
"bottleneck",
"gpus",
"fast",
"efficiently"
],
[
"greedy",
"fewer",
"regret",
"seek",
"acquire",
"bandits",
"guaranteed",
"conditioning",
"expectation",
"most"
],
[
"specific",
"recall",
"mini",
"feature",
"auxiliary",
"head",
"boosting",
"memory",
"already",
"cover"
],
[
"annotated",
"learns",
"leveraging",
"notably",
"learned",
"attention",
"forgetting",
"recall",
"retaining",
"aware"
],
[
"records",
"log",
"epsilon",
"logistic",
"minutes",
"metrics",
"lstm",
"length",
"softmax",
"measure"
],
[
"stochastic",
"likelihood",
"mcmc",
"carlo",
"priors",
"bayesian",
"probabilistic",
"densities",
"dirichlet",
"sampling"
],
[
"known",
"preprocessing",
"automated",
"frameworks",
"tools",
"comprehensive",
"meet",
"near",
"developments",
"here"
],
[
"sgd",
"convexity",
"hessian",
"minimax",
"convex",
"convergence",
"ascent",
"gradient",
"gradients",
"converges"
],
[
"highest",
"outperformed",
"hyperparameter",
"nearest",
"hyperparameters",
"minimum",
"bayesian",
"maximize",
"parameterized",
"minimize"
],
[
"hyperparameters",
"hyperparameter",
"potential",
"parametric",
"length",
"position",
"magnitude",
"depth",
"parameters",
"coordinates"
],
[
"linguistic",
"semantic",
"sentences",
"sentence",
"language",
"texts",
"corpus",
"textual",
"corpora",
"nlp"
],
[
"supervised",
"quantization",
"recognition",
"recognizing",
"discriminator",
"contained",
"deep",
"recognize",
"embeddings",
"labeled"
],
[
"cancer",
"medicine",
"clinical",
"healthcare",
"patients",
"diseases",
"medical",
"health",
"disease",
"aid"
],
[
"factorization",
"tensor",
"sparse",
"tensors",
"factor",
"matrix",
"matrices",
"lasso",
"besides",
"subspace"
],
[
"users",
"gender",
"appealing",
"discrimination",
"profiles",
"fairness",
"profile",
"candidate",
"personalized",
"specify"
],
[
"generalizing",
"suggesting",
"views",
"generalization",
"generalize",
"intuitive",
"hierarchical",
"generalizes",
"generalized",
"discovering"
]
] | 272.490411 | all-MiniLM-L6-v2 | 0.95 | -0.309468 | 0.174068 | 0.909696 |
ArXiv ML Papers | S³_angular | 44 | 20 | [
[
"ensemble",
"sensing",
"selected",
"trained",
"selection",
"chosen",
"dimensionality",
"training",
"settings",
"dataset"
],
[
"driving",
"tracking",
"gps",
"track",
"ranging",
"outside",
"vehicle",
"vehicles",
"camera",
"drive"
],
[
"hyperparameters",
"hyperparameter",
"parametric",
"potential",
"length",
"parameters",
"position",
"coordinates",
"depth",
"magnitude"
],
[
"autoencoders",
"modeling",
"nonlinear",
"manifold",
"modelling",
"modeled",
"differential",
"expressive",
"differentially",
"differentiable"
],
[
"records",
"log",
"epsilon",
"logistic",
"lstm",
"metrics",
"softmax",
"minutes",
"length",
"measure"
],
[
"tools",
"automated",
"comprehensive",
"preprocessing",
"known",
"frameworks",
"near",
"meet",
"developments",
"here"
],
[
"overfitting",
"nonparametric",
"asymptotically",
"outliers",
"outlier",
"fitting",
"smaller",
"cnns",
"statistically",
"quantify"
],
[
"accuracy",
"machines",
"trees",
"deterministic",
"outperform",
"classes",
"class",
"trainable",
"outperforms",
"trained"
],
[
"adversarial",
"vulnerability",
"adversary",
"malicious",
"attacker",
"exploits",
"threat",
"attacks",
"adversarially",
"adversaries"
],
[
"discrimination",
"profiles",
"appealing",
"users",
"fairness",
"profile",
"candidate",
"gender",
"personalized",
"specify"
],
[
"weather",
"forecast",
"predicting",
"forecasts",
"predict",
"forecasting",
"predicted",
"prediction",
"autoregressive",
"predictions"
],
[
"vertex",
"graph",
"graphs",
"nodes",
"vertices",
"networks",
"node",
"edges",
"neighbors",
"edge"
],
[
"dimensional",
"dimensionality",
"embeddings",
"approximating",
"approximations",
"approximates",
"learnable",
"embedding",
"approximated",
"person"
],
[
"guarantees",
"estimate",
"capability",
"projections",
"estimates",
"projected",
"limitations",
"depth",
"uncertainties",
"capabilities"
],
[
"gradient",
"minimax",
"convex",
"convexity",
"convergence",
"ascent",
"sgd",
"hessian",
"gradients",
"converges"
],
[
"wise",
"symmetric",
"overfitting",
"linear",
"naive",
"superiority",
"trained",
"vectors",
"euclidean",
"training"
],
[
"minimize",
"nearest",
"highest",
"bayesian",
"parameterized",
"hyperparameter",
"outperformed",
"hyperparameters",
"maximize",
"minimum"
],
[
"attention",
"forgetting",
"learned",
"learns",
"recall",
"annotated",
"leveraging",
"aware",
"retaining",
"notably"
],
[
"bottleneck",
"accelerators",
"builds",
"faster",
"gpu",
"inefficient",
"gpus",
"throughput",
"fast",
"build"
],
[
"audio",
"voice",
"acoustic",
"sound",
"frequencies",
"record",
"speaker",
"recorded",
"music",
"performing"
]
] | 243.913567 | all-MiniLM-L6-v2 | 0.955 | -0.322826 | 0.189364 | 0.895932 |
ArXiv ML Papers | S³_angular | 45 | 20 | [
[
"selection",
"chosen",
"selected",
"sensing",
"settings",
"ensemble",
"trained",
"dimensionality",
"dimensional",
"training"
],
[
"camera",
"tracking",
"gps",
"driving",
"vehicles",
"vehicle",
"ranging",
"track",
"drive",
"videos"
],
[
"fitting",
"outlier",
"outperforming",
"nonparametric",
"asymptotically",
"statistically",
"overfitting",
"outliers",
"smaller",
"cnns"
],
[
"insights",
"guarantees",
"limitations",
"capability",
"estimates",
"relying",
"confidence",
"projected",
"imputation",
"overhead"
],
[
"discriminator",
"supervised",
"recognizing",
"contained",
"labeled",
"recognize",
"recognition",
"quantization",
"distinguishing",
"unconstrained"
],
[
"seek",
"fewer",
"bandits",
"acquire",
"conditioning",
"greedy",
"regret",
"guaranteed",
"expectation",
"most"
],
[
"selects",
"fact",
"information",
"select",
"bayes",
"inference",
"inferring",
"relational",
"knowledge",
"bayesian"
],
[
"potential",
"parametric",
"hyperparameters",
"hyperparameter",
"parameters",
"length",
"position",
"coordinates",
"magnitude",
"attempts"
],
[
"log",
"records",
"lstm",
"minutes",
"epsilon",
"logistic",
"metrics",
"length",
"softmax",
"healthcare"
],
[
"convolutions",
"convolutional",
"layer",
"neurons",
"softmax",
"gaussian",
"backpropagation",
"deep",
"networks",
"neural"
],
[
"factorization",
"tensor",
"matrices",
"matrix",
"factor",
"sparse",
"tensors",
"besides",
"lasso",
"subspace"
],
[
"boosting",
"mini",
"head",
"specific",
"already",
"cover",
"auxiliary",
"feature",
"naive",
"minimizing"
],
[
"views",
"generalization",
"generalizing",
"suggesting",
"generalize",
"discovering",
"ideas",
"hierarchical",
"intuitive",
"generalizes"
],
[
"eeg",
"behavioral",
"behaviors",
"phenomena",
"dynamical",
"energy",
"complexities",
"behaviour",
"brain",
"dynamics"
],
[
"likelihood",
"bayesian",
"distributions",
"mcmc",
"dirichlet",
"probabilistic",
"stochastic",
"densities",
"carlo",
"priors"
],
[
"profiles",
"discrimination",
"appealing",
"personalized",
"fairness",
"profile",
"users",
"recommender",
"gender",
"candidate"
],
[
"vulnerability",
"adversary",
"threat",
"attacks",
"attacker",
"adversarially",
"malicious",
"adversarial",
"exploits",
"adversaries"
],
[
"embeddings",
"dimensional",
"learnable",
"trainable",
"approximates",
"approximating",
"approximations",
"dimensionality",
"embedding",
"dimension"
],
[
"compute",
"computes",
"computation",
"implementation",
"defined",
"algorithms",
"computed",
"implementations",
"computations",
"implementing"
],
[
"crucial",
"accelerators",
"bottleneck",
"throughput",
"faster",
"hardware",
"inefficient",
"gpu",
"builds",
"gpus"
]
] | 327.616405 | all-MiniLM-L6-v2 | 0.975 | -0.318709 | 0.172582 | 0.905875 |
ArXiv ML Papers | S³_angular | 46 | 20 | [
[
"limitations",
"estimates",
"estimate",
"guarantees",
"capability",
"projected",
"uncertainties",
"depth",
"projections",
"capabilities"
],
[
"profiles",
"users",
"fairness",
"discrimination",
"appealing",
"profile",
"gender",
"candidate",
"specify",
"personalized"
],
[
"learned",
"resonance",
"distillation",
"ensemble",
"ensembles",
"leveraging",
"learnt",
"learns",
"learning",
"tuning"
],
[
"factorization",
"matrix",
"matrices",
"tensor",
"sparse",
"factor",
"tensors",
"lasso",
"besides",
"subspace"
],
[
"learns",
"recall",
"annotated",
"leveraging",
"attention",
"notably",
"learned",
"forgetting",
"aware",
"retaining"
],
[
"node",
"nodes",
"vertex",
"edges",
"vertices",
"graph",
"neighbors",
"networks",
"graphs",
"edge"
],
[
"intermediate",
"offline",
"marginal",
"planning",
"completion",
"jointly",
"whole",
"retaining",
"pooling",
"broader"
],
[
"privacy",
"private",
"federated",
"shared",
"sharing",
"distributed",
"decentralized",
"independently",
"cloud",
"external"
],
[
"ensemble",
"trained",
"selected",
"selection",
"sensing",
"chosen",
"training",
"settings",
"dimensionality",
"dataset"
],
[
"convexity",
"sgd",
"minimax",
"hessian",
"ascent",
"gradient",
"convex",
"convergence",
"gradients",
"converges"
],
[
"convolutional",
"neuron",
"deep",
"neural",
"softmax",
"cnn",
"imagenet",
"backpropagation",
"cnns",
"rnns"
],
[
"inefficient",
"throughput",
"faster",
"builds",
"fast",
"gpus",
"gpu",
"accelerators",
"bottleneck",
"build"
],
[
"bandits",
"seek",
"greedy",
"fewer",
"regret",
"most",
"expectation",
"guaranteed",
"conditioning",
"acquire"
],
[
"generative",
"augmentation",
"producing",
"gan",
"gans",
"produces",
"generating",
"adapted",
"multimodal",
"transform"
],
[
"behaviour",
"behavioral",
"brain",
"complexities",
"dynamical",
"energy",
"physical",
"behavior",
"eeg",
"behaviors"
],
[
"likelihood",
"stochastic",
"bayesian",
"probabilistic",
"mcmc",
"densities",
"carlo",
"priors",
"dirichlet",
"sampling"
],
[
"reinforcement",
"controller",
"control",
"play",
"exploration",
"controlling",
"controlled",
"agents",
"action",
"planning"
],
[
"cancer",
"healthcare",
"medicine",
"clinical",
"disease",
"medical",
"patients",
"diseases",
"health",
"aid"
],
[
"quantization",
"contained",
"recognizing",
"discriminator",
"recognition",
"deep",
"supervised",
"labeled",
"recognize",
"embeddings"
],
[
"acoustic",
"sound",
"audio",
"music",
"speaker",
"recorded",
"voice",
"record",
"performing",
"frequencies"
]
] | 317.560862 | all-MiniLM-L6-v2 | 0.965 | -0.27723 | 0.174713 | 0.897798 |
ArXiv ML Papers | S³_angular | 43 | 30 | [
[
"laplacian",
"perturbations",
"perturbation",
"convergence",
"variance",
"converges",
"promising",
"landscape",
"converge",
"nearly"
],
[
"sequential",
"spatiotemporal",
"temporal",
"autoregressive",
"rnns",
"periodic",
"sequence",
"lstm",
"sequences",
"recurrent"
],
[
"task",
"estimator",
"measurement",
"measurements",
"estimators",
"estimate",
"measuring",
"measure",
"sensors",
"estimation"
],
[
"viewed",
"have",
"related",
"views",
"subsequently",
"intuitive",
"similar",
"relevant",
"tend",
"novel"
],
[
"biased",
"equally",
"discriminate",
"representative",
"discrimination",
"candidates",
"equal",
"bias",
"fairness",
"biases"
],
[
"neuron",
"evolving",
"weights",
"logistic",
"networks",
"flow",
"chain",
"nets",
"mcmc",
"neural"
],
[
"deeper",
"pipelines",
"pac",
"deep",
"pipeline",
"forecasts",
"forecast",
"whole",
"stacked",
"horizon"
],
[
"outlier",
"overfitting",
"fitting",
"outliers",
"cnns",
"bottleneck",
"statistically",
"asymptotically",
"deviations",
"outperforming"
],
[
"learns",
"machine",
"supervised",
"leads",
"dataset",
"datasets",
"learners",
"instance",
"batch",
"machines"
],
[
"imagenet",
"recognition",
"images",
"classification",
"svm",
"image",
"background",
"classified",
"recognizing",
"classifying"
],
[
"prior",
"posterior",
"priors",
"gaussian",
"densities",
"bayes",
"likelihood",
"bayesian",
"probabilistic",
"distributions"
],
[
"carlo",
"particle",
"quantized",
"11",
"quantum",
"randomly",
"71",
"10",
"ucb",
"discovered"
],
[
"diseases",
"clinical",
"patients",
"medicine",
"healthcare",
"medical",
"disease",
"health",
"cancer",
"diagnosis"
],
[
"feed",
"solely",
"lasso",
"includes",
"attention",
"alone",
"extended",
"ad",
"highlight",
"first"
],
[
"interpretability",
"interpretable",
"explanations",
"explaining",
"uncertainties",
"rigorously",
"rigorous",
"reliability",
"explainable",
"robustness"
],
[
"classification",
"labels",
"supervised",
"labeled",
"classifying",
"classifiers",
"discriminative",
"classifier",
"labeling",
"label"
],
[
"forecast",
"forecasts",
"forecasting",
"weather",
"vehicle",
"traffic",
"gps",
"transportation",
"autoregressive",
"prediction"
],
[
"role",
"representations",
"underlying",
"tensor",
"tensors",
"respectively",
"vectors",
"each",
"vector",
"multivariate"
],
[
"privacy",
"private",
"federated",
"accessible",
"public",
"publicly",
"alone",
"sharing",
"addresses",
"access"
],
[
"predictor",
"hand",
"have",
"information",
"details",
"kernel",
"armed",
"later",
"learn",
"predictors"
],
[
"combines",
"learners",
"fusion",
"learnt",
"aware",
"learns",
"stream",
"maintain",
"coarse",
"learner"
],
[
"search",
"bound",
"retrieval",
"seek",
"bandits",
"expectation",
"bandit",
"searching",
"randomly",
"sampling"
],
[
"intermediate",
"completion",
"consist",
"termed",
"dedicated",
"labels",
"projected",
"compress",
"interpolation",
"relative"
],
[
"classes",
"class",
"phases",
"phase",
"arises",
"examples",
"training",
"learning",
"abstract",
"inducing"
],
[
"gan",
"generative",
"generating",
"gans",
"generator",
"generate",
"autoencoders",
"generates",
"generated",
"autoencoder"
],
[
"making",
"projects",
"project",
"produced",
"tools",
"developers",
"creating",
"visualization",
"developments",
"producing"
],
[
"tuning",
"predicting",
"filters",
"track",
"music",
"sensing",
"prediction",
"ensemble",
"ensembles",
"tracking"
],
[
"partial",
"field",
"mixture",
"head",
"normalizing",
"varepsilon",
"variant",
"horizon",
"additive",
"specific"
],
[
"hessian",
"minimize",
"sgd",
"convex",
"convexity",
"ascent",
"minimax",
"gradient",
"minimization",
"descent"
],
[
"tracking",
"camera",
"track",
"ranging",
"pose",
"poses",
"3d",
"gps",
"points",
"capturing"
]
] | 351.27492 | all-MiniLM-L6-v2 | 0.943333 | -0.307117 | 0.172498 | 0.88861 |
ArXiv ML Papers | S³_angular | 44 | 30 | [
[
"encodes",
"associated",
"encode",
"encoded",
"molecular",
"protein",
"encoding",
"encoder",
"domains",
"quickly"
],
[
"longer",
"https",
"hand",
"long",
"dependent",
"predictor",
"via",
"far",
"future",
"internet"
],
[
"bounds",
"approximations",
"approximating",
"approximates",
"threshold",
"especially",
"approximation",
"arbitrarily",
"function",
"depth"
],
[
"partial",
"specific",
"variant",
"field",
"head",
"mixture",
"minimizing",
"normalizing",
"mini",
"scenarios"
],
[
"tracking",
"vision",
"pose",
"poses",
"scenes",
"view",
"assist",
"ranging",
"moving",
"perspective"
],
[
"gaining",
"subsequent",
"nets",
"towards",
"toward",
"chain",
"sequential",
"tend",
"order",
"forward"
],
[
"public",
"private",
"publicly",
"privacy",
"federated",
"accessible",
"alone",
"hidden",
"decentralized",
"access"
],
[
"quantum",
"datasets",
"dataset",
"states",
"infer",
"collection",
"discovery",
"inferring",
"discovering",
"selects"
],
[
"constraint",
"constraints",
"logic",
"relational",
"conditional",
"inference",
"constrained",
"inputs",
"unconstrained",
"reasoning"
],
[
"healthcare",
"clinical",
"medical",
"diagnosis",
"patients",
"disease",
"diseases",
"medicine",
"health",
"cancer"
],
[
"track",
"ensembles",
"music",
"ensemble",
"sensing",
"prediction",
"filters",
"tuning",
"tune",
"tracking"
],
[
"bottleneck",
"overfitting",
"outliers",
"fitting",
"cnns",
"convolutions",
"throughput",
"outlier",
"convolution",
"asymptotically"
],
[
"classification",
"imagenet",
"classifier",
"recognition",
"classifying",
"svm",
"classified",
"classifiers",
"images",
"classify"
],
[
"tree",
"outperform",
"forest",
"classifier",
"outperforming",
"outperformed",
"boosting",
"outperforms",
"classifiers",
"validated"
],
[
"gan",
"gans",
"generative",
"generation",
"generate",
"generating",
"generates",
"generator",
"autoencoders",
"generated"
],
[
"differentiation",
"differential",
"backpropagation",
"nonlinear",
"differentiable",
"hessian",
"gradient",
"derivatives",
"perturbed",
"coordinate"
],
[
"city",
"iot",
"transportation",
"planning",
"urban",
"traffic",
"area",
"driving",
"drive",
"neighborhood"
],
[
"wearable",
"person",
"kernel",
"eeg",
"devices",
"activity",
"device",
"form",
"human",
"identified"
],
[
"likelihood",
"bayesian",
"densities",
"mcmc",
"posterior",
"bayes",
"distributional",
"probabilistic",
"distributions",
"mixture"
],
[
"attacker",
"vulnerability",
"adversarial",
"exploits",
"adversary",
"exploit",
"security",
"intrusion",
"attacks",
"adversarially"
],
[
"supervised",
"labeling",
"label",
"labeled",
"unlabeled",
"discriminative",
"classification",
"labels",
"learning",
"classifiers"
],
[
"projects",
"producing",
"creating",
"developments",
"produced",
"developers",
"making",
"created",
"project",
"tools"
],
[
"metrics",
"metric",
"distances",
"similarity",
"corresponds",
"dimension",
"distance",
"dimensionality",
"dimensional",
"euclidean"
],
[
"lstm",
"sequences",
"sequence",
"events",
"occurrence",
"rnns",
"recurrent",
"records",
"record",
"temporal"
],
[
"subsequently",
"related",
"similar",
"viewed",
"massive",
"tend",
"relevant",
"rapidly",
"have",
"seen"
],
[
"iteration",
"completion",
"forgetting",
"iteratively",
"iterative",
"iterations",
"navigation",
"discover",
"autoencoder",
"expansion"
],
[
"predicted",
"forecasting",
"forecast",
"general",
"forecasts",
"weather",
"predict",
"predicting",
"prediction",
"predictions"
],
[
"intuition",
"intuitive",
"generalize",
"generalizing",
"explanations",
"brain",
"cognitive",
"neural",
"generalization",
"describe"
],
[
"generalizes",
"datasets",
"supervised",
"generalization",
"learns",
"generalize",
"dataset",
"ai",
"machine",
"hyperparameters"
],
[
"inducing",
"class",
"out",
"phases",
"classes",
"phase",
"learning",
"arises",
"examples",
"training"
]
] | 340.881999 | all-MiniLM-L6-v2 | 0.953333 | -0.283111 | 0.193428 | 0.894101 |
ArXiv ML Papers | S³_angular | 45 | 30 | [
[
"contained",
"labeled",
"unconstrained",
"structured",
"labeling",
"unlabeled",
"tight",
"annotation",
"annotations",
"supervised"
],
[
"sequences",
"sequential",
"sequence",
"recurrent",
"rnns",
"minimax",
"lstm",
"softmax",
"attention",
"rnn"
],
[
"bound",
"bandits",
"coverage",
"retrieval",
"searching",
"seek",
"expectation",
"bandit",
"considerable",
"search"
],
[
"discrimination",
"fairness",
"bias",
"biases",
"discriminate",
"representative",
"equally",
"equal",
"candidate",
"candidates"
],
[
"classifying",
"classifiers",
"supervised",
"labeled",
"regularization",
"classifier",
"classification",
"discriminative",
"unlabeled",
"backpropagation"
],
[
"armed",
"boosting",
"recall",
"forest",
"covid",
"identified",
"represent",
"identify",
"raw",
"identifying"
],
[
"downstream",
"adaptation",
"stream",
"transferring",
"across",
"transfer",
"relu",
"transferred",
"leveraging",
"incorporating"
],
[
"area",
"urban",
"iot",
"planning",
"city",
"bounded",
"car",
"drive",
"driving",
"neighborhood"
],
[
"massive",
"similar",
"related",
"previously",
"have",
"subsequently",
"descent",
"same",
"viewed",
"tend"
],
[
"energy",
"generator",
"distillation",
"generates",
"mining",
"generating",
"compressed",
"magnetic",
"resonance",
"imbalanced"
],
[
"records",
"history",
"historical",
"measure",
"occurrence",
"temporal",
"recurrent",
"periodic",
"record",
"events"
],
[
"field",
"mixture",
"head",
"mini",
"partial",
"specific",
"variant",
"minimizing",
"minimax",
"normalizing"
],
[
"distances",
"dimensionality",
"dimensional",
"embeddings",
"metrics",
"metric",
"similarity",
"distance",
"embedding",
"dimension"
],
[
"decentralized",
"federated",
"distributed",
"private",
"privacy",
"accessible",
"external",
"sharing",
"shared",
"independently"
],
[
"machine",
"machines",
"samples",
"instance",
"dataset",
"supervised",
"datasets",
"learners",
"learning",
"learns"
],
[
"gan",
"generative",
"generating",
"generate",
"gans",
"generator",
"generated",
"generates",
"generation",
"autoencoders"
],
[
"training",
"grounded",
"cnn",
"assessment",
"gnn",
"trained",
"augment",
"relu",
"reliability",
"dnn"
],
[
"unsupervised",
"partitioning",
"cluster",
"clustering",
"partition",
"group",
"clusters",
"instances",
"groups",
"regions"
],
[
"static",
"invariant",
"invariance",
"dependencies",
"efficient",
"components",
"quantization",
"aspect",
"symmetric",
"person"
],
[
"flow",
"weights",
"transition",
"evolving",
"mcmc",
"classify",
"branch",
"logistic",
"classifier",
"divergence"
],
[
"fitting",
"bottleneck",
"cnns",
"outlier",
"outliers",
"overfitting",
"asymptotically",
"statistically",
"approximations",
"quantify"
],
[
"imagenet",
"background",
"recognition",
"image",
"svm",
"images",
"convolutional",
"convolution",
"classification",
"convolutions"
],
[
"deep",
"deeper",
"horizon",
"pipeline",
"pipelines",
"dense",
"depth",
"pac",
"projected",
"generalizing"
],
[
"quantum",
"datasets",
"infer",
"inferring",
"discovery",
"states",
"dataset",
"collection",
"selects",
"determined"
],
[
"interpolation",
"dedicated",
"completion",
"optical",
"intermediate",
"compress",
"consist",
"termed",
"projected",
"retaining"
],
[
"convergence",
"converge",
"converges",
"perturbation",
"perturbations",
"laplacian",
"vary",
"near",
"promising",
"variance"
],
[
"explanations",
"cognitive",
"intuitive",
"intuition",
"brain",
"generalization",
"neural",
"generalizing",
"generalize",
"abstraction"
],
[
"variational",
"likelihood",
"priors",
"distributional",
"posterior",
"bayesian",
"densities",
"gaussian",
"distributions",
"mixture"
],
[
"classes",
"class",
"inducing",
"training",
"out",
"learning",
"phase",
"phases",
"learnt",
"arises"
],
[
"prediction",
"predicting",
"bottleneck",
"optimizing",
"selection",
"ensemble",
"ensembles",
"tuning",
"boosting",
"sensing"
]
] | 313.46132 | all-MiniLM-L6-v2 | 0.93 | -0.306811 | 0.165869 | 0.883074 |
ArXiv ML Papers | S³_angular | 46 | 30 | [
[
"iteration",
"incremental",
"iterations",
"forgetting",
"iterative",
"iteratively",
"completion",
"discover",
"navigation",
"learns"
],
[
"stronger",
"configuration",
"enhance",
"capable",
"superiority",
"deployment",
"confirm",
"configurations",
"enhances",
"guidance"
],
[
"convergence",
"converge",
"converges",
"perturbation",
"promising",
"perturbations",
"laplacian",
"near",
"variance",
"nearly"
],
[
"rewards",
"play",
"reward",
"reinforcement",
"games",
"game",
"exploration",
"controller",
"plays",
"strategies"
],
[
"textual",
"texts",
"english",
"bert",
"language",
"corpora",
"nlp",
"sentences",
"linguistic",
"sentence"
],
[
"layer",
"parametric",
"functions",
"parameters",
"theorem",
"parameter",
"hyperparameters",
"atari",
"hyperparameter",
"optimum"
],
[
"convolutions",
"recognition",
"allowed",
"convolution",
"images",
"imagenet",
"image",
"svm",
"convolutional",
"background"
],
[
"subspaces",
"inverse",
"sparse",
"solving",
"solves",
"cover",
"resolution",
"subspace",
"matrices",
"denoising"
],
[
"discriminative",
"labeled",
"classifiers",
"regularization",
"backpropagation",
"classifier",
"supervised",
"classification",
"classifying",
"unlabeled"
],
[
"generative",
"gan",
"generating",
"gans",
"generated",
"generate",
"generator",
"autoencoders",
"generates",
"generation"
],
[
"vertex",
"graph",
"graphs",
"vertices",
"nodes",
"edges",
"node",
"edge",
"networks",
"neighbors"
],
[
"dedicated",
"completion",
"intermediate",
"termed",
"compress",
"consist",
"interpolation",
"retaining",
"optical",
"projected"
],
[
"hand",
"information",
"longer",
"predictor",
"such",
"later",
"https",
"long",
"predictors",
"kernel"
],
[
"qualitative",
"monitoring",
"reliability",
"fail",
"learners",
"uncertainty",
"uncertainties",
"expert",
"validity",
"uncertain"
],
[
"machine",
"datasets",
"dataset",
"supervised",
"instance",
"batch",
"learns",
"learning",
"training",
"learners"
],
[
"gaussian",
"variational",
"densities",
"priors",
"distributional",
"likelihood",
"distributions",
"bayesian",
"posterior",
"mixture"
],
[
"person",
"wearable",
"eeg",
"devices",
"device",
"activity",
"human",
"kernel",
"identify",
"identified"
],
[
"acoustic",
"voice",
"speaker",
"audio",
"sound",
"music",
"speech",
"tune",
"spectrum",
"noise"
],
[
"convexity",
"gradient",
"minimax",
"ascent",
"convex",
"optimization",
"minimize",
"hessian",
"minimization",
"gradients"
],
[
"distances",
"dimensional",
"dimensionality",
"embeddings",
"metrics",
"metric",
"similarity",
"embedding",
"dimension",
"distance"
],
[
"describe",
"generalizing",
"brain",
"generalization",
"generalize",
"intuition",
"intuitive",
"explanations",
"neural",
"cognitive"
],
[
"classes",
"out",
"phase",
"inducing",
"phases",
"class",
"training",
"learning",
"arises",
"learnt"
],
[
"federated",
"private",
"privacy",
"shared",
"decentralized",
"distributed",
"sharing",
"accessible",
"external",
"independently"
],
[
"ranging",
"3d",
"pose",
"driving",
"vision",
"vehicles",
"tracking",
"robotics",
"coordinates",
"moving"
],
[
"general",
"forecast",
"predicted",
"forecasting",
"forecasts",
"weather",
"predict",
"predicting",
"prediction",
"predictions"
],
[
"quantized",
"approximations",
"thresholding",
"approximates",
"approximating",
"approximation",
"threshold",
"interpretable",
"interpretability",
"trees"
],
[
"clinical",
"medical",
"healthcare",
"medicine",
"disease",
"diseases",
"patients",
"cancer",
"health",
"diagnosis"
],
[
"bandit",
"possibilities",
"expectation",
"bandits",
"retrieval",
"seek",
"searching",
"search",
"bound",
"coverage"
],
[
"area",
"drive",
"planning",
"car",
"city",
"driving",
"urban",
"iot",
"transportation",
"traffic"
],
[
"computational",
"algorithmic",
"algorithms",
"library",
"approaches",
"dynamically",
"dynamic",
"constructed",
"deterministic",
"visualization"
]
] | 327.507221 | all-MiniLM-L6-v2 | 0.973333 | -0.250205 | 0.195052 | 0.890368 |
ArXiv ML Papers | S³_angular | 43 | 40 | [
[
"teacher",
"learner",
"student",
"trained",
"learns",
"studying",
"learnt",
"learners",
"learnable",
"revisit"
],
[
"adversary",
"security",
"attack",
"attacker",
"attacks",
"adversarial",
"adversarially",
"exploit",
"malicious",
"intrusion"
],
[
"stronger",
"enhance",
"enhanced",
"weaker",
"advanced",
"level",
"enhances",
"ease",
"adapted",
"away"
],
[
"softmax",
"classification",
"algorithm",
"classifiers",
"classifier",
"randomized",
"svm",
"encode",
"ranking",
"similarity"
],
[
"viewed",
"similar",
"convex",
"neurons",
"grid",
"tend",
"fields",
"represents",
"field",
"convexity"
],
[
"iteration",
"discover",
"inform",
"iteratively",
"iterations",
"guide",
"without",
"dynamically",
"build",
"while"
],
[
"nlp",
"literature",
"semantic",
"textual",
"texts",
"text",
"document",
"corpus",
"dictionary",
"words"
],
[
"ai",
"robots",
"machine",
"tasks",
"robot",
"machines",
"robotic",
"task",
"autonomous",
"intelligence"
],
[
"represents",
"represent",
"represented",
"covid",
"place",
"runtime",
"interpretable",
"interpret",
"representing",
"environments"
],
[
"medicine",
"healthcare",
"disease",
"diseases",
"clinical",
"health",
"patients",
"cancer",
"medical",
"covid"
],
[
"determine",
"by",
"lead",
"consensus",
"co",
"cnns",
"on",
"follow",
"cnn",
"frac"
],
[
"pseudo",
"patterns",
"sequences",
"trick",
"rigorous",
"approximating",
"benchmarks",
"smoothing",
"sequence",
"speedups"
],
[
"bandits",
"bandit",
"searching",
"seek",
"search",
"strategy",
"obtains",
"greedy",
"optimal",
"choosing"
],
[
"reconstruction",
"decoding",
"matrix",
"matrices",
"imputation",
"reconstructed",
"denoising",
"reconstructing",
"sparse",
"decompose"
],
[
"efforts",
"acceleration",
"100",
"vector",
"vectors",
"magnitude",
"fail",
"imbalance",
"attempts",
"aims"
],
[
"unsupervised",
"imagenet",
"cnns",
"cnn",
"recognition",
"convolutions",
"bernoulli",
"representations",
"probabilistic",
"convolution"
],
[
"reliability",
"guarantee",
"guarantees",
"markov",
"recall",
"discriminator",
"state",
"safety",
"bayes",
"positive"
],
[
"recognize",
"identified",
"facial",
"identify",
"black",
"face",
"identification",
"white",
"identifying",
"faces"
],
[
"gans",
"gan",
"generating",
"generative",
"generator",
"generation",
"generate",
"generated",
"generates",
"producing"
],
[
"neighborhood",
"clusters",
"groups",
"cluster",
"clustering",
"segmentation",
"locally",
"communities",
"unsupervised",
"partitioning"
],
[
"vectors",
"tensor",
"tensors",
"respectively",
"underlying",
"each",
"vector",
"multivariate",
"manner",
"quite"
],
[
"devise",
"wearable",
"surrogate",
"device",
"physical",
"running",
"devices",
"eeg",
"activity",
"walk"
],
[
"intuition",
"brain",
"explanations",
"neurons",
"neuron",
"neural",
"eeg",
"intelligence",
"cognitive",
"explore"
],
[
"insights",
"discover",
"deep",
"benchmark",
"benchmarks",
"explored",
"coverage",
"challenging",
"tests",
"metrics"
],
[
"ensemble",
"combination",
"named",
"manually",
"subsets",
"fusion",
"hybrid",
"split",
"subset",
"combinations"
],
[
"classifying",
"classifiers",
"labels",
"labeling",
"labeled",
"supervised",
"label",
"unlabeled",
"classifier",
"classification"
],
[
"digital",
"define",
"definition",
"qualitative",
"rigorous",
"rigorously",
"empirically",
"empirical",
"naive",
"physical"
],
[
"mri",
"gpus",
"extracted",
"imaging",
"cancer",
"extraction",
"segment",
"stages",
"region",
"segmentation"
],
[
"determined",
"dimensions",
"dimensionality",
"inferring",
"depth",
"deep",
"entire",
"dimension",
"dimensional",
"resource"
],
[
"forecasts",
"forecasting",
"predicted",
"predictions",
"weather",
"forecast",
"predict",
"predicts",
"prediction",
"predicting"
],
[
"corpus",
"learning",
"deterministic",
"corpora",
"learnt",
"ones",
"preliminary",
"theoretical",
"practice",
"recognize"
],
[
"particle",
"fourier",
"quantized",
"quantization",
"quantum",
"resonance",
"showed",
"observable",
"in",
"physics"
],
[
"networks",
"graphs",
"nodes",
"vertex",
"node",
"vertices",
"edges",
"edge",
"graph",
"spectral"
],
[
"predicts",
"predict",
"predictive",
"confidence",
"predicted",
"calibrated",
"predictors",
"predictions",
"predictor",
"trust"
],
[
"vehicle",
"gps",
"transportation",
"car",
"traffic",
"driving",
"drive",
"transport",
"navigation",
"vehicles"
],
[
"estimation",
"97",
"98",
"fitting",
"asymptotically",
"ml",
"asymptotic",
"96",
"likelihood",
"throughput"
],
[
"semantic",
"logic",
"symbolic",
"computation",
"solvers",
"recursive",
"diagnostic",
"solver",
"constraint",
"formulate"
],
[
"euclidean",
"constraint",
"constrain",
"linearly",
"constraints",
"unconstrained",
"exactly",
"symmetric",
"dimension",
"linear"
],
[
"thresholding",
"threshold",
"minimum",
"softmax",
"trees",
"minimal",
"tree",
"approximations",
"soft",
"maximum"
],
[
"classification",
"boost",
"boosting",
"samples",
"dataset",
"collect",
"actively",
"learns",
"popular",
"sampling"
]
] | 260.62405 | all-MiniLM-L6-v2 | 0.9175 | -0.296678 | 0.198891 | 0.894106 |
ArXiv ML Papers | S³_angular | 44 | 40 | [
[
"safety",
"risk",
"reliability",
"positive",
"certain",
"probabilistic",
"recall",
"curse",
"guarantees",
"supervision"
],
[
"standard",
"regular",
"batch",
"primary",
"activation",
"distribution",
"uniformly",
"tend",
"distributions",
"regularized"
],
[
"combine",
"combined",
"specific",
"setup",
"objective",
"fusion",
"combines",
"combination",
"combinations",
"dual"
],
[
"inspection",
"test",
"tests",
"patches",
"reports",
"patch",
"testing",
"developers",
"development",
"tested"
],
[
"attempts",
"magnitude",
"vector",
"fail",
"acceleration",
"vectors",
"imbalance",
"vs",
"topic",
"100"
],
[
"vertices",
"spectral",
"vertex",
"edges",
"nodes",
"node",
"networks",
"edge",
"graphs",
"graph"
],
[
"scales",
"leads",
"lead",
"tensors",
"tensor",
"oriented",
"performs",
"manner",
"easily",
"manifold"
],
[
"deterministic",
"paradigm",
"biological",
"theoretical",
"corpus",
"recognizing",
"practice",
"recognize",
"corpora",
"abstract"
],
[
"neuron",
"intractable",
"eeg",
"propagation",
"paths",
"brain",
"neurons",
"cognitive",
"neural",
"exhaustive"
],
[
"function",
"predicts",
"algorithm",
"predict",
"predictions",
"optimizing",
"optimal",
"optimizer",
"predicting",
"algorithms"
],
[
"ensures",
"characteristic",
"characteristics",
"curse",
"character",
"fidelity",
"transferring",
"cannot",
"tractable",
"factors"
],
[
"gps",
"transportation",
"traffic",
"driving",
"vehicle",
"navigation",
"drive",
"car",
"vehicles",
"transport"
],
[
"feature",
"generalized",
"norm",
"lasso",
"features",
"extent",
"trivial",
"exactly",
"alleviate",
"bounded"
],
[
"model",
"encoder",
"autoencoder",
"trainable",
"encode",
"autoencoders",
"encoders",
"modelling",
"encodes",
"models"
],
[
"gaussian",
"likelihood",
"empirical",
"laplacian",
"distributions",
"densities",
"density",
"diffusion",
"distribution",
"spatiotemporal"
],
[
"termed",
"completion",
"dedicated",
"complete",
"interpolation",
"maintaining",
"meaningful",
"marginal",
"formulate",
"compress"
],
[
"perturbation",
"forgetting",
"perturbed",
"perturbations",
"stability",
"predicts",
"overfitting",
"regret",
"imbalanced",
"consistently"
],
[
"generates",
"generative",
"gans",
"gan",
"generating",
"generate",
"generated",
"generator",
"generation",
"producing"
],
[
"infer",
"confidence",
"millions",
"experimental",
"researchers",
"tested",
"experimentally",
"cnn",
"estimated",
"aim"
],
[
"stored",
"compressed",
"truncated",
"rigorous",
"storing",
"embeddings",
"numerically",
"unstructured",
"embedding",
"variants"
],
[
"algorithms",
"algorithm",
"algorithmic",
"autoencoder",
"autoencoders",
"model",
"personalized",
"learning",
"software",
"models"
],
[
"quantization",
"quantum",
"particle",
"quantized",
"fourier",
"showed",
"resonance",
"in",
"observable",
"states"
],
[
"fairness",
"biased",
"bias",
"representative",
"discriminate",
"discrimination",
"biases",
"equal",
"unbiased",
"fair"
],
[
"seek",
"bandits",
"bandit",
"obtains",
"search",
"searching",
"strategy",
"greedy",
"thought",
"optimal"
],
[
"grid",
"tend",
"represents",
"similar",
"fields",
"neurons",
"field",
"viewed",
"view",
"descent"
],
[
"dependencies",
"imitation",
"person",
"invariance",
"abstraction",
"perceptual",
"naturally",
"distinguish",
"components",
"aspect"
],
[
"tensor",
"decomposition",
"denoising",
"mri",
"matrix",
"matrices",
"basis",
"tensors",
"imputation",
"factorization"
],
[
"insufficient",
"drastically",
"dramatically",
"minimum",
"minimal",
"sparse",
"gnn",
"modal",
"severely",
"difficulty"
],
[
"depth",
"chosen",
"representations",
"characteristic",
"dimension",
"dimensional",
"dimensionality",
"dimensions",
"representation",
"characteristics"
],
[
"softmax",
"classifiers",
"discriminative",
"priors",
"classifying",
"classify",
"unlabeled",
"distinguishing",
"conventional",
"classification"
],
[
"forecasts",
"weather",
"predictions",
"prediction",
"predicted",
"forecast",
"forecasting",
"predict",
"predicting",
"predicts"
],
[
"decisions",
"classification",
"music",
"videos",
"video",
"convolutions",
"analytics",
"visualization",
"quantitative",
"convolution"
],
[
"rather",
"deeper",
"dependent",
"passing",
"way",
"best",
"information",
"authors",
"better",
"aiming"
],
[
"cnn",
"pipelines",
"deep",
"cnns",
"pipeline",
"resnet",
"deeper",
"rich",
"acquire",
"dense"
],
[
"multi",
"multivariate",
"components",
"network",
"networks",
"vectors",
"net",
"width",
"covariates",
"softmax"
],
[
"reinforcement",
"behavioral",
"actions",
"action",
"games",
"policy",
"behaviors",
"play",
"enforce",
"rewards"
],
[
"randomized",
"measure",
"gaussian",
"spectral",
"classifier",
"softmax",
"normalized",
"accuracy",
"svm",
"recognition"
],
[
"similarity",
"comparable",
"distances",
"distance",
"nearest",
"compares",
"ranking",
"similarities",
"metric",
"metrics"
],
[
"measurement",
"task",
"tasks",
"operates",
"uncertainties",
"operate",
"ensuring",
"projections",
"projection",
"reliability"
],
[
"rnns",
"recurrent",
"sequences",
"lstm",
"temporal",
"sequence",
"sequential",
"autoregressive",
"rnn",
"shows"
]
] | 285.377113 | all-MiniLM-L6-v2 | 0.9175 | -0.323559 | 0.19416 | 0.887283 |
ArXiv ML Papers | S³_angular | 45 | 40 | [
[
"equal",
"discrimination",
"unbiased",
"discriminate",
"fairness",
"biases",
"biased",
"bias",
"representative",
"fair"
],
[
"tensors",
"tensor",
"decomposition",
"basis",
"matrices",
"matrix",
"vectors",
"vector",
"3d",
"sublinear"
],
[
"accuracy",
"reliable",
"reliability",
"uncertainties",
"uncertainty",
"accurately",
"precision",
"consistently",
"consistent",
"accuracies"
],
[
"attacks",
"adversary",
"attack",
"adversarially",
"adversarial",
"attacker",
"exploit",
"security",
"intrusion",
"exploits"
],
[
"samples",
"dataset",
"screening",
"datasets",
"sample",
"finds",
"tested",
"benchmark",
"batch",
"discover"
],
[
"rnn",
"sequences",
"lstm",
"sequence",
"sequential",
"recurrent",
"rnns",
"autoregressive",
"temporal",
"future"
],
[
"scalability",
"potential",
"shape",
"marginal",
"oriented",
"assist",
"width",
"scalable",
"wide",
"multivariate"
],
[
"dimensionality",
"kernels",
"kernel",
"perturbations",
"perturbation",
"estimates",
"spectral",
"gaussian",
"laplacian",
"landscape"
],
[
"weather",
"forecasts",
"predictions",
"predicted",
"predict",
"forecast",
"forecasting",
"predicting",
"prediction",
"predicts"
],
[
"predict",
"predictions",
"predictive",
"predicts",
"randomized",
"algorithm",
"predicting",
"prediction",
"function",
"optimal"
],
[
"description",
"estimate",
"embedding",
"embeddings",
"needs",
"log",
"estimation",
"estimator",
"solvers",
"estimates"
],
[
"abstraction",
"perceptual",
"invariance",
"signals",
"dependencies",
"naturally",
"person",
"efficient",
"scenes",
"environmental"
],
[
"goals",
"objectives",
"objective",
"projection",
"projections",
"fidelity",
"goal",
"multivariate",
"capabilities",
"incorporate"
],
[
"recommender",
"recommendations",
"suggests",
"ranking",
"factorization",
"recommendation",
"suggest",
"rank",
"percentage",
"profile"
],
[
"imagenet",
"mnist",
"representations",
"cnn",
"convolutions",
"unsupervised",
"cnns",
"poses",
"bernoulli",
"sensing"
],
[
"simulation",
"monte",
"posterior",
"bayesian",
"probabilistic",
"markov",
"mcmc",
"carlo",
"simulations",
"rnns"
],
[
"modeled",
"dynamics",
"surface",
"rigorous",
"numerically",
"manifold",
"modeling",
"dynamical",
"nonlinear",
"simulations"
],
[
"sound",
"music",
"voice",
"audio",
"speaker",
"acoustic",
"recorded",
"speech",
"tune",
"frequencies"
],
[
"regularized",
"kernel",
"backpropagation",
"regularization",
"normalizing",
"normalization",
"normalized",
"factorization",
"principal",
"par"
],
[
"node",
"graphs",
"graph",
"vertex",
"edge",
"vertices",
"edges",
"nodes",
"networks",
"spectral"
],
[
"modal",
"insufficient",
"considerable",
"gnn",
"gnns",
"sufficiently",
"substantially",
"minimal",
"sufficient",
"approximates"
],
[
"these",
"segmentation",
"segment",
"extraction",
"stage",
"edges",
"context",
"stages",
"period",
"periods"
],
[
"acceleration",
"magnitude",
"attempts",
"vector",
"vectors",
"fail",
"reference",
"100",
"aims",
"effort"
],
[
"explaining",
"interpretable",
"explanations",
"interpretability",
"describe",
"understanding",
"interpret",
"interpretation",
"visual",
"describes"
],
[
"cnns",
"rich",
"cnn",
"resnet",
"filters",
"dense",
"pipelines",
"filtering",
"pipeline",
"deep"
],
[
"quantum",
"quantized",
"quantization",
"particle",
"resonance",
"in",
"physics",
"fourier",
"observable",
"discovered"
],
[
"diversity",
"weak",
"diverse",
"individually",
"weakly",
"broadly",
"mnist",
"ensemble",
"global",
"ensembles"
],
[
"videos",
"music",
"digital",
"video",
"decisions",
"identification",
"classification",
"visualization",
"identify",
"analytics"
],
[
"robots",
"ai",
"machine",
"autonomous",
"robotic",
"machines",
"robot",
"tasks",
"intelligent",
"artificial"
],
[
"rather",
"take",
"pass",
"dependent",
"information",
"passing",
"deeper",
"derive",
"best",
"operator"
],
[
"processes",
"serve",
"softmax",
"kernels",
"core",
"decompose",
"weights",
"relu",
"decomposition",
"generator"
],
[
"attention",
"rank",
"rnn",
"ranging",
"boosting",
"trainable",
"rnns",
"overfitting",
"inputs",
"localization"
],
[
"lead",
"goals",
"reach",
"offs",
"scoring",
"between",
"determine",
"measuring",
"scores",
"pairwise"
],
[
"system",
"classified",
"detect",
"basic",
"ucb",
"auc",
"sensors",
"matrices",
"languages",
"systems"
],
[
"gradient",
"priors",
"backpropagation",
"nonparametric",
"gradients",
"prior",
"hessian",
"differentiation",
"softmax",
"method"
],
[
"bottleneck",
"likelihood",
"asymptotic",
"asymptotically",
"97",
"98",
"fitting",
"ml",
"inferring",
"throughput"
],
[
"dimensionality",
"characteristic",
"dimensional",
"inferring",
"depth",
"collected",
"dimensions",
"entire",
"dimension",
"resource"
],
[
"classifier",
"softmax",
"randomized",
"measure",
"encode",
"classifiers",
"classification",
"spectral",
"algorithm",
"classified"
],
[
"brain",
"neurons",
"eeg",
"neuron",
"neural",
"mri",
"cognitive",
"propagation",
"analytically",
"volume"
],
[
"model",
"modeled",
"modeling",
"models",
"simulation",
"calibration",
"networks",
"suited",
"trained",
"modelling"
]
] | 355.194545 | all-MiniLM-L6-v2 | 0.9 | -0.310509 | 0.202086 | 0.894465 |
ArXiv ML Papers | S³_angular | 46 | 40 | [
[
"boost",
"boosting",
"giving",
"extracts",
"mainly",
"interest",
"natural",
"specifically",
"particular",
"extract"
],
[
"blind",
"rather",
"pass",
"cross",
"deeper",
"passing",
"dependent",
"derive",
"information",
"visual"
],
[
"mri",
"extraction",
"segmentation",
"segment",
"extracted",
"imaging",
"region",
"cancer",
"stages",
"ct"
],
[
"generalizes",
"generalize",
"generalization",
"incremental",
"generalizing",
"granularity",
"termed",
"parameterized",
"multimodal",
"softmax"
],
[
"music",
"classification",
"digital",
"visualization",
"video",
"decisions",
"videos",
"classifier",
"style",
"probabilistic"
],
[
"contexts",
"scenes",
"context",
"grounded",
"contextual",
"pixels",
"parts",
"convolutions",
"distinguish",
"spectral"
],
[
"analyze",
"equations",
"reconstructed",
"reconstruction",
"nonlinear",
"analytical",
"simulations",
"solvers",
"denoising",
"inverse"
],
[
"energy",
"tail",
"bandwidth",
"resnet",
"evolution",
"yielding",
"bounded",
"nets",
"specified",
"light"
],
[
"ranging",
"rnns",
"gps",
"rnn",
"attention",
"boosting",
"rank",
"trainable",
"vertex",
"inputs"
],
[
"cnn",
"off",
"cnns",
"mnist",
"convolutions",
"imagenet",
"probabilistic",
"unbiased",
"networks",
"unsupervised"
],
[
"dedicated",
"completion",
"termed",
"meaningful",
"maintaining",
"complete",
"compress",
"baselines",
"comprehensive",
"marginal"
],
[
"generated",
"generative",
"generator",
"gans",
"generating",
"generates",
"gan",
"generate",
"generation",
"producing"
],
[
"diverse",
"diversity",
"weak",
"individually",
"global",
"mixed",
"multiple",
"broadly",
"multi",
"stacked"
],
[
"rnns",
"recurrent",
"sequence",
"lstm",
"sequences",
"sequential",
"rnn",
"temporal",
"autoregressive",
"industry"
],
[
"gradients",
"hessian",
"gradient",
"classifying",
"classifiers",
"backpropagation",
"softmax",
"labeled",
"supervised",
"discriminative"
],
[
"trained",
"predictors",
"predictor",
"predictive",
"predict",
"modeled",
"model",
"predicts",
"models",
"predicted"
],
[
"neurons",
"tend",
"similar",
"fields",
"grid",
"represents",
"convex",
"field",
"viewed",
"tilde"
],
[
"dynamical",
"dynamics",
"invariant",
"embeddings",
"stability",
"stationary",
"express",
"embedding",
"represented",
"expressive"
],
[
"obtains",
"seek",
"bandits",
"strategy",
"greedy",
"bandit",
"search",
"searching",
"optimal",
"choosing"
],
[
"additive",
"combining",
"field",
"mixture",
"mini",
"formulation",
"partial",
"included",
"separately",
"composed"
],
[
"criterion",
"gaussian",
"layer",
"activations",
"activation",
"parameters",
"parameter",
"hyperparameter",
"parametric",
"optimum"
],
[
"laplacian",
"gaussian",
"near",
"estimators",
"estimates",
"estimate",
"kernel",
"kernels",
"landscape",
"approximately"
],
[
"softmax",
"decomposition",
"serve",
"core",
"decompose",
"weights",
"relu",
"kernels",
"processes",
"generator"
],
[
"norm",
"lasso",
"constrain",
"regularization",
"trivial",
"setting",
"feature",
"sparse",
"sparsity",
"constrained"
],
[
"optimizing",
"maximization",
"optimisation",
"optimization",
"optimize",
"maximizing",
"optimized",
"optimizer",
"objective",
"function"
],
[
"behavioral",
"action",
"enforce",
"actions",
"policy",
"policies",
"behaviors",
"reinforcement",
"critic",
"games"
],
[
"transferred",
"transfer",
"transferring",
"domains",
"adaptation",
"adapting",
"domain",
"across",
"unlabeled",
"retaining"
],
[
"gradient",
"descent",
"hessian",
"sgd",
"ascent",
"backpropagation",
"accelerated",
"convex",
"convexity",
"gradients"
],
[
"predictions",
"weather",
"prediction",
"predicted",
"forecast",
"forecasting",
"forecasts",
"predicts",
"predict",
"predicting"
],
[
"autoencoder",
"autoencoders",
"latent",
"stochastic",
"variational",
"bayesian",
"likelihood",
"modeled",
"modeling",
"generative"
],
[
"walk",
"device",
"surrogate",
"activity",
"devices",
"devise",
"physical",
"running",
"eeg",
"wearable"
],
[
"observation",
"supervised",
"estimating",
"estimator",
"sampled",
"estimators",
"sampling",
"observations",
"observed",
"estimation"
],
[
"bias",
"discriminate",
"equal",
"fairness",
"discrimination",
"unbiased",
"biases",
"biased",
"representative",
"fair"
],
[
"cognitive",
"brain",
"intelligence",
"perception",
"focusing",
"neural",
"neuron",
"perceptual",
"neurons",
"attention"
],
[
"deployment",
"service",
"serve",
"iot",
"infrastructure",
"multi",
"server",
"task",
"message",
"setup"
],
[
"leads",
"machine",
"robots",
"machines",
"robotic",
"lead",
"tools",
"tensors",
"perform",
"tasks"
],
[
"recommender",
"recommendation",
"recommendations",
"suggests",
"ranking",
"profile",
"factorization",
"suggest",
"personalized",
"percentage"
],
[
"boundaries",
"dense",
"paths",
"expansion",
"densities",
"density",
"concrete",
"volume",
"numerically",
"boundary"
],
[
"interventions",
"requiring",
"involve",
"causal",
"involved",
"impacts",
"inference",
"aid",
"involving",
"partially"
],
[
"vector",
"magnitude",
"vectors",
"acceleration",
"attempts",
"efforts",
"fail",
"vs",
"topic",
"effort"
]
] | 273.248837 | all-MiniLM-L6-v2 | 0.9275 | -0.315199 | 0.183221 | 0.895347 |
ArXiv ML Papers | S³_angular | 43 | 50 | [
[
"pipeline",
"directed",
"downstream",
"pipelines",
"flows",
"flow",
"superior",
"predictive",
"cnn",
"cnns"
],
[
"grid",
"subspace",
"represent",
"representation",
"matrix",
"matrices",
"vs",
"may",
"subspaces",
"represents"
],
[
"25",
"exploring",
"integrate",
"integrating",
"integration",
"intelligent",
"ease",
"directions",
"50",
"explored"
],
[
"other",
"convolutional",
"essential",
"offers",
"ideas",
"importance",
"things",
"utility",
"broader",
"contrastive"
],
[
"discriminate",
"biases",
"bias",
"fairness",
"discrimination",
"biased",
"equally",
"representative",
"candidate",
"fair"
],
[
"product",
"understand",
"interpretability",
"interpretable",
"industry",
"credit",
"mean",
"valued",
"understanding",
"financial"
],
[
"convolution",
"convolutions",
"kernel",
"kernels",
"multivariate",
"orthogonal",
"gaussian",
"convolutional",
"process",
"target"
],
[
"layer",
"networks",
"net",
"neural",
"backpropagation",
"neuron",
"neurons",
"layers",
"network",
"inputs"
],
[
"aid",
"sources",
"interventions",
"causal",
"involve",
"associated",
"involved",
"inter",
"evidence",
"trigger"
],
[
"outperform",
"roughly",
"popularity",
"nearly",
"scale",
"perturbation",
"drastically",
"highly",
"perturbations",
"outperforms"
],
[
"replacing",
"transfer",
"transferring",
"transferred",
"adapting",
"adapted",
"adapt",
"adaptation",
"changing",
"transformation"
],
[
"policies",
"reinforcement",
"policy",
"actions",
"games",
"critic",
"enforce",
"action",
"behavioral",
"textit"
],
[
"throughput",
"hardware",
"computing",
"computes",
"computation",
"gpus",
"faster",
"gpu",
"runtime",
"processing"
],
[
"languages",
"basic",
"ucb",
"basis",
"detect",
"matrices",
"molecular",
"policies",
"dictionary",
"sensors"
],
[
"distances",
"compares",
"nearest",
"compare",
"metrics",
"metric",
"distance",
"similarity",
"comparable",
"comparison"
],
[
"extensive",
"sampling",
"extending",
"expansion",
"constructing",
"extends",
"arbitrary",
"trees",
"boundaries",
"extent"
],
[
"variant",
"represents",
"frequently",
"convex",
"convexity",
"image",
"fields",
"naive",
"similar",
"almost"
],
[
"dynamics",
"dynamic",
"dynamical",
"stationary",
"transitions",
"formally",
"transition",
"transformations",
"invariant",
"preserving"
],
[
"embedding",
"imagenet",
"threat",
"representations",
"energy",
"camera",
"demand",
"half",
"notions",
"bounded"
],
[
"specified",
"specific",
"range",
"optimizer",
"minutes",
"min",
"minimum",
"optimization",
"optimisation",
"optimal"
],
[
"adversary",
"attacker",
"adversarial",
"adversarially",
"attacks",
"security",
"adversaries",
"exploits",
"malicious",
"attack"
],
[
"restricted",
"limitation",
"able",
"allow",
"allowed",
"character",
"case",
"cannot",
"can",
"root"
],
[
"unlabeled",
"label",
"classified",
"labels",
"labeling",
"classification",
"labeled",
"classify",
"supervised",
"classifying"
],
[
"robot",
"robotics",
"pose",
"poses",
"robotic",
"coordinate",
"joint",
"position",
"coordinates",
"robots"
],
[
"partial",
"intrinsic",
"extraction",
"evaluate",
"distinguish",
"confirm",
"verify",
"phases",
"assess",
"yield"
],
[
"bandits",
"searching",
"seek",
"search",
"bandit",
"exploitation",
"obtains",
"market",
"finds",
"price"
],
[
"structural",
"protein",
"find",
"molecular",
"discover",
"step",
"added",
"become",
"steps",
"generated"
],
[
"completion",
"shot",
"through",
"poisoning",
"descent",
"down",
"exhaustive",
"fail",
"reflect",
"air"
],
[
"monte",
"mcmc",
"markov",
"simulation",
"carlo",
"stochastic",
"simulations",
"randomly",
"sampling",
"simulated"
],
[
"raw",
"software",
"spatio",
"empirical",
"corpus",
"tend",
"multimodal",
"contextual",
"tool",
"vanilla"
],
[
"wide",
"risks",
"divergence",
"specific",
"marginal",
"risk",
"width",
"potential",
"margin",
"areas"
],
[
"dimension",
"selects",
"dimensional",
"datasets",
"dimensionality",
"representations",
"capacity",
"dataset",
"inputs",
"dimensions"
],
[
"channel",
"raw",
"decoder",
"deeper",
"compression",
"degrade",
"encoded",
"information",
"channels",
"reconstructed"
],
[
"nlp",
"textual",
"translation",
"texts",
"corpora",
"linguistic",
"sentences",
"english",
"corpus",
"text"
],
[
"pruning",
"pruned",
"reduced",
"minimal",
"reduces",
"distinct",
"compressed",
"batch",
"reduce",
"compress"
],
[
"flow",
"process",
"bottleneck",
"throughput",
"scene",
"scenes",
"content",
"topic",
"body",
"processes"
],
[
"dependent",
"get",
"last",
"pass",
"passing",
"take",
"attributes",
"takes",
"lightweight",
"gives"
],
[
"ensemble",
"ensembles",
"bridge",
"streams",
"continuous",
"stream",
"identification",
"entries",
"combine",
"incremental"
],
[
"encode",
"encodes",
"code",
"encoders",
"autoencoders",
"encoder",
"autoencoder",
"variational",
"codes",
"encoded"
],
[
"embeddings",
"relu",
"core",
"poses",
"weights",
"embedding",
"embedded",
"critic",
"soft",
"energy"
],
[
"federated",
"distributed",
"communicate",
"protocol",
"cooperative",
"shared",
"centralized",
"collaborative",
"communications",
"communication"
],
[
"vertices",
"vertex",
"graphs",
"graph",
"edge",
"nodes",
"node",
"edges",
"networks",
"graphical"
],
[
"clinical",
"medical",
"patients",
"diagnosis",
"medicine",
"cancer",
"mri",
"diagnostic",
"practitioners",
"healthcare"
],
[
"monte",
"defined",
"termed",
"defines",
"initialization",
"dynamically",
"characterized",
"cell",
"describes",
"compact"
],
[
"style",
"proposes",
"formalism",
"imitation",
"texts",
"manipulation",
"manner",
"patterns",
"traditional",
"fashion"
],
[
"boosting",
"tree",
"forest",
"classifier",
"boost",
"classifiers",
"trees",
"classify",
"xgboost",
"rate"
],
[
"recommendation",
"factorization",
"suggest",
"recommender",
"recommendations",
"profile",
"suggests",
"personalized",
"ranking",
"profiles"
],
[
"outliers",
"iot",
"outlier",
"anomaly",
"anomalies",
"event",
"intrusion",
"detect",
"sensor",
"deviation"
],
[
"71",
"uniform",
"remains",
"quality",
"hundreds",
"thousands",
"website",
"full",
"uniformly",
"estimates"
],
[
"nonlinear",
"differential",
"differentiation",
"derivatives",
"hessian",
"differentially",
"integrating",
"optimization",
"differentiable",
"gradient"
]
] | 274.481598 | all-MiniLM-L6-v2 | 0.946 | -0.330391 | 0.187578 | 0.900631 |
ArXiv ML Papers | S³_angular | 44 | 50 | [
[
"lstm",
"sequences",
"recurrent",
"rnns",
"forecasting",
"forecasts",
"sequential",
"sequence",
"records",
"temporal"
],
[
"protein",
"molecular",
"find",
"become",
"added",
"discover",
"comes",
"structural",
"obtained",
"execute"
],
[
"publicly",
"conventional",
"privacy",
"independent",
"hidden",
"federated",
"private",
"natural",
"public",
"independently"
],
[
"another",
"on",
"twitter",
"post",
"fast",
"lead",
"https",
"video",
"goal",
"scoring"
],
[
"transform",
"aggregation",
"preserving",
"transformers",
"operators",
"transforms",
"dynamical",
"transformations",
"invariant",
"transformation"
],
[
"euclidean",
"squared",
"lipschitz",
"norm",
"criterion",
"lasso",
"enables",
"settings",
"extent",
"bound"
],
[
"features",
"identify",
"trick",
"additionally",
"ways",
"collect",
"feature",
"boosting",
"boost",
"identified"
],
[
"gan",
"gans",
"generator",
"hilbert",
"generate",
"generated",
"generates",
"generating",
"generation",
"generative"
],
[
"head",
"mini",
"mixture",
"partial",
"minimax",
"composed",
"scenes",
"discriminative",
"aspects",
"proposals"
],
[
"embeddings",
"density",
"defend",
"embedded",
"embedding",
"represent",
"incorporate",
"inequality",
"densities",
"integrate"
],
[
"posterior",
"bayesian",
"priors",
"prior",
"bayes",
"priori",
"likelihood",
"assuming",
"conditioned",
"mcmc"
],
[
"expand",
"expansion",
"detailed",
"incremental",
"inform",
"massive",
"suite",
"iteration",
"exploration",
"iterations"
],
[
"empirically",
"extensively",
"researchers",
"ensembles",
"sample",
"observations",
"empirical",
"samples",
"sampling",
"estimating"
],
[
"formulate",
"modeled",
"computationally",
"models",
"modeling",
"gpu",
"faster",
"profiles",
"including",
"structure"
],
[
"gradient",
"fusion",
"go",
"fidelity",
"mutual",
"mixed",
"compares",
"share",
"gnns",
"bleu"
],
[
"factor",
"latent",
"tensors",
"tensor",
"pytorch",
"vector",
"factorization",
"vectors",
"factors",
"magnetic"
],
[
"detect",
"ensemble",
"videos",
"predictors",
"recommender",
"detecting",
"supervised",
"recommendation",
"audio",
"video"
],
[
"factorization",
"suggesting",
"profile",
"recommender",
"recommendations",
"recommendation",
"suggest",
"personalized",
"suggests",
"ranking"
],
[
"industry",
"interpretability",
"interpretable",
"describes",
"description",
"product",
"understand",
"service",
"describe",
"mean"
],
[
"embedding",
"raw",
"embeddings",
"embedded",
"soft",
"relu",
"weights",
"core",
"poses",
"critic"
],
[
"person",
"known",
"deterministic",
"historical",
"adaptively",
"predictor",
"dependencies",
"squared",
"predictors",
"inefficient"
],
[
"neural",
"cognitive",
"intelligence",
"intractable",
"eeg",
"neurons",
"brain",
"neuron",
"mri",
"coherent"
],
[
"averaging",
"measure",
"aggregate",
"score",
"valued",
"np",
"outcomes",
"measures",
"nonnegative",
"dedicated"
],
[
"gives",
"pass",
"get",
"dependent",
"lightweight",
"last",
"expected",
"passing",
"attributes",
"give"
],
[
"fourier",
"quantum",
"hilbert",
"quantization",
"quantized",
"resonance",
"answering",
"observable",
"computed",
"discovery"
],
[
"estimated",
"nearest",
"minimum",
"softmax",
"estimation",
"near",
"minutes",
"proposes",
"approximation",
"min"
],
[
"camera",
"images",
"imagenet",
"energy",
"recognition",
"notions",
"representations",
"want",
"bernoulli",
"supervised"
],
[
"tractable",
"modal",
"minimax",
"sparse",
"mini",
"threshold",
"severe",
"described",
"minimum",
"consider"
],
[
"assessment",
"gan",
"scores",
"surrogate",
"highest",
"scoring",
"assess",
"assessing",
"rank",
"ranking"
],
[
"cnns",
"guarantees",
"markov",
"observing",
"ensuring",
"security",
"detection",
"vision",
"cnn",
"detecting"
],
[
"filtering",
"values",
"83",
"numbers",
"list",
"71",
"filters",
"filter",
"numerical",
"numerically"
],
[
"objective",
"optimisation",
"objectives",
"optimize",
"optimizes",
"optimizing",
"optimization",
"optimized",
"multi",
"maximizing"
],
[
"decomposition",
"normalization",
"normalized",
"principal",
"regularization",
"normalizing",
"characterizing",
"factorization",
"characterize",
"representative"
],
[
"which",
"solving",
"contrastive",
"11",
"other",
"frequencies",
"dependencies",
"dependency",
"offers",
"readily"
],
[
"randomized",
"algorithm",
"transformer",
"boosting",
"classifier",
"classifiers",
"svm",
"recognition",
"ranking",
"ensemble"
],
[
"forecasting",
"forecast",
"predicted",
"predictions",
"weather",
"forecasts",
"predict",
"predicts",
"prediction",
"agnostic"
],
[
"similar",
"represents",
"recently",
"image",
"convexity",
"variant",
"fields",
"frequently",
"convex",
"subsequently"
],
[
"fitting",
"candidate",
"candidates",
"98",
"inferring",
"model",
"having",
"models",
"blind",
"attributed"
],
[
"dnns",
"dnn",
"auc",
"infinite",
"hard",
"difficulty",
"ratio",
"names",
"10",
"epsilon"
],
[
"computation",
"gpu",
"computing",
"throughput",
"runtime",
"performance",
"hardware",
"faster",
"processing",
"cloud"
],
[
"identification",
"profiles",
"determined",
"methodology",
"identifies",
"validate",
"validation",
"identify",
"identified",
"validity"
],
[
"pipelines",
"pipeline",
"flows",
"directed",
"nas",
"downstream",
"flow",
"superior",
"largest",
"rich"
],
[
"attention",
"specifically",
"eye",
"namely",
"leading",
"vision",
"wise",
"perceptual",
"summary",
"focusing"
],
[
"bandwidth",
"diffusion",
"terms",
"traffic",
"mixture",
"formulation",
"flow",
"granularity",
"streams",
"gradients"
],
[
"patients",
"medicine",
"healthcare",
"medical",
"diseases",
"disease",
"clinical",
"health",
"patient",
"practitioners"
],
[
"gradient",
"convexity",
"convex",
"descent",
"sgd",
"hessian",
"ascent",
"gradients",
"convergence",
"backpropagation"
],
[
"program",
"detecting",
"implementations",
"detect",
"solvers",
"recursive",
"algorithm",
"algorithmic",
"algorithms",
"solver"
],
[
"who",
"82",
"include",
"includes",
"coefficients",
"estimator",
"selected",
"83",
"many",
"how"
],
[
"evaluations",
"wasserstein",
"analyse",
"assessing",
"hessian",
"contrastive",
"horizon",
"perspectives",
"assessment",
"rise"
],
[
"oriented",
"potential",
"specific",
"vary",
"wide",
"width",
"specialized",
"there",
"certain",
"connections"
]
] | 293.511892 | all-MiniLM-L6-v2 | 0.906 | -0.34121 | 0.172703 | 0.899664 |
ArXiv ML Papers | S³_angular | 45 | 50 | [
[
"determined",
"ensembles",
"continuous",
"defined",
"collect",
"numerically",
"generator",
"entries",
"collected",
"identification"
],
[
"roughly",
"nearly",
"popularity",
"scale",
"highly",
"outperform",
"drastically",
"significantly",
"scaling",
"extremely"
],
[
"representations",
"imagenet",
"convolutions",
"convolution",
"energy",
"robots",
"recognition",
"bernoulli",
"machine",
"convolutional"
],
[
"characterized",
"close",
"modal",
"check",
"gnn",
"gender",
"sparse",
"threshold",
"described",
"messages"
],
[
"regions",
"segmentation",
"cluster",
"clustering",
"clusters",
"unsupervised",
"partitioning",
"segment",
"subset",
"counterparts"
],
[
"scenes",
"bottleneck",
"flow",
"content",
"scene",
"recovery",
"process",
"throughput",
"brain",
"box"
],
[
"layer",
"networks",
"backpropagation",
"neurons",
"neuron",
"layers",
"neural",
"net",
"network",
"projections"
],
[
"normalization",
"normalizing",
"principal",
"normalized",
"characterizing",
"decomposition",
"invariant",
"characterize",
"regularization",
"attribute"
],
[
"technology",
"projections",
"contexts",
"procedures",
"task",
"measurements",
"decade",
"technologies",
"projection",
"context"
],
[
"discriminative",
"head",
"composed",
"mixture",
"minimax",
"proposals",
"aspects",
"mini",
"partial",
"scenes"
],
[
"termed",
"cell",
"monte",
"defined",
"defines",
"dynamically",
"initialization",
"defining",
"characterized",
"cells"
],
[
"predicted",
"validation",
"validated",
"validate",
"calibrated",
"accuracy",
"calibration",
"expect",
"predictions",
"scoring"
],
[
"patient",
"patients",
"medical",
"healthcare",
"clinical",
"medicine",
"diagnosis",
"cancer",
"mri",
"practitioners"
],
[
"sgd",
"gradient",
"hessian",
"descent",
"convexity",
"gradients",
"convex",
"convergence",
"ascent",
"backpropagation"
],
[
"on",
"post",
"another",
"goals",
"fast",
"lead",
"scoring",
"goal",
"leading",
"strategies"
],
[
"subspace",
"subspaces",
"vectors",
"represent",
"matrices",
"matrix",
"grid",
"vs",
"representation",
"representations"
],
[
"dynamical",
"dynamic",
"dynamics",
"dynamically",
"stationary",
"system",
"nonlinear",
"formally",
"preserving",
"systems"
],
[
"learns",
"learners",
"learner",
"student",
"study",
"learning",
"teacher",
"learnable",
"learnt",
"studying"
],
[
"molecular",
"biological",
"genetic",
"finding",
"find",
"structural",
"protein",
"discover",
"added",
"discovery"
],
[
"relu",
"poses",
"core",
"embeddings",
"soft",
"raw",
"weights",
"embedded",
"embedding",
"critic"
],
[
"production",
"industry",
"product",
"development",
"industrial",
"service",
"such",
"process",
"products",
"management"
],
[
"lstm",
"recurrent",
"rnns",
"sequences",
"sequence",
"sequential",
"rnn",
"temporal",
"recall",
"pattern"
],
[
"discriminate",
"bias",
"biases",
"fairness",
"discrimination",
"biased",
"candidate",
"equally",
"fair",
"representative"
],
[
"distinguishing",
"generalization",
"generalize",
"generalizing",
"softmax",
"summarize",
"distinguish",
"learning",
"lasso",
"randomly"
],
[
"games",
"behavioral",
"critic",
"action",
"actions",
"policy",
"policies",
"enforce",
"reinforcement",
"interventions"
],
[
"fourier",
"observable",
"resonance",
"answering",
"hilbert",
"quantization",
"quantized",
"quantum",
"discovery",
"particle"
],
[
"includes",
"who",
"include",
"82",
"themselves",
"was",
"incorporates",
"how",
"selected",
"usually"
],
[
"mixed",
"fidelity",
"coupled",
"mutual",
"go",
"gnns",
"fusion",
"share",
"mismatch",
"gradient"
],
[
"optimisation",
"optimizing",
"optimized",
"algorithm",
"optimize",
"optimization",
"heuristic",
"optimizer",
"optimal",
"algorithms"
],
[
"enables",
"lasso",
"alleviate",
"norm",
"criterion",
"regularization",
"operators",
"euclidean",
"lipschitz",
"generalized"
],
[
"able",
"codes",
"curse",
"valid",
"allowed",
"cannot",
"restricted",
"risks",
"deal",
"can"
],
[
"down",
"completion",
"poisoning",
"fail",
"descent",
"reflect",
"shot",
"exhaustive",
"air",
"efforts"
],
[
"predictors",
"videos",
"recognition",
"recommendation",
"detection",
"detecting",
"recommender",
"detect",
"supervised",
"ensemble"
],
[
"hessian",
"differentiation",
"differential",
"differentiable",
"implicit",
"derivatives",
"curve",
"nonlinear",
"simulations",
"equations"
],
[
"basic",
"dictionary",
"basis",
"detection",
"ucb",
"calibration",
"detect",
"sensors",
"languages",
"matrices"
],
[
"benchmark",
"remains",
"quality",
"website",
"respect",
"uniform",
"imagenet",
"highlights",
"thousands",
"treat"
],
[
"namely",
"adopted",
"promise",
"sketch",
"specifically",
"than",
"adopt",
"cannot",
"expect",
"amenable"
],
[
"robot",
"pose",
"robots",
"robotic",
"position",
"poses",
"robotics",
"coordinate",
"assisted",
"capable"
],
[
"shared",
"communicate",
"collaborative",
"cooperative",
"federated",
"distributed",
"protocol",
"centralized",
"communications",
"server"
],
[
"methodology",
"discriminative",
"known",
"solver",
"intelligence",
"ai",
"intelligent",
"controller",
"adaptively",
"person"
],
[
"kernel",
"gaussian",
"kernels",
"orthogonal",
"parametric",
"hyperparameter",
"convolution",
"manifold",
"convolutional",
"multivariate"
],
[
"other",
"analyse",
"evaluations",
"identified",
"insight",
"assessing",
"assessment",
"contrastive",
"perspectives",
"seen"
],
[
"risks",
"wide",
"vary",
"specific",
"oriented",
"width",
"potential",
"specialized",
"risk",
"connections"
],
[
"attacker",
"attacks",
"adversarially",
"adversarial",
"adversary",
"adversaries",
"security",
"exploits",
"attack",
"exploiting"
],
[
"compression",
"compress",
"prediction",
"predict",
"predicting",
"atari",
"projected",
"predicted",
"predictions",
"predicts"
],
[
"identified",
"image",
"fields",
"convex",
"subsequently",
"frequently",
"represents",
"convexity",
"grid",
"similar"
],
[
"formulate",
"models",
"modeling",
"modeled",
"structure",
"computationally",
"profiles",
"modelling",
"model",
"optimizes"
],
[
"running",
"activity",
"device",
"devices",
"wearable",
"sensor",
"sensing",
"sensors",
"activities",
"physical"
],
[
"style",
"manipulation",
"fashion",
"designs",
"art",
"manner",
"proposes",
"images",
"expressive",
"traditional"
],
[
"learning",
"faces",
"face",
"facial",
"learnt",
"training",
"practice",
"human",
"theory",
"recognition"
]
] | 324.579469 | all-MiniLM-L6-v2 | 0.916 | -0.334454 | 0.183758 | 0.888783 |
ArXiv ML Papers | S³_angular | 46 | 50 | [
[
"adversarial",
"attacks",
"adversarially",
"attacker",
"adversary",
"security",
"attack",
"exploits",
"exploiting",
"adversaries"
],
[
"projected",
"hyper",
"optimizer",
"optimizes",
"optimisation",
"hyperparameters",
"hyperparameter",
"parameterized",
"scoring",
"optimum"
],
[
"aspects",
"composed",
"scenes",
"mixture",
"partial",
"mini",
"head",
"minimax",
"aspect",
"proposals"
],
[
"labeled",
"labels",
"labeling",
"classification",
"classifier",
"classified",
"label",
"classify",
"unlabeled",
"classifying"
],
[
"superior",
"downstream",
"pipeline",
"dense",
"nas",
"flows",
"flow",
"directed",
"pipelines",
"rich"
],
[
"ai",
"adaptive",
"adaptively",
"known",
"person",
"ranking",
"predictor",
"predictors",
"estimating",
"intelligence"
],
[
"surprisingly",
"distortion",
"roughly",
"nearly",
"perturbations",
"rarely",
"almost",
"drastically",
"perturbation",
"outperform"
],
[
"subset",
"clusters",
"segmentation",
"unsupervised",
"segment",
"clustering",
"cluster",
"partitioning",
"counterparts",
"regions"
],
[
"down",
"completion",
"descent",
"fail",
"effort",
"poisoning",
"efforts",
"exhaustive",
"reflect",
"shot"
],
[
"sensor",
"event",
"outlier",
"iot",
"outliers",
"anomaly",
"monitoring",
"anomalies",
"intrusion",
"detect"
],
[
"clinical",
"medicine",
"patients",
"healthcare",
"medical",
"patient",
"disease",
"diagnosis",
"diseases",
"cancer"
],
[
"explanations",
"interpretability",
"interpretable",
"interpret",
"explaining",
"understand",
"explainable",
"mean",
"interpretation",
"understanding"
],
[
"embedded",
"query",
"views",
"include",
"densities",
"embeddings",
"separately",
"benefits",
"embedding",
"sqrt"
],
[
"pruning",
"without",
"nets",
"distinct",
"minimal",
"reduced",
"trivial",
"reduction",
"randomized",
"pruned"
],
[
"allowed",
"risks",
"penalty",
"safety",
"constraint",
"ensures",
"prevent",
"curse",
"able",
"risk"
],
[
"studying",
"study",
"learns",
"student",
"learnt",
"learners",
"learn",
"teacher",
"learner",
"learnable"
],
[
"tests",
"patches",
"testing",
"patch",
"test",
"tested",
"experimental",
"coverage",
"experimentally",
"screening"
],
[
"gnn",
"during",
"threshold",
"check",
"severe",
"consider",
"described",
"sparse",
"mini",
"modal"
],
[
"up",
"boost",
"boosting",
"collect",
"additionally",
"selected",
"aggregated",
"feature",
"actually",
"collected"
],
[
"personalized",
"profile",
"suggest",
"suggesting",
"suggests",
"recommendation",
"recommendations",
"recommender",
"factorization",
"profiles"
],
[
"uncertainties",
"accuracy",
"uncertainty",
"validated",
"validate",
"validation",
"reliability",
"accurately",
"reliable",
"rigorously"
],
[
"filtering",
"numbers",
"83",
"71",
"values",
"list",
"numerical",
"numerically",
"filters",
"filter"
],
[
"obtains",
"bandit",
"bandits",
"market",
"search",
"price",
"searching",
"seek",
"exploitation",
"finds"
],
[
"smoothness",
"robust",
"robustness",
"smoothing",
"smooth",
"focused",
"noisy",
"focus",
"focusing",
"tracking"
],
[
"systems",
"invariant",
"mathematical",
"scales",
"outputs",
"stability",
"units",
"stable",
"sublinear",
"weighted"
],
[
"cnns",
"guarantees",
"markov",
"cnn",
"bayes",
"detection",
"observing",
"ensuring",
"inferring",
"detecting"
],
[
"forecast",
"forecasts",
"weather",
"landscape",
"forest",
"nature",
"reflect",
"forecasting",
"outside",
"ground"
],
[
"matrix",
"subspace",
"subspaces",
"basis",
"matrices",
"grid",
"linear",
"represent",
"representation",
"describes"
],
[
"initialization",
"dynamically",
"cell",
"characterized",
"cnn",
"defining",
"termed",
"defines",
"monte",
"defined"
],
[
"challenging",
"difficulty",
"difficult",
"dnn",
"dnns",
"struggle",
"challenges",
"precision",
"epsilon",
"recall"
],
[
"style",
"expressive",
"variation",
"variations",
"fashion",
"art",
"manipulation",
"predefined",
"designing",
"coding"
],
[
"nlp",
"sentences",
"textual",
"texts",
"linguistic",
"corpora",
"corpus",
"text",
"english",
"translation"
],
[
"strategies",
"goal",
"video",
"on",
"twitter",
"another",
"https",
"com",
"post",
"goals"
],
[
"respect",
"capturing",
"remains",
"highlights",
"93",
"captures",
"quality",
"rl",
"71",
"free"
],
[
"granularity",
"metrics",
"similarity",
"comparable",
"distance",
"similar",
"distances",
"metric",
"space",
"meaningful"
],
[
"dimensional",
"inputs",
"representation",
"representations",
"selects",
"dimensionality",
"dimension",
"dimensions",
"retrieval",
"values"
],
[
"discrimination",
"biases",
"bias",
"discriminate",
"fairness",
"biased",
"representative",
"fair",
"candidate",
"equal"
],
[
"aid",
"involve",
"causal",
"interventions",
"involved",
"counterfactual",
"trigger",
"sources",
"associated",
"impacts"
],
[
"pose",
"assisted",
"robots",
"robotics",
"robot",
"robotic",
"poses",
"position",
"autonomous",
"machines"
],
[
"svm",
"tend",
"vanilla",
"contextual",
"raw",
"multimodal",
"spatio",
"tool",
"software",
"corpus"
],
[
"bounds",
"bound",
"complexity",
"arbitrarily",
"estimate",
"theorem",
"bounded",
"limitations",
"extent",
"asymptotically"
],
[
"98",
"fitting",
"candidate",
"candidates",
"having",
"scenario",
"idea",
"attributed",
"inferring",
"our"
],
[
"distinguishing",
"generalization",
"classifying",
"randomly",
"softmax",
"distinguish",
"learning",
"classifiers",
"generalize",
"generalizing"
],
[
"potential",
"width",
"specialized",
"oriented",
"wide",
"vary",
"specific",
"utilization",
"certain",
"near"
],
[
"speaker",
"voice",
"music",
"acoustic",
"speech",
"audio",
"sound",
"tune",
"recorded",
"performing"
],
[
"generates",
"generated",
"gan",
"gans",
"generator",
"generative",
"generation",
"generating",
"producing",
"generate"
],
[
"mcmc",
"monte",
"stochastic",
"markov",
"simulation",
"simulations",
"randomly",
"sampling",
"carlo",
"probabilistic"
],
[
"rnn",
"rnns",
"earlier",
"behind",
"backpropagation",
"predict",
"directly",
"neighboring",
"tilde",
"nearest"
],
[
"driving",
"transportation",
"traffic",
"urban",
"gps",
"transport",
"vehicle",
"vehicles",
"navigation",
"car"
],
[
"forecast",
"autoregressive",
"recurrent",
"sequences",
"sequence",
"forecasts",
"sequential",
"temporal",
"lstm",
"forecasting"
]
] | 296.18473 | all-MiniLM-L6-v2 | 0.968 | -0.330665 | 0.188685 | 0.907237 |
ArXiv ML Papers | S³_combined | 43 | 10 | [
[
"aiming",
"outperforming",
"accuracy",
"outperform",
"estimator",
"optimizer",
"screening",
"tuning",
"calibrated",
"overfitting"
],
[
"benchmarks",
"streaming",
"speedup",
"cloud",
"throughput",
"bandwidth",
"bottleneck",
"faster",
"streams",
"scalable"
],
[
"regularization",
"regularized",
"minimax",
"sparse",
"bandits",
"leveraging",
"lasso",
"minimize",
"optimal",
"optimality"
],
[
"softmax",
"rnn",
"forecast",
"forecasting",
"forecasts",
"backpropagation",
"predicts",
"prediction",
"predicting",
"predict"
],
[
"shape",
"sensing",
"discovery",
"discovering",
"lasso",
"dataset",
"discover",
"convex",
"supervised",
"optimization"
],
[
"iterative",
"selecting",
"guide",
"rapid",
"retrieval",
"planning",
"iteratively",
"completion",
"steps",
"optimizer"
],
[
"fitting",
"compression",
"dataset",
"diagnostic",
"datasets",
"denoising",
"prevalence",
"outliers",
"outlier",
"transformer"
],
[
"statistical",
"clustering",
"encountered",
"analytics",
"ranking",
"criteria",
"prevalence",
"estimating",
"popularity",
"statistics"
],
[
"annotated",
"datasets",
"annotations",
"dataset",
"predictive",
"annotation",
"automated",
"predict",
"categories",
"supervised"
],
[
"likelihood",
"simulations",
"priors",
"generative",
"variational",
"prior",
"bayesian",
"modeling",
"inference",
"monte"
]
] | 270.239487 | all-MiniLM-L6-v2 | 0.92 | -0.327841 | 0.207018 | 0.887979 |
ArXiv ML Papers | S³_combined | 44 | 10 | [
[
"models",
"forecasts",
"forecasting",
"forecast",
"generative",
"autoregressive",
"modelling",
"simulations",
"likelihood",
"rnns"
],
[
"generation",
"observing",
"automated",
"interactive",
"events",
"occur",
"planning",
"scenarios",
"outcomes",
"involve"
],
[
"predicting",
"classify",
"classifiers",
"learns",
"classifier",
"classification",
"classifying",
"prediction",
"predict",
"predictive"
],
[
"corpus",
"languages",
"textual",
"texts",
"linguistic",
"speech",
"language",
"corpora",
"sentences",
"speaker"
],
[
"reward",
"rewards",
"reinforcement",
"strategy",
"strategies",
"exploration",
"behavioral",
"control",
"critic",
"bandit"
],
[
"ranking",
"estimating",
"bandits",
"averaging",
"commonly",
"clustering",
"estimation",
"statistical",
"oracle",
"empirical"
],
[
"labeling",
"leveraging",
"retrieval",
"supervised",
"labeled",
"broadly",
"regularized",
"deep",
"bias",
"bandits"
],
[
"security",
"attacks",
"adversarial",
"adversarially",
"adversaries",
"adversary",
"attacker",
"exploits",
"privacy",
"vulnerability"
],
[
"streaming",
"faster",
"bandwidth",
"cloud",
"speedup",
"benchmarks",
"throughput",
"bottleneck",
"scalable",
"streams"
],
[
"vertices",
"nodes",
"networks",
"graphs",
"graph",
"vertex",
"node",
"relational",
"communities",
"neighbors"
]
] | 270.332015 | all-MiniLM-L6-v2 | 0.99 | -0.26402 | 0.194077 | 0.882392 |
ArXiv ML Papers | S³_combined | 45 | 10 | [
[
"rnn",
"reinforcement",
"learnt",
"learns",
"atari",
"trained",
"strategy",
"future",
"neuron",
"learned"
],
[
"classifier",
"predictors",
"predicting",
"predictive",
"supervised",
"classifying",
"classify",
"classification",
"classifiers",
"prediction"
],
[
"workflow",
"streaming",
"cloud",
"database",
"speedup",
"xgboost",
"streams",
"pipeline",
"iot",
"boost"
],
[
"algorithm",
"cluster",
"clustering",
"algorithms",
"ranking",
"selecting",
"analytics",
"criteria",
"estimating",
"methodology"
],
[
"manifold",
"dynamics",
"robotic",
"robotics",
"motion",
"recognition",
"physics",
"controller",
"objects",
"classify"
],
[
"annotation",
"annotations",
"labeled",
"reward",
"distinguishing",
"labeling",
"discriminate",
"superiority",
"demonstrate",
"rewards"
],
[
"thresholding",
"recognition",
"symmetric",
"norm",
"discriminator",
"convex",
"detection",
"convexity",
"detector",
"detecting"
],
[
"gan",
"3d",
"traffic",
"covid",
"pixel",
"infrastructure",
"believe",
"cloud",
"coverage",
"simulations"
],
[
"attacker",
"attacks",
"vulnerability",
"security",
"interpretability",
"noise",
"vulnerable",
"interpretable",
"adversaries",
"intrusion"
],
[
"communities",
"networks",
"graph",
"graphs",
"node",
"nodes",
"vertex",
"vertices",
"neighbors",
"relational"
]
] | 275.361716 | all-MiniLM-L6-v2 | 0.97 | -0.325967 | 0.170095 | 0.871481 |
ArXiv ML Papers | S³_combined | 46 | 10 | [
[
"ranking",
"estimating",
"clustering",
"prevalence",
"bandits",
"statistical",
"popularity",
"commonly",
"criteria",
"approximately"
],
[
"sensing",
"datasets",
"classification",
"discovery",
"discovering",
"dataset",
"calibration",
"shape",
"discover",
"classifier"
],
[
"completion",
"iterative",
"rapid",
"planning",
"iteratively",
"guide",
"selecting",
"steps",
"retrieval",
"unstructured"
],
[
"scalable",
"speedup",
"benchmarks",
"cloud",
"bottleneck",
"throughput",
"streaming",
"bandwidth",
"faster",
"streams"
],
[
"learns",
"reinforcement",
"reward",
"rewards",
"critic",
"strategy",
"learnt",
"bandit",
"strategies",
"exploration"
],
[
"bayesian",
"likelihood",
"generative",
"priors",
"simulations",
"modelling",
"modeling",
"inference",
"models",
"prior"
],
[
"annotations",
"annotated",
"labeled",
"labeling",
"supervised",
"predictive",
"leveraging",
"datasets",
"guided",
"annotation"
],
[
"future",
"simulation",
"development",
"workflow",
"temporal",
"modelling",
"dynamics",
"developing",
"forecast",
"behaviors"
],
[
"vision",
"capturing",
"scenes",
"3d",
"observing",
"capture",
"camera",
"scene",
"pixels",
"exhibit"
],
[
"networks",
"graphs",
"node",
"nodes",
"graph",
"vertex",
"vertices",
"communities",
"relational",
"neighbors"
]
] | 273.927626 | all-MiniLM-L6-v2 | 0.98 | -0.293128 | 0.175996 | 0.878624 |
ArXiv ML Papers | S³_combined | 43 | 20 | [
[
"outlier",
"statistically",
"asymptotically",
"overfitting",
"deviations",
"asymptotic",
"fitting",
"granularity",
"outliers",
"nonparametric"
],
[
"distillation",
"generates",
"tuning",
"ensemble",
"ensembles",
"optimizes",
"learns",
"resonance",
"learned",
"generating"
],
[
"exploit",
"adversarial",
"adversary",
"security",
"attacker",
"attacks",
"exploits",
"vulnerability",
"adversarially",
"intrusion"
],
[
"cluster",
"classifier",
"svm",
"clusters",
"randomized",
"clustering",
"classification",
"algorithm",
"segmentation",
"batch"
],
[
"builds",
"accelerators",
"throughput",
"faster",
"benchmarks",
"runtime",
"build",
"speedup",
"hardware",
"optimized"
],
[
"expectation",
"bandits",
"channels",
"greedy",
"bandit",
"hypothesis",
"regret",
"conditioning",
"seek",
"acquire"
],
[
"classifier",
"vision",
"classifiers",
"recognition",
"minimizing",
"auxiliary",
"memory",
"head",
"boosting",
"optimizer"
],
[
"annotated",
"attention",
"learns",
"learned",
"forgetting",
"leveraging",
"recall",
"collect",
"retaining",
"collected"
],
[
"records",
"log",
"metrics",
"epsilon",
"logistic",
"length",
"lstm",
"healthcare",
"minutes",
"measure"
],
[
"bayes",
"probabilistic",
"likelihood",
"stochastic",
"distributions",
"bayesian",
"densities",
"variational",
"markov",
"sampling"
],
[
"tools",
"automated",
"preprocessing",
"implementations",
"frameworks",
"interactive",
"oracle",
"modification",
"encountered",
"correspond"
],
[
"gradient",
"hessian",
"minimax",
"convexity",
"convex",
"gradients",
"sgd",
"ascent",
"convergence",
"perturbation"
],
[
"outperformed",
"hyperparameters",
"minimum",
"highest",
"bayesian",
"nearest",
"bayes",
"hyperparameter",
"maximize",
"minimize"
],
[
"coordinate",
"coordinates",
"attempts",
"hyperparameters",
"magnitude",
"parametric",
"potential",
"length",
"parameters",
"comparisons"
],
[
"nlp",
"textual",
"texts",
"sentences",
"corpus",
"corpora",
"linguistic",
"semantic",
"text",
"language"
],
[
"labeled",
"recognition",
"distinguishing",
"discriminator",
"quantization",
"supervised",
"recognizing",
"labeling",
"embeddings",
"embedding"
],
[
"patients",
"medicine",
"medical",
"healthcare",
"clinical",
"diseases",
"diagnostic",
"disease",
"cancer",
"diagnosis"
],
[
"tensor",
"sparse",
"factor",
"matrix",
"matrices",
"factorization",
"factors",
"tensors",
"subspace",
"subspaces"
],
[
"profiles",
"personalized",
"gender",
"fairness",
"preference",
"appealing",
"profile",
"discrimination",
"users",
"bias"
],
[
"intuitive",
"generalization",
"generalizing",
"discovering",
"views",
"hierarchical",
"suggesting",
"granularity",
"reasoning",
"intuition"
]
] | 369.953521 | all-MiniLM-L6-v2 | 0.95 | -0.311224 | 0.192271 | 0.914789 |
ArXiv ML Papers | S³_combined | 44 | 20 | [
[
"dimensionality",
"sensing",
"selection",
"training",
"trained",
"chosen",
"selected",
"bandwidth",
"settings",
"ensemble"
],
[
"ranging",
"capturing",
"vehicle",
"gps",
"camera",
"driving",
"vision",
"vehicles",
"tracking",
"traffic"
],
[
"attempts",
"coordinates",
"magnitude",
"hyperparameters",
"potential",
"length",
"parametric",
"parameters",
"coordinate",
"vectors"
],
[
"differentiable",
"nonlinear",
"autoencoders",
"modelling",
"modeling",
"manifold",
"expressive",
"differential",
"3d",
"formulated"
],
[
"log",
"logistic",
"records",
"metrics",
"length",
"healthcare",
"epsilon",
"lstm",
"minutes",
"measure"
],
[
"interactive",
"oracle",
"preprocessing",
"automated",
"frameworks",
"tools",
"modification",
"implementations",
"encountered",
"correspond"
],
[
"outliers",
"outlier",
"asymptotically",
"statistically",
"fitting",
"overfitting",
"asymptotic",
"deviations",
"smaller",
"granularity"
],
[
"classes",
"classifier",
"deterministic",
"classifiers",
"class",
"trees",
"accuracy",
"differentiable",
"adaptation",
"machines"
],
[
"exploit",
"adversarial",
"adversary",
"security",
"attacker",
"attacks",
"exploits",
"vulnerability",
"adversarially",
"intrusion"
],
[
"fairness",
"preference",
"users",
"appealing",
"discrimination",
"profile",
"profiles",
"personalized",
"bias",
"gender"
],
[
"forecast",
"predicted",
"weather",
"predicting",
"prediction",
"forecasts",
"autoregressive",
"forecasting",
"predict",
"regression"
],
[
"vertex",
"graph",
"graphs",
"vertices",
"nodes",
"networks",
"edges",
"node",
"edge",
"communities"
],
[
"invariance",
"approximations",
"embedding",
"dimensional",
"dimensionality",
"embeddings",
"approximated",
"approximates",
"approximating",
"learnable"
],
[
"estimation",
"projections",
"estimate",
"capability",
"confidence",
"guarantees",
"limitations",
"uncertainties",
"estimates",
"capabilities"
],
[
"convergence",
"gradient",
"minimax",
"gradients",
"convexity",
"hessian",
"convex",
"sgd",
"ascent",
"perturbation"
],
[
"superiority",
"vectors",
"linear",
"symmetric",
"naive",
"determine",
"linearly",
"matrices",
"wise",
"trained"
],
[
"nearest",
"highest",
"bayesian",
"outperformed",
"hyperparameter",
"minimum",
"hyperparameters",
"bayes",
"maximize",
"minimize"
],
[
"annotated",
"attention",
"learns",
"recall",
"leveraging",
"collect",
"retaining",
"forgetting",
"learned",
"notably"
],
[
"faster",
"benchmarks",
"throughput",
"build",
"builds",
"runtime",
"speedup",
"accelerators",
"hardware",
"implementations"
],
[
"speaker",
"frequencies",
"recorded",
"voice",
"acoustic",
"audio",
"sound",
"frequency",
"speech",
"music"
]
] | 299.589109 | all-MiniLM-L6-v2 | 0.965 | -0.314669 | 0.19289 | 0.904729 |
ArXiv ML Papers | S³_combined | 45 | 20 | [
[
"selected",
"sensing",
"dimensionality",
"chosen",
"settings",
"selection",
"ensemble",
"bandwidth",
"trained",
"dimensional"
],
[
"vehicle",
"vision",
"tracking",
"vehicles",
"capturing",
"gps",
"ranging",
"camera",
"driving",
"3d"
],
[
"statistically",
"classifiers",
"overfitting",
"outlier",
"outliers",
"asymptotic",
"asymptotically",
"fitting",
"deviations",
"nonparametric"
],
[
"insights",
"confidence",
"overhead",
"imputation",
"capability",
"guarantees",
"estimates",
"limitations",
"capabilities",
"relying"
],
[
"quantization",
"recognition",
"distinguishing",
"recognizing",
"discriminator",
"labeling",
"labeled",
"supervised",
"structured",
"unlabeled"
],
[
"seek",
"expectation",
"bandits",
"greedy",
"conditioning",
"probability",
"regret",
"hypothesis",
"bandit",
"choosing"
],
[
"bayesian",
"bayes",
"knowledge",
"inference",
"logic",
"relational",
"inferring",
"selects",
"probabilities",
"relations"
],
[
"potential",
"coordinates",
"parameters",
"parametric",
"attempts",
"hyperparameters",
"magnitude",
"length",
"vectors",
"comparisons"
],
[
"epsilon",
"metrics",
"logistic",
"healthcare",
"log",
"lstm",
"records",
"minutes",
"length",
"activity"
],
[
"gaussian",
"backpropagation",
"uncertainty",
"neurons",
"neural",
"neuron",
"networks",
"projections",
"softmax",
"uncertainties"
],
[
"factor",
"matrix",
"factorization",
"tensor",
"matrices",
"sparse",
"factors",
"tensors",
"subspaces",
"subspace"
],
[
"classifier",
"classifiers",
"auxiliary",
"minimizing",
"vision",
"optimizer",
"head",
"recognition",
"decoder",
"perceptual"
],
[
"generalizing",
"discovering",
"intuitive",
"generalization",
"suggesting",
"views",
"intuition",
"hierarchical",
"generalize",
"widespread"
],
[
"eeg",
"behaviors",
"phenomena",
"dynamics",
"brain",
"behaviour",
"cognitive",
"behavioral",
"behavior",
"phenomenon"
],
[
"posterior",
"bayesian",
"bayes",
"distributions",
"carlo",
"likelihood",
"probabilistic",
"densities",
"stochastic",
"markov"
],
[
"appealing",
"preference",
"discrimination",
"personalized",
"profiles",
"fairness",
"profile",
"bias",
"gender",
"users"
],
[
"vulnerability",
"adversarially",
"adversarial",
"adversary",
"exploits",
"attacker",
"attacks",
"exploit",
"security",
"intrusion"
],
[
"dimensional",
"approximating",
"approximates",
"invariance",
"learnable",
"imitation",
"approximations",
"embeddings",
"dimensionality",
"embedding"
],
[
"implementations",
"implementation",
"computation",
"computational",
"implementing",
"compute",
"computations",
"computationally",
"algorithms",
"algorithm"
],
[
"faster",
"builds",
"hardware",
"throughput",
"accelerators",
"gpu",
"benchmarks",
"bottleneck",
"optimized",
"gpus"
]
] | 297.369547 | all-MiniLM-L6-v2 | 0.96 | -0.329073 | 0.189117 | 0.902874 |
ArXiv ML Papers | S³_combined | 46 | 20 | [
[
"uncertainties",
"estimates",
"guarantees",
"capability",
"estimation",
"projections",
"capabilities",
"limitations",
"confidence",
"estimate"
],
[
"appealing",
"discrimination",
"gender",
"preference",
"fairness",
"profiles",
"profile",
"personalized",
"users",
"bias"
],
[
"learns",
"generates",
"ensembles",
"distillation",
"resonance",
"ensemble",
"tuning",
"optimizes",
"generating",
"learned"
],
[
"sparse",
"tensor",
"matrix",
"matrices",
"factors",
"factorization",
"factor",
"tensors",
"subspace",
"subspaces"
],
[
"forgetting",
"attention",
"annotated",
"recall",
"leveraging",
"collect",
"learns",
"learned",
"retaining",
"notably"
],
[
"networks",
"graph",
"nodes",
"vertices",
"graphs",
"node",
"edge",
"vertex",
"edges",
"communities"
],
[
"cognitive",
"completion",
"intermediate",
"planning",
"jointly",
"consist",
"interpolation",
"marginal",
"labeling",
"offline"
],
[
"decentralized",
"distributed",
"federated",
"shared",
"privacy",
"private",
"independently",
"sharing",
"cloud",
"cooperative"
],
[
"selected",
"sensing",
"selection",
"ensemble",
"dimensionality",
"chosen",
"trained",
"training",
"settings",
"bandwidth"
],
[
"convexity",
"gradient",
"hessian",
"convex",
"gradients",
"convergence",
"minimax",
"sgd",
"ascent",
"perturbation"
],
[
"cnn",
"imagenet",
"neural",
"backpropagation",
"neurons",
"neuron",
"cnns",
"rnns",
"softmax",
"convolutional"
],
[
"benchmarks",
"builds",
"accelerators",
"throughput",
"faster",
"build",
"runtime",
"hardware",
"speedup",
"implementations"
],
[
"greedy",
"regret",
"expectation",
"seek",
"bandits",
"conditioning",
"bandit",
"channels",
"acquire",
"choosing"
],
[
"multimodal",
"adaptation",
"generator",
"synthesis",
"generative",
"generating",
"generates",
"gans",
"gan",
"generated"
],
[
"behavioral",
"behaviors",
"cognitive",
"behaviour",
"eeg",
"dynamics",
"brain",
"phenomena",
"behavior",
"complexities"
],
[
"likelihood",
"stochastic",
"densities",
"bayesian",
"probabilistic",
"distributions",
"bayes",
"variational",
"carlo",
"markov"
],
[
"exploration",
"controlling",
"reinforcement",
"control",
"controller",
"controlled",
"planning",
"play",
"action",
"agents"
],
[
"clinical",
"medical",
"medicine",
"disease",
"cancer",
"diseases",
"diagnosis",
"diagnostic",
"healthcare",
"patients"
],
[
"recognizing",
"recognition",
"labeling",
"discriminator",
"distinguishing",
"quantization",
"labeled",
"supervised",
"embeddings",
"embedding"
],
[
"audio",
"recorded",
"acoustic",
"voice",
"speaker",
"sound",
"frequency",
"frequencies",
"speech",
"music"
]
] | 317.90938 | all-MiniLM-L6-v2 | 0.96 | -0.261488 | 0.206414 | 0.909145 |
ArXiv ML Papers | S³_combined | 43 | 30 | [
[
"diffusion",
"convergence",
"landscape",
"perturbations",
"perturbation",
"researchers",
"variance",
"laplacian",
"converge",
"dynamics"
],
[
"recurrent",
"sequences",
"sequence",
"lstm",
"temporal",
"periodic",
"spatiotemporal",
"rnns",
"autoregressive",
"records"
],
[
"measurement",
"measurements",
"measure",
"estimator",
"measuring",
"estimation",
"estimate",
"task",
"tasks",
"estimators"
],
[
"convexity",
"viewed",
"increasingly",
"tend",
"subsequently",
"intuitive",
"similar",
"convex",
"phenomenon",
"novel"
],
[
"fairness",
"equally",
"biases",
"biased",
"bias",
"representative",
"discriminate",
"discrimination",
"equal",
"strictly"
],
[
"weights",
"networks",
"neuron",
"evolving",
"movement",
"chain",
"neural",
"transition",
"rnn",
"flow"
],
[
"pipelines",
"forecasts",
"deep",
"forecast",
"pipeline",
"pac",
"ensembles",
"ensemble",
"allocation",
"predicts"
],
[
"overfitting",
"fitting",
"outliers",
"outlier",
"deviations",
"asymptotically",
"deviation",
"statistically",
"asymptotic",
"outperforming"
],
[
"leads",
"machine",
"learners",
"samples",
"learns",
"supervised",
"dataset",
"batch",
"datasets",
"sample"
],
[
"recognition",
"images",
"classifying",
"image",
"classification",
"imagenet",
"classifier",
"svm",
"convolutional",
"background"
],
[
"bayesian",
"likelihood",
"densities",
"bayes",
"posterior",
"gaussian",
"probabilistic",
"priors",
"distributions",
"distributional"
],
[
"quantum",
"quantized",
"ucb",
"probability",
"particle",
"compressed",
"randomly",
"carlo",
"10",
"uniformly"
],
[
"clinical",
"medical",
"patients",
"disease",
"healthcare",
"diseases",
"medicine",
"mri",
"diagnosis",
"cancer"
],
[
"feed",
"themselves",
"attention",
"effect",
"solely",
"follow",
"targeted",
"influence",
"controlling",
"effects"
],
[
"explanations",
"interpretability",
"uncertainties",
"interpretable",
"uncertainty",
"accuracy",
"reliability",
"robustness",
"interpretation",
"explaining"
],
[
"labeled",
"classifier",
"labeling",
"classifying",
"classify",
"classifiers",
"supervised",
"classification",
"labels",
"backpropagation"
],
[
"forecasting",
"forecasts",
"weather",
"forecast",
"gps",
"traffic",
"predicting",
"prediction",
"autoregressive",
"transportation"
],
[
"tensor",
"tensors",
"multivariate",
"role",
"respectively",
"vectors",
"underlying",
"representations",
"vector",
"each"
],
[
"private",
"privacy",
"public",
"accessible",
"publicly",
"federated",
"decentralized",
"independently",
"shared",
"sharing"
],
[
"information",
"hand",
"predictors",
"predictor",
"learn",
"regret",
"details",
"have",
"armed",
"https"
],
[
"learns",
"combines",
"fusion",
"centralized",
"distributed",
"collaborative",
"learner",
"learners",
"stream",
"learnt"
],
[
"greedy",
"seek",
"searching",
"bandits",
"bandit",
"sampling",
"bound",
"randomly",
"retrieval",
"expectation"
],
[
"labels",
"consist",
"optical",
"labeling",
"interpolation",
"projected",
"completion",
"compress",
"marginal",
"intermediate"
],
[
"phases",
"phase",
"classes",
"distributional",
"theory",
"notions",
"class",
"signals",
"corpus",
"arises"
],
[
"gan",
"generator",
"generating",
"gans",
"generative",
"generate",
"generates",
"generated",
"autoencoder",
"autoencoders"
],
[
"visualization",
"manipulation",
"expressive",
"process",
"interactive",
"analyse",
"developments",
"creating",
"making",
"produced"
],
[
"track",
"sensing",
"prediction",
"tracking",
"ensembles",
"music",
"ensemble",
"predicting",
"perception",
"tuning"
],
[
"ensembles",
"mixture",
"fusion",
"field",
"combining",
"normalizing",
"head",
"ensemble",
"optimality",
"minimizing"
],
[
"sgd",
"hessian",
"convexity",
"convex",
"minimize",
"minimization",
"gradient",
"gradients",
"optimization",
"minimax"
],
[
"gps",
"camera",
"track",
"tracking",
"ranging",
"3d",
"capturing",
"highlights",
"poses",
"pose"
]
] | 333.277372 | all-MiniLM-L6-v2 | 0.916667 | -0.305697 | 0.185225 | 0.88857 |
ArXiv ML Papers | S³_combined | 44 | 30 | [
[
"domains",
"molecular",
"encodes",
"encoded",
"encode",
"encoder",
"encoding",
"protein",
"decoding",
"autoencoder"
],
[
"predictor",
"dependent",
"hand",
"fidelity",
"https",
"longer",
"via",
"information",
"future",
"cross"
],
[
"bounds",
"approximates",
"thresholding",
"threshold",
"approximation",
"approximations",
"arbitrarily",
"optimality",
"approximating",
"approximated"
],
[
"normalizing",
"mixture",
"ensembles",
"minimizing",
"optimizer",
"field",
"fusion",
"combining",
"head",
"optimality"
],
[
"tracking",
"vision",
"ranging",
"pose",
"poses",
"3d",
"camera",
"scenes",
"perspective",
"moving"
],
[
"tend",
"sequential",
"proportional",
"subsequent",
"towards",
"forward",
"chain",
"toward",
"target",
"transferred"
],
[
"independently",
"privacy",
"federated",
"publicly",
"accessible",
"public",
"decentralized",
"private",
"alone",
"sharing"
],
[
"datasets",
"quantum",
"discovering",
"dimensionality",
"discovery",
"dataset",
"collection",
"sampling",
"selects",
"carlo"
],
[
"logic",
"relational",
"constraints",
"constraint",
"constrained",
"conditional",
"probabilistic",
"inference",
"relations",
"reasoning"
],
[
"patients",
"diseases",
"healthcare",
"clinical",
"medical",
"diagnosis",
"medicine",
"disease",
"mri",
"diagnostic"
],
[
"sensing",
"ensembles",
"track",
"tracking",
"music",
"ensemble",
"prediction",
"filters",
"sampled",
"tuning"
],
[
"overfitting",
"fitting",
"outlier",
"asymptotically",
"outliers",
"bottleneck",
"cnns",
"convolutions",
"convolution",
"throughput"
],
[
"classifying",
"svm",
"classifier",
"recognition",
"classification",
"images",
"classifiers",
"classify",
"imagenet",
"classified"
],
[
"tree",
"classifiers",
"classifier",
"reliable",
"forest",
"interpretability",
"accuracy",
"outperform",
"outperforming",
"outperformed"
],
[
"generator",
"gan",
"generating",
"generative",
"gans",
"generate",
"generates",
"generated",
"generation",
"autoencoders"
],
[
"nonlinear",
"derivatives",
"hessian",
"differentiation",
"differential",
"differentiable",
"gradient",
"backpropagation",
"calibration",
"perturbed"
],
[
"area",
"traffic",
"urban",
"bandwidth",
"distance",
"iot",
"planning",
"driving",
"transportation",
"city"
],
[
"activity",
"wearable",
"eeg",
"behavioral",
"signals",
"behaviors",
"device",
"kernel",
"devices",
"signal"
],
[
"bayesian",
"likelihood",
"posterior",
"bayes",
"distributions",
"probabilistic",
"densities",
"distributional",
"mcmc",
"stochastic"
],
[
"exploit",
"exploits",
"vulnerability",
"attacks",
"adversarially",
"security",
"adversarial",
"intrusion",
"adversary",
"attacker"
],
[
"supervised",
"labeled",
"labeling",
"labels",
"label",
"unlabeled",
"classifiers",
"classification",
"classifier",
"classifying"
],
[
"manipulation",
"developments",
"analyse",
"technologies",
"research",
"visualization",
"creating",
"interactive",
"expressive",
"process"
],
[
"similarity",
"metric",
"distance",
"distances",
"metrics",
"corresponds",
"dimensional",
"dimension",
"dimensionality",
"dimensions"
],
[
"rnns",
"recurrent",
"records",
"record",
"sequences",
"temporal",
"occurrence",
"events",
"sequence",
"lstm"
],
[
"convexity",
"emerging",
"viewed",
"similar",
"tend",
"increasingly",
"subsequently",
"novel",
"convex",
"intuitive"
],
[
"navigation",
"iterations",
"reconstructing",
"forgetting",
"discover",
"autoencoder",
"iteration",
"iteratively",
"iterative",
"exploration"
],
[
"predicting",
"weather",
"forecasts",
"predict",
"forecast",
"forecasting",
"autoregressive",
"prediction",
"predicted",
"predicts"
],
[
"neural",
"generalizing",
"perception",
"intuitive",
"intuition",
"explanations",
"cognitive",
"brain",
"neuron",
"neurons"
],
[
"generalizes",
"learns",
"supervised",
"generalization",
"datasets",
"algorithms",
"generalize",
"dataset",
"ai",
"hyperparameters"
],
[
"corpus",
"phases",
"inducing",
"class",
"transitions",
"phase",
"classes",
"distributional",
"distributions",
"theory"
]
] | 237.027429 | all-MiniLM-L6-v2 | 0.93 | -0.285436 | 0.206243 | 0.902851 |
ArXiv ML Papers | S³_combined | 45 | 30 | [
[
"mechanisms",
"labeled",
"unlabeled",
"labeling",
"structured",
"experiments",
"annotation",
"imitation",
"contained",
"quantization"
],
[
"sequences",
"sequential",
"sequence",
"attention",
"recurrent",
"rnns",
"lstm",
"rnn",
"encode",
"minimax"
],
[
"bandits",
"retrieval",
"seek",
"bandit",
"expectation",
"coverage",
"searching",
"sampling",
"bound",
"considerable"
],
[
"fairness",
"biases",
"representative",
"bias",
"discrimination",
"discriminate",
"heterogeneity",
"biased",
"appealing",
"equal"
],
[
"classifiers",
"supervised",
"classifier",
"classifying",
"classification",
"labeled",
"classify",
"labeling",
"regularization",
"backpropagation"
],
[
"curse",
"emerging",
"boosting",
"boost",
"armed",
"identified",
"forest",
"xgboost",
"covid",
"recall"
],
[
"across",
"transfer",
"transferring",
"transferred",
"adaptation",
"transformation",
"leveraging",
"stream",
"target",
"combines"
],
[
"planning",
"urban",
"area",
"iot",
"traffic",
"bandwidth",
"autoencoder",
"bounded",
"distance",
"city"
],
[
"convexity",
"massive",
"tend",
"subsequently",
"related",
"emerging",
"viewed",
"similar",
"intuitive",
"convex"
],
[
"energy",
"generating",
"generates",
"generator",
"distillation",
"compressed",
"mining",
"resonance",
"imbalanced",
"optimizes"
],
[
"periodic",
"events",
"record",
"temporal",
"occurrence",
"records",
"measure",
"recurrent",
"measures",
"history"
],
[
"ensembles",
"mixture",
"field",
"minimizing",
"ensemble",
"normalizing",
"decoder",
"fusion",
"head",
"mini"
],
[
"similarity",
"dimensionality",
"embeddings",
"metrics",
"distances",
"distance",
"metric",
"dimension",
"dimensional",
"embedding"
],
[
"privacy",
"federated",
"distributed",
"private",
"decentralized",
"shared",
"independently",
"sharing",
"public",
"centralized"
],
[
"samples",
"dataset",
"supervised",
"machine",
"learns",
"learners",
"datasets",
"sample",
"machines",
"batch"
],
[
"gans",
"gan",
"generative",
"generating",
"generate",
"generation",
"generates",
"generated",
"autoencoders",
"generator"
],
[
"cnn",
"assessment",
"learned",
"reliability",
"trained",
"training",
"augment",
"grounded",
"gnn",
"dnn"
],
[
"segmentation",
"aggregation",
"partition",
"clusters",
"clustering",
"cluster",
"partitioning",
"aggregated",
"instances",
"group"
],
[
"invariant",
"symmetric",
"variational",
"notions",
"quantization",
"invariance",
"covariance",
"static",
"dependencies",
"arbitrary"
],
[
"branch",
"ensemble",
"logistic",
"evolving",
"transition",
"classifier",
"classify",
"mcmc",
"classifiers",
"probabilistic"
],
[
"outliers",
"fitting",
"asymptotically",
"outlier",
"overfitting",
"approximations",
"deviations",
"statistically",
"asymptotic",
"bottleneck"
],
[
"svm",
"convolution",
"image",
"images",
"convolutional",
"recognition",
"imagenet",
"classification",
"art",
"gaussian"
],
[
"pipeline",
"dense",
"deep",
"pipelines",
"pac",
"forecast",
"forecasts",
"horizon",
"deeper",
"allocation"
],
[
"selects",
"quantum",
"datasets",
"discovery",
"discovering",
"sampling",
"carlo",
"frequencies",
"dimensionality",
"probabilities"
],
[
"consist",
"optical",
"compress",
"interpolation",
"completion",
"intermediate",
"labeling",
"collaborative",
"labels",
"dedicated"
],
[
"perturbations",
"diffusion",
"converges",
"convergence",
"converge",
"simulations",
"variances",
"researchers",
"variance",
"perturbation"
],
[
"explanations",
"brain",
"cognitive",
"perception",
"generalizing",
"neural",
"abstraction",
"intuition",
"intuitive",
"generalization"
],
[
"posterior",
"distributional",
"bayesian",
"likelihood",
"densities",
"gaussian",
"distributions",
"variational",
"bayes",
"distribution"
],
[
"demonstrating",
"distributional",
"classes",
"phase",
"phases",
"class",
"distributions",
"inducing",
"corpus",
"learners"
],
[
"ensembles",
"selection",
"tuning",
"ensemble",
"boosting",
"prediction",
"predicting",
"optimizing",
"track",
"sensing"
]
] | 290.914814 | all-MiniLM-L6-v2 | 0.906667 | -0.315412 | 0.172839 | 0.89272 |
ArXiv ML Papers | S³_combined | 46 | 30 | [
[
"iteratively",
"iteration",
"forgetting",
"iterations",
"iterative",
"incremental",
"discover",
"learns",
"posterior",
"navigation"
],
[
"superiority",
"enhancing",
"calibration",
"enhance",
"deployment",
"configuration",
"configurations",
"enhances",
"build",
"enable"
],
[
"perturbations",
"converges",
"converge",
"convergence",
"perturbation",
"researchers",
"laplacian",
"simulations",
"diffusion",
"variance"
],
[
"reward",
"rewards",
"play",
"exploration",
"reinforcement",
"controller",
"game",
"agents",
"games",
"strategy"
],
[
"nlp",
"corpora",
"textual",
"texts",
"linguistic",
"text",
"sentences",
"semantic",
"corpus",
"literature"
],
[
"parameters",
"coordinates",
"parameter",
"parametric",
"optimum",
"curve",
"theorem",
"hyperparameter",
"neuron",
"hyperparameters"
],
[
"images",
"convolution",
"convolutional",
"svm",
"recognition",
"image",
"imagenet",
"art",
"convolutions",
"tensor"
],
[
"sparse",
"resolution",
"cover",
"matrix",
"matrices",
"solves",
"subspaces",
"denoising",
"subspace",
"imaging"
],
[
"labeling",
"classifiers",
"classifier",
"regularization",
"supervised",
"backpropagation",
"labeled",
"classifying",
"classification",
"classify"
],
[
"generative",
"gans",
"gan",
"generation",
"generator",
"autoencoders",
"generate",
"generated",
"generates",
"generating"
],
[
"nodes",
"vertex",
"vertices",
"graphs",
"graph",
"edges",
"edge",
"networks",
"trees",
"node"
],
[
"labels",
"compress",
"interpolation",
"completion",
"sampled",
"consist",
"labeling",
"intermediate",
"meaningful",
"optical"
],
[
"information",
"predictor",
"hand",
"own",
"estimates",
"predictors",
"longer",
"policy",
"https",
"publicly"
],
[
"validity",
"monitoring",
"reliability",
"uncertain",
"reliable",
"uncertainty",
"qualitative",
"uncertainties",
"fail",
"expert"
],
[
"learns",
"machine",
"learners",
"supervised",
"dataset",
"datasets",
"samples",
"learning",
"batch",
"training"
],
[
"densities",
"likelihood",
"bayesian",
"distributions",
"gaussian",
"posterior",
"distributional",
"variational",
"bayes",
"distribution"
],
[
"signal",
"eeg",
"device",
"devices",
"signals",
"activation",
"behavioral",
"activity",
"wearable",
"behaviors"
],
[
"audio",
"voice",
"acoustic",
"speech",
"music",
"speaker",
"sound",
"noise",
"frequencies",
"spectrum"
],
[
"convex",
"convexity",
"optimization",
"gradients",
"hessian",
"minimization",
"minimize",
"optimizer",
"gradient",
"optimisation"
],
[
"dimensionality",
"embeddings",
"similarity",
"metrics",
"distances",
"metric",
"distance",
"dimension",
"dimensional",
"embedding"
],
[
"brain",
"intuition",
"cognitive",
"generalizing",
"intuitive",
"perception",
"abstraction",
"explanations",
"neural",
"generalization"
],
[
"distributional",
"corpus",
"transient",
"phase",
"phases",
"inducing",
"classes",
"class",
"distributions",
"learners"
],
[
"privacy",
"shared",
"distributed",
"independently",
"decentralized",
"private",
"federated",
"public",
"centralized",
"sharing"
],
[
"tracking",
"ranging",
"coordinates",
"vision",
"3d",
"gps",
"robotics",
"track",
"motion",
"pose"
],
[
"weather",
"forecasts",
"forecast",
"forecasting",
"predicting",
"autoregressive",
"prediction",
"predict",
"predicted",
"predicts"
],
[
"approximating",
"thresholding",
"threshold",
"interpretability",
"quantized",
"approximates",
"approximations",
"interpretable",
"approximation",
"trees"
],
[
"disease",
"clinical",
"diseases",
"medical",
"diagnosis",
"healthcare",
"mri",
"patients",
"medicine",
"cancer"
],
[
"bandits",
"retrieval",
"seek",
"sampling",
"expectation",
"searching",
"coverage",
"bandit",
"greedy",
"bound"
],
[
"area",
"iot",
"traffic",
"urban",
"bandwidth",
"transportation",
"planning",
"distance",
"city",
"driving"
],
[
"synthesis",
"computationally",
"computational",
"algorithm",
"algorithmic",
"algorithms",
"deterministic",
"visualization",
"preprocessing",
"manipulation"
]
] | 334.535942 | all-MiniLM-L6-v2 | 0.963333 | -0.25198 | 0.216324 | 0.899347 |
ArXiv ML Papers | S³_combined | 43 | 40 | [
[
"learnt",
"teacher",
"revisit",
"trained",
"studying",
"learnable",
"learner",
"student",
"learners",
"learns"
],
[
"exploits",
"exploit",
"attack",
"attacker",
"attacks",
"security",
"adversary",
"adversarial",
"intrusion",
"adversarially"
],
[
"weaker",
"enhance",
"enhanced",
"enhances",
"stronger",
"exploring",
"superiority",
"advanced",
"sensitivity",
"susceptible"
],
[
"randomized",
"similarity",
"encode",
"algorithm",
"classifier",
"ranking",
"accuracy",
"classifiers",
"encoder",
"entropy"
],
[
"convex",
"tend",
"grid",
"neurons",
"convexity",
"represents",
"similar",
"viewed",
"fields",
"attribute"
],
[
"iterations",
"discover",
"iteration",
"inform",
"dynamically",
"build",
"generate",
"iteratively",
"without",
"guide"
],
[
"texts",
"textual",
"literature",
"nlp",
"text",
"semantic",
"linguistic",
"document",
"corpus",
"dictionary"
],
[
"ai",
"robot",
"tasks",
"robots",
"machines",
"robotic",
"automated",
"machine",
"task",
"intelligence"
],
[
"represent",
"represents",
"covid",
"interpretable",
"interpret",
"representing",
"represented",
"environments",
"runtime",
"embeddings"
],
[
"medicine",
"diseases",
"clinical",
"patients",
"disease",
"healthcare",
"medical",
"health",
"covid",
"patient"
],
[
"follow",
"principles",
"consensus",
"publicly",
"https",
"determine",
"significant",
"principle",
"public",
"reach"
],
[
"sequences",
"trick",
"approximating",
"patterns",
"smoothing",
"rigorous",
"smooth",
"regularized",
"pseudo",
"accuracies"
],
[
"seek",
"choosing",
"bandits",
"bandit",
"greedy",
"search",
"searching",
"retrieval",
"strategy",
"optimal"
],
[
"reconstructing",
"reconstructed",
"denoising",
"imputation",
"sparse",
"compressed",
"decoding",
"reconstruction",
"matrix",
"resolution"
],
[
"imbalance",
"vector",
"fail",
"vectors",
"acceleration",
"magnitude",
"efforts",
"inequality",
"attempts",
"aims"
],
[
"unsupervised",
"probabilistic",
"convolutions",
"representations",
"bernoulli",
"recognition",
"cnn",
"poses",
"imagenet",
"convolution"
],
[
"guarantee",
"guarantees",
"reliability",
"bayes",
"probabilistic",
"risk",
"markov",
"discriminator",
"safety",
"recall"
],
[
"recognize",
"face",
"facial",
"recognizing",
"faces",
"contrast",
"present",
"recognition",
"identified",
"identifying"
],
[
"gan",
"gans",
"generates",
"generating",
"generation",
"generative",
"generator",
"generate",
"producing",
"generated"
],
[
"cluster",
"clustering",
"segmentation",
"clusters",
"unsupervised",
"communities",
"partitioning",
"phases",
"regions",
"neighborhood"
],
[
"respectively",
"tensor",
"vectors",
"tensors",
"multivariate",
"vector",
"correlated",
"parallel",
"dimensionality",
"manner"
],
[
"eeg",
"devices",
"wearable",
"device",
"tested",
"activity",
"physical",
"body",
"running",
"arm"
],
[
"neurons",
"neuron",
"cognitive",
"intelligence",
"intuition",
"neural",
"brain",
"eeg",
"explanations",
"perception"
],
[
"explored",
"benchmarks",
"discover",
"tests",
"benchmark",
"deep",
"challenging",
"insights",
"coverage",
"metrics"
],
[
"combine",
"manually",
"hybrid",
"ensemble",
"split",
"combination",
"combining",
"combines",
"fusion",
"combinations"
],
[
"label",
"supervised",
"classifier",
"classification",
"classifying",
"classifiers",
"labels",
"labeling",
"labeled",
"unlabeled"
],
[
"qualitative",
"rigorous",
"digital",
"empirically",
"uncertainty",
"empirical",
"definition",
"flows",
"naive",
"rigorously"
],
[
"segmentation",
"extraction",
"imaging",
"segment",
"region",
"regions",
"stages",
"mri",
"screening",
"diagnosis"
],
[
"characteristic",
"dimensionality",
"dimensions",
"dimensional",
"dimension",
"depth",
"characteristics",
"sampling",
"determined",
"combinatorial"
],
[
"weather",
"forecast",
"forecasting",
"forecasts",
"predicted",
"predictions",
"predict",
"predicts",
"autoregressive",
"predicting"
],
[
"structures",
"theoretical",
"corpora",
"distributional",
"theory",
"deterministic",
"corpus",
"class",
"practice",
"speech"
],
[
"resonance",
"fourier",
"particle",
"quantization",
"observable",
"quantized",
"quantum",
"physics",
"accelerators",
"subspace"
],
[
"graphs",
"graph",
"vertex",
"edges",
"nodes",
"vertices",
"edge",
"node",
"networks",
"spectral"
],
[
"predict",
"confidence",
"predicted",
"predictive",
"accuracy",
"predicts",
"calibration",
"trust",
"calibrated",
"predictions"
],
[
"navigation",
"gps",
"vehicle",
"transportation",
"vehicles",
"drive",
"driving",
"transport",
"traffic",
"car"
],
[
"asymptotically",
"asymptotic",
"fitting",
"latency",
"estimation",
"throughput",
"98",
"likelihood",
"blind",
"measuring"
],
[
"abstraction",
"constraint",
"solvers",
"solver",
"symbolic",
"logic",
"semantic",
"diagnostic",
"relational",
"reasoning"
],
[
"linearly",
"euclidean",
"constrain",
"constraints",
"constraint",
"constrained",
"dual",
"symmetric",
"dimension",
"optimality"
],
[
"trees",
"thresholding",
"tree",
"softmax",
"threshold",
"minimum",
"approximations",
"minimal",
"approximation",
"interpretability"
],
[
"classification",
"boost",
"boosting",
"sampling",
"samples",
"popular",
"sample",
"actively",
"collect",
"dataset"
]
] | 297.297021 | all-MiniLM-L6-v2 | 0.9275 | -0.305678 | 0.21308 | 0.905852 |
ArXiv ML Papers | S³_combined | 44 | 40 | [
[
"risk",
"positive",
"reliability",
"safety",
"guarantees",
"probabilistic",
"risks",
"covariance",
"supervision",
"markov"
],
[
"regularized",
"tend",
"batch",
"parametric",
"distributions",
"distribution",
"uniformly",
"regular",
"standard",
"preferences"
],
[
"combine",
"combines",
"dual",
"fusion",
"combination",
"combining",
"multi",
"simultaneously",
"combinations",
"combined"
],
[
"testing",
"development",
"developers",
"patches",
"inspection",
"test",
"tests",
"tested",
"experimental",
"patch"
],
[
"fail",
"vectors",
"vector",
"imbalance",
"acceleration",
"causes",
"magnitude",
"accelerate",
"attempts",
"decreases"
],
[
"graph",
"graphs",
"edge",
"node",
"vertices",
"edges",
"spectral",
"vertex",
"nodes",
"networks"
],
[
"tensors",
"tensor",
"scales",
"leads",
"oriented",
"lead",
"manner",
"robotic",
"vertices",
"accelerators"
],
[
"deterministic",
"paradigm",
"theory",
"theoretical",
"corpora",
"corpus",
"parallel",
"biological",
"hypothesis",
"complexity"
],
[
"intuition",
"cognitive",
"eeg",
"brain",
"neuron",
"neurons",
"neural",
"paths",
"propagation",
"intuitive"
],
[
"optimal",
"predictions",
"predicting",
"predict",
"algorithms",
"algorithm",
"predicts",
"optimizing",
"optimizer",
"optimization"
],
[
"curse",
"fidelity",
"ensures",
"characteristic",
"transferring",
"characteristics",
"factors",
"character",
"transferred",
"tractable"
],
[
"gps",
"traffic",
"navigation",
"driving",
"transportation",
"vehicle",
"transport",
"drive",
"vehicles",
"trajectory"
],
[
"feature",
"lasso",
"generalized",
"features",
"sparsity",
"extent",
"bounded",
"alleviate",
"norm",
"trivial"
],
[
"autoencoder",
"encode",
"encodes",
"encoders",
"trainable",
"encoder",
"modelling",
"autoencoders",
"model",
"models"
],
[
"diffusion",
"densities",
"distribution",
"distributions",
"laplacian",
"gaussian",
"likelihood",
"distributional",
"processes",
"covariance"
],
[
"maintaining",
"consist",
"marginal",
"interpolation",
"consistent",
"structured",
"meaningful",
"resolution",
"compress",
"completion"
],
[
"perturbations",
"stability",
"stable",
"variance",
"perturbation",
"perturbed",
"predicts",
"imbalanced",
"catastrophic",
"coupled"
],
[
"generate",
"generation",
"producing",
"generator",
"generative",
"generating",
"generates",
"gan",
"gans",
"generated"
],
[
"experimental",
"tested",
"millions",
"gan",
"confidence",
"researchers",
"experimentally",
"accuracy",
"findings",
"infer"
],
[
"truncated",
"rigorous",
"stored",
"generate",
"unstructured",
"storing",
"compressed",
"embeddings",
"embedding",
"consist"
],
[
"autoencoder",
"autoencoders",
"algorithm",
"algorithmic",
"personalized",
"model",
"algorithms",
"models",
"automated",
"modeled"
],
[
"physics",
"in",
"fourier",
"resonance",
"quantum",
"quantized",
"quantization",
"observable",
"particle",
"showed"
],
[
"biased",
"discriminate",
"fairness",
"biases",
"discrimination",
"bias",
"equal",
"representative",
"appealing",
"unbiased"
],
[
"greedy",
"market",
"bandit",
"search",
"searching",
"exploration",
"seek",
"strategy",
"bandits",
"combinatorial"
],
[
"tend",
"neurons",
"convex",
"represents",
"convexity",
"fields",
"similar",
"grid",
"viewed",
"field"
],
[
"notions",
"invariance",
"imitation",
"abstraction",
"dependencies",
"covariance",
"signals",
"perceptual",
"behaviors",
"aspect"
],
[
"factorization",
"decomposition",
"matrices",
"tensor",
"tensors",
"matrix",
"denoising",
"mri",
"imputation",
"basis"
],
[
"compressed",
"severe",
"insufficient",
"sparse",
"substantially",
"dramatically",
"drastically",
"minimal",
"variational",
"severely"
],
[
"depth",
"representation",
"characteristic",
"representations",
"dimension",
"dimensional",
"dimensionality",
"dimensions",
"characteristics",
"transformers"
],
[
"conventional",
"classifying",
"priors",
"softmax",
"classify",
"classifiers",
"distinguishing",
"classifier",
"hessian",
"bayesian"
],
[
"weather",
"forecasts",
"forecasting",
"forecast",
"predicted",
"prediction",
"predicting",
"predict",
"predicts",
"predictions"
],
[
"videos",
"music",
"video",
"convolutions",
"decisions",
"classification",
"audio",
"visualization",
"quantitative",
"features"
],
[
"visual",
"dependent",
"uncertainties",
"blind",
"passing",
"authors",
"information",
"transforms",
"cross",
"rather"
],
[
"dense",
"resnet",
"downstream",
"decentralized",
"cnn",
"pipeline",
"pipelines",
"acquire",
"cnns",
"spread"
],
[
"multivariate",
"covariates",
"jointly",
"multi",
"width",
"vectors",
"network",
"networks",
"convex",
"components"
],
[
"reinforcement",
"behaviors",
"action",
"behavioral",
"actions",
"policy",
"enforce",
"rewards",
"policies",
"interventions"
],
[
"randomized",
"recognition",
"gaussian",
"accuracy",
"normalized",
"measure",
"spectral",
"classifier",
"encode",
"softmax"
],
[
"distances",
"similarity",
"distance",
"nearest",
"ranking",
"comparable",
"metrics",
"compares",
"similarities",
"metric"
],
[
"reliability",
"processes",
"uncertainties",
"measurement",
"uncertainty",
"task",
"tasks",
"machines",
"ensuring",
"projection"
],
[
"sequences",
"recurrent",
"temporal",
"sequence",
"lstm",
"autoregressive",
"rnns",
"sequential",
"rnn",
"forecasting"
]
] | 348.466742 | all-MiniLM-L6-v2 | 0.89 | -0.318662 | 0.207202 | 0.893793 |
ArXiv ML Papers | S³_combined | 45 | 40 | [
[
"bias",
"biases",
"discrimination",
"unbiased",
"fairness",
"discriminate",
"representative",
"equal",
"biased",
"assess"
],
[
"vectors",
"decomposition",
"matrix",
"vector",
"tensors",
"matrices",
"tensor",
"factorization",
"3d",
"multivariate"
],
[
"uncertainty",
"reliability",
"consistent",
"uncertainties",
"accuracy",
"reliable",
"consistency",
"precision",
"consistently",
"uncertain"
],
[
"intrusion",
"adversary",
"security",
"adversarially",
"adversarial",
"exploit",
"exploits",
"attack",
"attacker",
"attacks"
],
[
"samples",
"benchmark",
"dataset",
"datasets",
"finds",
"tested",
"sample",
"screening",
"batch",
"discover"
],
[
"temporal",
"rnns",
"autoregressive",
"sequence",
"sequences",
"lstm",
"recurrent",
"rnn",
"sequential",
"forecasting"
],
[
"marginal",
"potential",
"scalability",
"shape",
"scalable",
"oriented",
"flexibility",
"jointly",
"multivariate",
"risk"
],
[
"estimates",
"perturbations",
"dimensionality",
"laplacian",
"perturbation",
"gaussian",
"kernels",
"landscape",
"spectral",
"covariance"
],
[
"forecast",
"forecasts",
"forecasting",
"weather",
"predicts",
"predicted",
"predict",
"predictions",
"predicting",
"prediction"
],
[
"algorithm",
"randomized",
"predict",
"predicts",
"predicting",
"predictive",
"prediction",
"encode",
"predictions",
"optimal"
],
[
"needs",
"estimate",
"embeddings",
"estimates",
"embedding",
"description",
"solvers",
"estimation",
"estimator",
"estimators"
],
[
"perceptual",
"covariance",
"invariance",
"signals",
"dependencies",
"notions",
"perception",
"abstraction",
"behavioral",
"imitation"
],
[
"goals",
"fidelity",
"projection",
"objectives",
"projections",
"multivariate",
"capabilities",
"mri",
"incorporate",
"objective"
],
[
"appealing",
"recommendations",
"popularity",
"recommendation",
"recommender",
"personalized",
"factorization",
"ranking",
"rank",
"suggests"
],
[
"mnist",
"poses",
"convolutions",
"unsupervised",
"representations",
"imagenet",
"cnn",
"cnns",
"bernoulli",
"pose"
],
[
"simulation",
"mcmc",
"carlo",
"bayesian",
"markov",
"monte",
"probabilistic",
"posterior",
"simulations",
"stochastic"
],
[
"rigorous",
"differentially",
"smoothing",
"modeling",
"dynamical",
"nonlinear",
"modeled",
"dynamics",
"manifold",
"smoothness"
],
[
"speaker",
"audio",
"sound",
"speech",
"voice",
"acoustic",
"music",
"frequencies",
"tune",
"interplay"
],
[
"normalizing",
"backpropagation",
"normalized",
"normalization",
"regularization",
"regularized",
"kernel",
"transformation",
"factorization",
"invariant"
],
[
"vertex",
"graph",
"graphs",
"vertices",
"nodes",
"edge",
"node",
"edges",
"networks",
"spectral"
],
[
"sufficiently",
"substantially",
"insufficient",
"gnn",
"considerable",
"approximates",
"dramatically",
"minimal",
"sqrt",
"enormous"
],
[
"edges",
"segment",
"periods",
"stage",
"stages",
"segmentation",
"extraction",
"inspection",
"period",
"labeling"
],
[
"attempts",
"efforts",
"aims",
"fail",
"magnitude",
"imbalance",
"accelerate",
"vectors",
"vector",
"acceleration"
],
[
"explanations",
"interpretable",
"explaining",
"interpretability",
"interpret",
"interpretation",
"visual",
"describes",
"describe",
"understanding"
],
[
"pipelines",
"pipeline",
"spread",
"resnet",
"filters",
"decentralized",
"filtering",
"dense",
"filter",
"rich"
],
[
"quantum",
"quantization",
"resonance",
"fourier",
"particle",
"physics",
"quantized",
"observable",
"discovered",
"in"
],
[
"individually",
"diverse",
"diversity",
"ensemble",
"weakly",
"weak",
"broadly",
"combines",
"experimental",
"ensembles"
],
[
"classification",
"videos",
"decisions",
"visualization",
"music",
"video",
"digital",
"audio",
"probabilistic",
"quantitative"
],
[
"robot",
"robots",
"machines",
"ai",
"robotic",
"autonomous",
"tasks",
"task",
"machine",
"automated"
],
[
"dependent",
"transformation",
"transforms",
"information",
"derive",
"transform",
"rather",
"passing",
"encoded",
"blind"
],
[
"decompose",
"processes",
"weights",
"generator",
"process",
"kernels",
"decomposition",
"core",
"machines",
"softmax"
],
[
"vertex",
"attention",
"ranging",
"rank",
"trainable",
"rnn",
"rnns",
"ranking",
"boosting",
"inputs"
],
[
"reach",
"measurement",
"measuring",
"offs",
"examine",
"scores",
"scoring",
"goals",
"lead",
"measured"
],
[
"auc",
"system",
"ucb",
"languages",
"sensors",
"basic",
"classified",
"detect",
"sparse",
"systems"
],
[
"conventional",
"gradient",
"differentiation",
"gradients",
"softmax",
"nonparametric",
"hessian",
"priors",
"contrastive",
"derivatives"
],
[
"latency",
"98",
"asymptotically",
"asymptotic",
"fitting",
"likelihood",
"bottleneck",
"throughput",
"ml",
"measuring"
],
[
"collected",
"sampling",
"dimensionality",
"characteristic",
"dimensional",
"depth",
"dimensions",
"dimension",
"representations",
"representation"
],
[
"spectral",
"encode",
"measure",
"classifier",
"classifiers",
"ranking",
"softmax",
"algorithm",
"encoder",
"randomized"
],
[
"brain",
"neural",
"analytically",
"intuition",
"neuron",
"eeg",
"volume",
"neurons",
"mri",
"cognitive"
],
[
"simulation",
"suited",
"networks",
"modeled",
"modeling",
"model",
"modelling",
"calibration",
"models",
"ideal"
]
] | 296.637324 | all-MiniLM-L6-v2 | 0.89 | -0.321696 | 0.213028 | 0.902201 |
ArXiv ML Papers | S³_combined | 46 | 40 | [
[
"extracts",
"attention",
"mainly",
"interest",
"traffic",
"natural",
"boost",
"giving",
"boosting",
"beneficial"
],
[
"visual",
"cross",
"blind",
"rather",
"dependent",
"derive",
"passing",
"information",
"deeper",
"pass"
],
[
"stages",
"extraction",
"segment",
"imaging",
"mri",
"segmentation",
"region",
"screening",
"cancer",
"regions"
],
[
"generalize",
"multimodal",
"granularity",
"generalizing",
"parameterized",
"generalizes",
"generalization",
"incremental",
"automated",
"procedures"
],
[
"audio",
"decisions",
"visualization",
"videos",
"music",
"classification",
"digital",
"probabilistic",
"video",
"classifier"
],
[
"contexts",
"contextual",
"context",
"pixels",
"scenes",
"grounded",
"convolutions",
"spectral",
"distinguish",
"parts"
],
[
"equations",
"simulations",
"solvers",
"analytical",
"nonlinear",
"analyze",
"reconstruction",
"reconstructed",
"denoising",
"dynamics"
],
[
"energy",
"yielding",
"tail",
"bandwidth",
"evolution",
"bounded",
"resnet",
"evolving",
"continuously",
"consumption"
],
[
"ranging",
"attention",
"rnn",
"rank",
"trainable",
"autoencoder",
"rnns",
"vertex",
"gps",
"autoencoders"
],
[
"mnist",
"cnn",
"convolutions",
"probabilistic",
"imagenet",
"networks",
"unbiased",
"off",
"cnns",
"backpropagation"
],
[
"marginal",
"completion",
"interpolation",
"consistent",
"averaging",
"outcomes",
"consist",
"meaningful",
"compress",
"maintaining"
],
[
"gans",
"generates",
"generating",
"generation",
"producing",
"generative",
"generator",
"gan",
"generate",
"generated"
],
[
"multiple",
"individually",
"ensemble",
"diversity",
"diverse",
"mixed",
"combines",
"multi",
"broadly",
"heterogeneity"
],
[
"sequences",
"recurrent",
"temporal",
"lstm",
"rnn",
"rnns",
"sequential",
"sequence",
"autoregressive",
"forecasting"
],
[
"gradient",
"hessian",
"softmax",
"classifying",
"gradients",
"backpropagation",
"classifiers",
"contrastive",
"labeled",
"labeling"
],
[
"models",
"predicts",
"predictive",
"model",
"predict",
"modeled",
"modeling",
"predictors",
"predictor",
"predicted"
],
[
"neurons",
"represents",
"tend",
"grid",
"convexity",
"convex",
"views",
"viewed",
"fields",
"similar"
],
[
"stationary",
"dynamical",
"notions",
"nonlinear",
"expressive",
"embeddings",
"dynamics",
"invariant",
"stability",
"geometric"
],
[
"greedy",
"bandits",
"searching",
"market",
"seek",
"bandit",
"strategy",
"choosing",
"optimal",
"search"
],
[
"mixture",
"combining",
"separately",
"partial",
"formulation",
"combine",
"theoretic",
"ensembles",
"additive",
"composed"
],
[
"gaussian",
"parameters",
"parametric",
"distribution",
"activation",
"parameter",
"activations",
"posterior",
"criterion",
"distributions"
],
[
"gaussian",
"laplacian",
"estimate",
"estimates",
"asymptotically",
"landscape",
"estimators",
"kernels",
"expectation",
"variance"
],
[
"decompose",
"processes",
"generator",
"decomposition",
"variability",
"weights",
"core",
"softmax",
"serve",
"process"
],
[
"sparse",
"constrained",
"sparsity",
"setting",
"trivial",
"regularization",
"lasso",
"feature",
"constrain",
"regularized"
],
[
"optimization",
"maximizing",
"optimizer",
"optimized",
"optimize",
"maximization",
"optimisation",
"optimizing",
"constrained",
"optimizes"
],
[
"controlled",
"behavioral",
"reinforcement",
"enforce",
"behaviors",
"action",
"actions",
"policies",
"policy",
"behavior"
],
[
"adaptation",
"transferred",
"domains",
"transfer",
"transferring",
"across",
"complementary",
"transforms",
"pretrained",
"transformation"
],
[
"gradient",
"convex",
"sgd",
"hessian",
"accelerated",
"convexity",
"accelerate",
"backpropagation",
"gradients",
"descent"
],
[
"forecasting",
"predicting",
"predictions",
"weather",
"predicts",
"predicted",
"predict",
"forecasts",
"forecast",
"prediction"
],
[
"variational",
"stochastic",
"autoencoder",
"likelihood",
"latent",
"bayesian",
"autoencoders",
"modelling",
"modeled",
"modeling"
],
[
"physical",
"eeg",
"wearable",
"device",
"running",
"walk",
"tested",
"devices",
"activity",
"devise"
],
[
"sampling",
"estimator",
"sampled",
"observation",
"estimators",
"population",
"observations",
"observed",
"estimating",
"sample"
],
[
"discrimination",
"fairness",
"bias",
"biases",
"unbiased",
"discriminate",
"equal",
"biased",
"representative",
"assess"
],
[
"intelligence",
"brain",
"cognitive",
"perception",
"neuron",
"focusing",
"perceptual",
"attention",
"neural",
"eeg"
],
[
"infrastructure",
"iot",
"deployment",
"task",
"combines",
"simultaneously",
"tasks",
"communicate",
"multi",
"server"
],
[
"robotic",
"machines",
"robots",
"leads",
"tensors",
"tensor",
"tasks",
"machine",
"perform",
"tools"
],
[
"personalized",
"recommender",
"appealing",
"factorization",
"recommendations",
"recommendation",
"ranking",
"rank",
"profile",
"suggests"
],
[
"paths",
"divergence",
"boundary",
"boundaries",
"volume",
"pipelines",
"dense",
"partitioning",
"densities",
"density"
],
[
"causal",
"interventions",
"involve",
"effects",
"impacts",
"facilitate",
"interactions",
"covariates",
"involving",
"factors"
],
[
"magnitude",
"acceleration",
"vector",
"vectors",
"attempts",
"fail",
"efforts",
"causes",
"imbalance",
"topic"
]
] | 233.773731 | all-MiniLM-L6-v2 | 0.915 | -0.318967 | 0.195741 | 0.90393 |
ArXiv ML Papers | S³_combined | 43 | 50 | [
[
"predictive",
"downstream",
"flows",
"flow",
"directed",
"superior",
"pipeline",
"pipelines",
"centralized",
"causal"
],
[
"subspaces",
"subspace",
"represent",
"grid",
"embeddings",
"patient",
"matrices",
"matrix",
"representation",
"vs"
],
[
"exploring",
"explored",
"superiority",
"25",
"intelligent",
"integrating",
"directions",
"50",
"exploration",
"integrate"
],
[
"importance",
"convolutional",
"contrastive",
"frequencies",
"offers",
"complexities",
"convolution",
"ideas",
"utility",
"efforts"
],
[
"discriminate",
"biases",
"fairness",
"discrimination",
"bias",
"biased",
"representative",
"outcomes",
"fair",
"equal"
],
[
"interpretation",
"interpretability",
"industry",
"interpretable",
"credit",
"financial",
"understand",
"products",
"product",
"interpret"
],
[
"kernel",
"gaussian",
"convolutional",
"orthogonal",
"multivariate",
"convolution",
"process",
"covariance",
"convolutions",
"kernels"
],
[
"layers",
"backpropagation",
"neurons",
"neuron",
"neural",
"networks",
"layer",
"net",
"network",
"projections"
],
[
"involve",
"aid",
"interventions",
"associated",
"impacts",
"causal",
"counterfactual",
"sources",
"imputation",
"evidence"
],
[
"popularity",
"impacts",
"scaling",
"scales",
"perturbation",
"perturbations",
"scale",
"diffusion",
"roughly",
"magnitude"
],
[
"transfer",
"adaptation",
"adapting",
"transforms",
"transformation",
"transferring",
"transferred",
"adapt",
"transformations",
"replacing"
],
[
"behavioral",
"policy",
"reinforcement",
"policies",
"interventions",
"enforce",
"actions",
"critic",
"textit",
"action"
],
[
"accelerators",
"computing",
"hardware",
"throughput",
"processing",
"computation",
"gpu",
"implementations",
"performance",
"faster"
],
[
"languages",
"sensors",
"detect",
"ucb",
"tensors",
"basic",
"basis",
"tensor",
"dictionary",
"molecular"
],
[
"metric",
"distances",
"similarity",
"metrics",
"distance",
"comparable",
"nearest",
"compares",
"comparison",
"quantify"
],
[
"acoustic",
"extends",
"arbitrary",
"constructing",
"extending",
"sampling",
"trees",
"boundary",
"extensive",
"expansion"
],
[
"image",
"fields",
"convexity",
"naive",
"represents",
"convex",
"frequently",
"increasingly",
"grid",
"pattern"
],
[
"transitions",
"invariant",
"dynamic",
"dynamics",
"transformations",
"dynamical",
"preserving",
"transforms",
"flexible",
"flows"
],
[
"energy",
"bounded",
"demand",
"half",
"threat",
"imagenet",
"embedding",
"notions",
"representations",
"camera"
],
[
"optimization",
"constraint",
"constraints",
"minimum",
"constrained",
"optimisation",
"range",
"sensitivity",
"proposes",
"min"
],
[
"security",
"attacks",
"adversarial",
"adversarially",
"adversaries",
"exploit",
"adversary",
"attacker",
"threat",
"exploits"
],
[
"allowed",
"allow",
"restricted",
"factors",
"limitations",
"ability",
"able",
"allowing",
"limitation",
"character"
],
[
"label",
"labeling",
"labels",
"labeled",
"classify",
"classified",
"unlabeled",
"classification",
"taxonomy",
"categories"
],
[
"robotics",
"robot",
"robotic",
"pose",
"poses",
"joint",
"robots",
"position",
"objects",
"coordinate"
],
[
"phases",
"assess",
"spectral",
"confirm",
"extraction",
"partial",
"intrinsic",
"verify",
"distinguish",
"evaluate"
],
[
"exploitation",
"search",
"bandit",
"seek",
"searching",
"market",
"bandits",
"obtains",
"price",
"acquiring"
],
[
"protein",
"structural",
"discover",
"chain",
"molecular",
"generate",
"synthesis",
"genetic",
"generating",
"generated"
],
[
"targets",
"through",
"fail",
"descent",
"completion",
"reflect",
"down",
"signal",
"poisoning",
"acceleration"
],
[
"monte",
"mcmc",
"stochastic",
"markov",
"simulation",
"simulations",
"sampling",
"randomly",
"carlo",
"random"
],
[
"granularity",
"tend",
"multimodal",
"spatio",
"corpus",
"software",
"vanilla",
"contextual",
"empirical",
"raw"
],
[
"marginal",
"width",
"vary",
"covariates",
"risks",
"potential",
"risk",
"wide",
"divergence",
"dependence"
],
[
"dimension",
"capacity",
"selects",
"dimensionality",
"dimensional",
"datasets",
"representations",
"dataset",
"dimensions",
"retrieval"
],
[
"reconstructed",
"channel",
"encoded",
"compression",
"decoder",
"reconstruct",
"decoding",
"channels",
"distortion",
"decompose"
],
[
"textual",
"corpora",
"nlp",
"texts",
"linguistic",
"corpus",
"sentences",
"translation",
"text",
"semantic"
],
[
"pruning",
"compressed",
"compress",
"distinct",
"reduced",
"reduces",
"minimal",
"pruned",
"compression",
"reduce"
],
[
"content",
"process",
"flow",
"bottleneck",
"scene",
"scenes",
"throughput",
"processes",
"topic",
"diffusion"
],
[
"individual",
"dependent",
"attributes",
"lightweight",
"outcomes",
"last",
"passing",
"get",
"gives",
"pass"
],
[
"ensembles",
"stream",
"combines",
"combine",
"ensemble",
"continuous",
"streams",
"entries",
"collect",
"bridge"
],
[
"variational",
"encode",
"autoencoders",
"autoencoder",
"encoder",
"encoders",
"encodes",
"encoded",
"code",
"encoding"
],
[
"core",
"soft",
"embeddings",
"embedding",
"embedded",
"producing",
"relu",
"poses",
"energy",
"weights"
],
[
"cooperative",
"shared",
"centralized",
"collaborative",
"federated",
"distributed",
"communicate",
"communications",
"protocol",
"communication"
],
[
"edges",
"graph",
"graphs",
"node",
"vertices",
"nodes",
"vertex",
"edge",
"graphical",
"networks"
],
[
"patients",
"clinical",
"diagnosis",
"medical",
"mri",
"medicine",
"diagnostic",
"cancer",
"healthcare",
"imaging"
],
[
"describes",
"characterized",
"initialization",
"monte",
"defined",
"compact",
"dynamically",
"termed",
"defines",
"recurrent"
],
[
"proposes",
"manipulation",
"style",
"imitation",
"manner",
"formalism",
"designs",
"patterns",
"traditional",
"fashion"
],
[
"forest",
"boost",
"boosting",
"tree",
"trees",
"classifiers",
"branch",
"ensemble",
"classify",
"classifier"
],
[
"factorization",
"suggest",
"collaborative",
"ranking",
"suggests",
"recommender",
"preferences",
"recommendations",
"recommendation",
"personalized"
],
[
"outlier",
"anomaly",
"anomalies",
"iot",
"outliers",
"event",
"deviation",
"intrusion",
"particle",
"detect"
],
[
"uniform",
"full",
"quality",
"parameters",
"patches",
"remains",
"uniformly",
"thousands",
"hundreds",
"estimates"
],
[
"nonlinear",
"hessian",
"differentiation",
"differentially",
"derivatives",
"differential",
"differentiable",
"implicit",
"equations",
"dynamics"
]
] | 300.906645 | all-MiniLM-L6-v2 | 0.93 | -0.326993 | 0.196145 | 0.907047 |
ArXiv ML Papers | S³_combined | 44 | 50 | [
[
"records",
"temporal",
"sequences",
"sequential",
"lstm",
"demand",
"recurrent",
"forecast",
"forecasting",
"forecasts"
],
[
"genetic",
"molecular",
"protein",
"chain",
"structural",
"added",
"discover",
"generate",
"find",
"domain"
],
[
"publicly",
"federated",
"privacy",
"independent",
"independently",
"private",
"public",
"conventional",
"standard",
"differentially"
],
[
"goal",
"lead",
"fast",
"social",
"post",
"influence",
"https",
"twitter",
"video",
"public"
],
[
"transforms",
"aggregation",
"preserving",
"transformations",
"invariant",
"transform",
"transformation",
"dynamical",
"dynamics",
"operations"
],
[
"preferences",
"lipschitz",
"criterion",
"euclidean",
"bound",
"norm",
"regularization",
"lasso",
"enables",
"settings"
],
[
"feature",
"boost",
"collect",
"boosting",
"identified",
"features",
"character",
"collected",
"additionally",
"speedups"
],
[
"generator",
"generating",
"gan",
"generative",
"gans",
"generation",
"producing",
"generate",
"generates",
"hilbert"
],
[
"mixture",
"aspect",
"minimax",
"mini",
"aspects",
"composed",
"partial",
"ensembles",
"localized",
"head"
],
[
"embedding",
"embeddings",
"densities",
"density",
"inequality",
"embedded",
"incorporate",
"defend",
"capacity",
"channel"
],
[
"bayes",
"bayesian",
"posterior",
"priors",
"prior",
"densities",
"distributions",
"likelihood",
"priori",
"distributional"
],
[
"iteration",
"incremental",
"massive",
"expand",
"expansion",
"modes",
"inform",
"exploration",
"perturbations",
"iterations"
],
[
"empirically",
"sampling",
"samples",
"researchers",
"observations",
"ensembles",
"sample",
"empirical",
"experiments",
"systematically"
],
[
"formulate",
"modeled",
"models",
"computationally",
"modeling",
"model",
"structure",
"modelling",
"speedup",
"profiles"
],
[
"mutual",
"share",
"mixed",
"coupled",
"mismatch",
"fusion",
"combines",
"fidelity",
"gradient",
"backpropagation"
],
[
"tensor",
"tensors",
"factor",
"pytorch",
"latent",
"factorization",
"factors",
"vector",
"vectors",
"matrices"
],
[
"ensemble",
"detect",
"recommender",
"videos",
"predictors",
"audio",
"detecting",
"video",
"classifier",
"prediction"
],
[
"recommendation",
"recommender",
"suggests",
"personalized",
"suggest",
"factorization",
"recommendations",
"collaborative",
"ranking",
"profiles"
],
[
"interpretation",
"describes",
"interpretability",
"industry",
"service",
"interpretable",
"development",
"product",
"production",
"describe"
],
[
"embeddings",
"poses",
"core",
"relu",
"embedded",
"soft",
"embedding",
"raw",
"production",
"weights"
],
[
"individuals",
"deterministic",
"dependencies",
"adaptively",
"person",
"inefficient",
"predictor",
"controller",
"known",
"predictors"
],
[
"neuron",
"brain",
"cognitive",
"eeg",
"neural",
"neurons",
"intelligence",
"mri",
"intractable",
"coherent"
],
[
"nonnegative",
"score",
"quantify",
"outcomes",
"measure",
"averaging",
"measures",
"aggregate",
"marginal",
"dedicated"
],
[
"pass",
"gives",
"expected",
"outcomes",
"lightweight",
"get",
"dependent",
"attributes",
"passing",
"attribute"
],
[
"quantum",
"quantized",
"quantization",
"resonance",
"hilbert",
"fourier",
"spectrum",
"particle",
"discovery",
"signal"
],
[
"proposes",
"nearest",
"estimation",
"minimum",
"sensitivity",
"distance",
"estimated",
"near",
"softmax",
"minutes"
],
[
"camera",
"notions",
"energy",
"imagenet",
"demand",
"recognition",
"machines",
"want",
"representations",
"supervised"
],
[
"periods",
"severe",
"dramatically",
"compressed",
"sparse",
"tractable",
"restricted",
"threshold",
"minimax",
"characterized"
],
[
"surrogate",
"gan",
"highest",
"assessment",
"assessing",
"assess",
"rank",
"ranking",
"scores",
"scoring"
],
[
"guarantees",
"detection",
"markov",
"observing",
"bayes",
"detecting",
"cnns",
"ensuring",
"probabilistic",
"security"
],
[
"filters",
"values",
"numbers",
"filter",
"filtering",
"numerically",
"numerical",
"statistics",
"parameters",
"accelerate"
],
[
"optimizing",
"optimize",
"objectives",
"objective",
"multivariate",
"optimization",
"multi",
"optimizes",
"optimisation",
"maximizing"
],
[
"regularization",
"normalizing",
"normalization",
"normalized",
"characterizing",
"decomposition",
"factorization",
"maintaining",
"principal",
"characterize"
],
[
"dependency",
"frequencies",
"contrastive",
"frequency",
"solving",
"spectral",
"contrast",
"sign",
"spectrum",
"utility"
],
[
"boosting",
"randomized",
"classifier",
"algorithm",
"ensemble",
"classifiers",
"transformer",
"ranking",
"svm",
"recognition"
],
[
"weather",
"predictions",
"predicting",
"predicted",
"predicts",
"predict",
"forecasts",
"forecasting",
"forecast",
"prediction"
],
[
"image",
"convexity",
"convex",
"attribute",
"tend",
"represents",
"naive",
"grid",
"fields",
"view"
],
[
"models",
"model",
"fitting",
"candidate",
"latency",
"candidates",
"availability",
"having",
"joint",
"feasibility"
],
[
"dnn",
"dnns",
"auc",
"finite",
"infinite",
"imbalance",
"ratio",
"reliability",
"difficulty",
"epsilon"
],
[
"hardware",
"computation",
"computing",
"throughput",
"processing",
"accelerators",
"performance",
"runtime",
"implementations",
"cloud"
],
[
"reliable",
"methodology",
"validity",
"validate",
"profiles",
"validation",
"determined",
"identifies",
"counterfactual",
"identification"
],
[
"downstream",
"pipelines",
"centralized",
"directed",
"flow",
"pipeline",
"superior",
"flows",
"nas",
"predictive"
],
[
"attention",
"focusing",
"eye",
"specifically",
"focus",
"vision",
"perceptual",
"leading",
"summary",
"observing"
],
[
"granularity",
"flow",
"bandwidth",
"formulation",
"diffusion",
"traffic",
"terms",
"mixture",
"downstream",
"compression"
],
[
"medicine",
"patients",
"disease",
"medical",
"clinical",
"diseases",
"healthcare",
"patient",
"health",
"practitioners"
],
[
"sgd",
"ascent",
"descent",
"convexity",
"gradients",
"gradient",
"convex",
"backpropagation",
"hessian",
"convergence"
],
[
"detecting",
"program",
"algorithms",
"algorithmic",
"algorithm",
"detect",
"recursive",
"implementations",
"solvers",
"solver"
],
[
"estimator",
"82",
"coefficients",
"include",
"estimators",
"83",
"who",
"themselves",
"selected",
"coefficient"
],
[
"analyse",
"contrastive",
"assessing",
"wasserstein",
"contrast",
"hessian",
"assessment",
"evaluations",
"horizon",
"perspectives"
],
[
"potential",
"width",
"wide",
"vary",
"oriented",
"specialized",
"considerable",
"dependence",
"scalability",
"stable"
]
] | 319.947455 | all-MiniLM-L6-v2 | 0.892 | -0.334666 | 0.18104 | 0.90437 |
ArXiv ML Papers | S³_combined | 45 | 50 | [
[
"determined",
"ensembles",
"continuous",
"collect",
"generator",
"defined",
"numerically",
"consist",
"collected",
"ensemble"
],
[
"perturbation",
"popularity",
"perturbations",
"magnitude",
"scale",
"scaling",
"effect",
"impacts",
"impact",
"distortion"
],
[
"energy",
"imagenet",
"representations",
"convolution",
"recognition",
"convolutions",
"machine",
"robots",
"robotic",
"bernoulli"
],
[
"sparse",
"tractable",
"check",
"threshold",
"severe",
"modal",
"characterized",
"gender",
"dense",
"dramatically"
],
[
"segmentation",
"unsupervised",
"clusters",
"clustering",
"cluster",
"segment",
"partitioning",
"regions",
"separation",
"counterparts"
],
[
"flow",
"scene",
"bottleneck",
"process",
"scenes",
"throughput",
"recovery",
"content",
"lipschitz",
"brain"
],
[
"backpropagation",
"layer",
"networks",
"neurons",
"layers",
"neuron",
"neural",
"projections",
"projection",
"net"
],
[
"invariant",
"regularization",
"characterize",
"characterizing",
"transformations",
"normalization",
"decomposition",
"normalized",
"normalizing",
"maintaining"
],
[
"measurements",
"task",
"technologies",
"projection",
"projections",
"decade",
"measurement",
"procedures",
"tasks",
"technology"
],
[
"partial",
"mixture",
"mini",
"aspects",
"composed",
"minimax",
"ensembles",
"aspect",
"proposals",
"head"
],
[
"defines",
"monte",
"defined",
"dynamically",
"termed",
"describes",
"characterized",
"designed",
"initialization",
"compact"
],
[
"validation",
"validated",
"predicted",
"calibrated",
"validate",
"expect",
"accuracy",
"calibration",
"predictions",
"scoring"
],
[
"patient",
"diseases",
"diagnosis",
"healthcare",
"clinical",
"medical",
"medicine",
"mri",
"patients",
"disease"
],
[
"convex",
"gradient",
"descent",
"sgd",
"hessian",
"convexity",
"gradients",
"backpropagation",
"convergence",
"ascent"
],
[
"lead",
"post",
"fast",
"goal",
"goals",
"scoring",
"leading",
"https",
"curve",
"tail"
],
[
"linear",
"representation",
"subspace",
"subspaces",
"grid",
"matrix",
"matrices",
"representations",
"vectors",
"represent"
],
[
"dynamic",
"preserving",
"dynamics",
"transformations",
"flexible",
"dynamical",
"constrained",
"stationary",
"invariant",
"nonlinear"
],
[
"learner",
"studying",
"study",
"student",
"learns",
"learners",
"teacher",
"classes",
"studied",
"class"
],
[
"structural",
"biological",
"mechanisms",
"genetic",
"molecular",
"protein",
"chain",
"module",
"discover",
"biology"
],
[
"embedding",
"embeddings",
"core",
"relu",
"soft",
"poses",
"embedded",
"energy",
"weights",
"raw"
],
[
"industrial",
"process",
"industry",
"development",
"processes",
"production",
"product",
"service",
"products",
"developers"
],
[
"rnn",
"recurrent",
"lstm",
"sequential",
"sequences",
"recall",
"sequence",
"temporal",
"rnns",
"forecasts"
],
[
"representative",
"fairness",
"bias",
"biased",
"biases",
"discrimination",
"discriminate",
"fair",
"outcomes",
"candidate"
],
[
"randomly",
"softmax",
"generalization",
"generalizing",
"generalize",
"distinguishing",
"characteristic",
"distinguish",
"summarize",
"ensemble"
],
[
"policy",
"actions",
"action",
"reinforcement",
"interventions",
"behavioral",
"enforce",
"policies",
"behaviors",
"textit"
],
[
"quantization",
"resonance",
"particle",
"quantum",
"quantized",
"hilbert",
"fourier",
"observable",
"signal",
"spectrum"
],
[
"themselves",
"82",
"incorporates",
"coefficients",
"include",
"collect",
"includes",
"selected",
"detector",
"popular"
],
[
"fidelity",
"fusion",
"mixed",
"coupled",
"combines",
"complementary",
"mutual",
"mismatch",
"share",
"gradient"
],
[
"optimizing",
"optimisation",
"optimized",
"optimize",
"algorithm",
"heuristic",
"optimization",
"optimizer",
"heuristics",
"algorithms"
],
[
"lipschitz",
"regularization",
"enables",
"euclidean",
"lasso",
"norm",
"generalized",
"operators",
"preferences",
"criterion"
],
[
"allowed",
"restricted",
"encoded",
"risks",
"curse",
"codes",
"deal",
"risk",
"able",
"allow"
],
[
"down",
"completion",
"descent",
"reflect",
"fail",
"poisoning",
"efforts",
"shot",
"targets",
"impacts"
],
[
"videos",
"supervised",
"predictors",
"audio",
"detect",
"detecting",
"recommender",
"perception",
"ensemble",
"sensing"
],
[
"equations",
"nonlinear",
"differentiation",
"simulations",
"implicit",
"derivatives",
"curve",
"hessian",
"differentially",
"differential"
],
[
"ucb",
"sensors",
"basic",
"calibration",
"languages",
"molecular",
"tensors",
"tensor",
"basis",
"detect"
],
[
"calibration",
"quality",
"remains",
"benchmark",
"respect",
"highlights",
"uniform",
"imagenet",
"patches",
"treat"
],
[
"than",
"promise",
"sketch",
"expect",
"adopted",
"adopt",
"adopting",
"amenable",
"imitation",
"mean"
],
[
"robotics",
"robotic",
"robot",
"robots",
"joint",
"pose",
"poses",
"robust",
"position",
"hand"
],
[
"communicate",
"distributed",
"shared",
"cooperative",
"federated",
"communications",
"collaborative",
"centralized",
"protocol",
"communication"
],
[
"solver",
"methodology",
"controller",
"adaptively",
"ranking",
"ai",
"intelligence",
"discriminative",
"algorithmic",
"controlling"
],
[
"gaussian",
"kernel",
"multivariate",
"manifold",
"kernels",
"processes",
"covariates",
"parametric",
"orthogonal",
"convolutional"
],
[
"assessing",
"assessment",
"contrastive",
"examining",
"analyse",
"evaluations",
"analyze",
"insight",
"mri",
"perspectives"
],
[
"risk",
"width",
"wide",
"vary",
"risks",
"potential",
"covariates",
"specialized",
"flexibility",
"dependence"
],
[
"adversarial",
"attacks",
"attacker",
"exploits",
"adversary",
"adversaries",
"adversarially",
"exploiting",
"security",
"exploit"
],
[
"predicting",
"prediction",
"predict",
"projected",
"compress",
"atari",
"predictions",
"compression",
"predicts",
"embedding"
],
[
"fields",
"represents",
"convexity",
"convex",
"image",
"grid",
"frequently",
"subsequently",
"naive",
"attribute"
],
[
"computationally",
"structure",
"model",
"modeled",
"modeling",
"formulate",
"modelling",
"models",
"structures",
"profiles"
],
[
"wearable",
"devices",
"sensing",
"activity",
"physical",
"sensor",
"sensors",
"running",
"device",
"monitoring"
],
[
"manipulation",
"art",
"style",
"designs",
"fashion",
"designing",
"proposes",
"manner",
"imitation",
"formalism"
],
[
"learnt",
"facial",
"theory",
"training",
"practice",
"shifts",
"faces",
"face",
"phase",
"learning"
]
] | 303.456577 | all-MiniLM-L6-v2 | 0.91 | -0.344432 | 0.195031 | 0.89933 |
ArXiv ML Papers | S³_combined | 46 | 50 | [
[
"attacks",
"adversarial",
"attacker",
"adversarially",
"exploits",
"adversary",
"adversaries",
"security",
"exploit",
"attack"
],
[
"optimisation",
"optimizes",
"projected",
"hyperparameter",
"optimizer",
"hyperparameters",
"parameterized",
"hyper",
"projections",
"maximize"
],
[
"partial",
"mini",
"mixture",
"minimax",
"composed",
"aspect",
"aspects",
"ensembles",
"head",
"whole"
],
[
"labeling",
"labels",
"label",
"labeled",
"classification",
"unlabeled",
"categories",
"classify",
"taxonomy",
"classified"
],
[
"flows",
"pipeline",
"flow",
"downstream",
"pipelines",
"superior",
"directed",
"centralized",
"dense",
"filtering"
],
[
"rank",
"adaptively",
"predictor",
"predictors",
"ai",
"estimating",
"adaptive",
"approximating",
"ranking",
"solver"
],
[
"distortion",
"drastically",
"sparsity",
"perturbation",
"perturbations",
"perturbed",
"popularity",
"effect",
"effects",
"impact"
],
[
"clustering",
"segment",
"segmentation",
"unsupervised",
"partitioning",
"cluster",
"clusters",
"regions",
"counterparts",
"separation"
],
[
"reflect",
"descent",
"completion",
"down",
"fail",
"efforts",
"task",
"acceleration",
"targets",
"poisoning"
],
[
"outlier",
"anomaly",
"anomalies",
"outliers",
"iot",
"event",
"monitoring",
"intrusion",
"sensor",
"deviation"
],
[
"patients",
"medical",
"medicine",
"clinical",
"healthcare",
"patient",
"diseases",
"disease",
"practitioners",
"diagnosis"
],
[
"interpretable",
"interpretation",
"explaining",
"explanations",
"interpretability",
"credit",
"interpret",
"understand",
"interpreted",
"mean"
],
[
"multivariate",
"embeddings",
"query",
"capacity",
"densities",
"embedding",
"embedded",
"benefits",
"incorporate",
"separately"
],
[
"randomized",
"distinct",
"pruned",
"pruning",
"reduction",
"redundant",
"trivial",
"minimal",
"reduced",
"compressed"
],
[
"risks",
"constraint",
"penalty",
"constrained",
"safety",
"allowed",
"risk",
"ensures",
"constraints",
"constrain"
],
[
"teacher",
"learns",
"classes",
"learners",
"studying",
"student",
"learner",
"study",
"studied",
"class"
],
[
"testing",
"patches",
"patch",
"experimental",
"test",
"tested",
"tests",
"screening",
"experimentally",
"coverage"
],
[
"periods",
"tractable",
"severe",
"restricted",
"sparse",
"threshold",
"compressed",
"during",
"dramatically",
"minimal"
],
[
"boosting",
"collect",
"ensemble",
"aggregated",
"up",
"boost",
"feature",
"collected",
"ensembles",
"incremental"
],
[
"recommender",
"factorization",
"recommendation",
"recommendations",
"suggests",
"collaborative",
"personalized",
"suggest",
"profiles",
"ubiquitous"
],
[
"validate",
"accurately",
"accuracy",
"validation",
"reliable",
"uncertainties",
"validated",
"reliability",
"uncertainty",
"uncertain"
],
[
"filtering",
"values",
"filters",
"filter",
"numbers",
"numerically",
"statistics",
"numerical",
"accelerate",
"parameters"
],
[
"seek",
"bandits",
"bandit",
"market",
"exploration",
"exploitation",
"searching",
"price",
"obtains",
"search"
],
[
"smoothing",
"robust",
"smoothness",
"robustness",
"focused",
"smooth",
"noisy",
"focus",
"focusing",
"consistent"
],
[
"stability",
"stable",
"invariant",
"transformations",
"preserving",
"mathematical",
"systems",
"invariance",
"scales",
"transforms"
],
[
"guarantees",
"observing",
"ensuring",
"bayes",
"markov",
"randomized",
"probabilistic",
"coverage",
"detection",
"cnns"
],
[
"landscape",
"weather",
"forecast",
"forecasting",
"forecasts",
"forest",
"spectrum",
"reflect",
"nature",
"outside"
],
[
"subspace",
"matrix",
"matrices",
"subspaces",
"linear",
"grid",
"basis",
"describes",
"representation",
"randomized"
],
[
"defines",
"dynamically",
"termed",
"monte",
"defined",
"recurrent",
"characterized",
"continuously",
"initialization",
"compact"
],
[
"difficulty",
"dnn",
"precision",
"challenging",
"dnns",
"challenges",
"reliability",
"forgetting",
"struggle",
"epsilon"
],
[
"expressive",
"variations",
"variation",
"manipulation",
"designing",
"designs",
"style",
"coding",
"fashion",
"art"
],
[
"semantic",
"text",
"corpora",
"literature",
"nlp",
"corpus",
"texts",
"textual",
"sentences",
"linguistic"
],
[
"social",
"post",
"twitter",
"https",
"goals",
"video",
"goal",
"methodology",
"strategies",
"achieving"
],
[
"quality",
"treat",
"respect",
"calibration",
"captures",
"remains",
"highlights",
"capturing",
"capture",
"rate"
],
[
"granularity",
"distance",
"comparable",
"metrics",
"metric",
"meaningful",
"distances",
"similarity",
"space",
"molecular"
],
[
"inputs",
"dimensional",
"dimension",
"dimensionality",
"representations",
"representation",
"allocation",
"selects",
"retrieval",
"subspaces"
],
[
"discrimination",
"bias",
"fairness",
"biases",
"discriminate",
"biased",
"fair",
"outcomes",
"representative",
"equal"
],
[
"causal",
"interventions",
"impacts",
"involve",
"aid",
"consequence",
"trigger",
"associated",
"counterfactual",
"effects"
],
[
"robots",
"robot",
"robotic",
"machines",
"robust",
"pose",
"poses",
"robotics",
"objects",
"assisted"
],
[
"tool",
"raw",
"corpus",
"contextual",
"software",
"vanilla",
"spatio",
"tend",
"multimodal",
"empirical"
],
[
"limitations",
"arbitrarily",
"bound",
"asymptotically",
"complexity",
"sampling",
"estimate",
"bounded",
"bounds",
"extent"
],
[
"having",
"model",
"models",
"feasibility",
"latency",
"candidates",
"fitting",
"joint",
"candidate",
"phenomenon"
],
[
"randomly",
"generalizing",
"distinguish",
"generalization",
"classifying",
"distinguishing",
"ensemble",
"softmax",
"classifiers",
"learning"
],
[
"vary",
"oriented",
"width",
"wide",
"potential",
"dependence",
"scalability",
"specialized",
"risk",
"variances"
],
[
"voice",
"sound",
"audio",
"speech",
"acoustic",
"speaker",
"music",
"tune",
"frequencies",
"noise"
],
[
"gans",
"gan",
"generative",
"generates",
"generate",
"generation",
"generating",
"producing",
"generator",
"generated"
],
[
"markov",
"stochastic",
"monte",
"randomly",
"simulations",
"simulation",
"mcmc",
"sampling",
"carlo",
"probabilistic"
],
[
"predict",
"rnn",
"backpropagation",
"rnns",
"neighboring",
"tilde",
"nearest",
"directly",
"behind",
"earlier"
],
[
"gps",
"transportation",
"traffic",
"driving",
"navigation",
"transport",
"trajectories",
"vehicle",
"urban",
"vehicles"
],
[
"temporal",
"forecast",
"autoregressive",
"recurrent",
"forecasts",
"lstm",
"forecasting",
"sequences",
"sequential",
"sequence"
]
] | 312.315722 | all-MiniLM-L6-v2 | 0.95 | -0.326175 | 0.190984 | 0.910355 |
ArXiv ML Papers | CombinedTM | 43 | 10 | [
[
"of",
"matrix",
"rank",
"we",
"problem",
"that",
"for",
"is",
"algorithm",
"in"
],
[
"performance",
"the",
"and",
"to",
"network",
"model",
"based",
"detection",
"is",
"on"
],
[
"by",
"are",
"in",
"the",
"of",
"to",
"as",
"is",
"data",
"this"
],
[
"graph",
"domain",
"node",
"datasets",
"graphs",
"knowledge",
"source",
"propose",
"art",
"state"
],
[
"xgboost",
"stream",
"population",
"neurons",
"up",
"obtain",
"inference",
"impractical",
"inspection",
"root"
],
[
"induced",
"net",
"thus",
"layer",
"units",
"faster",
"computational",
"architecture",
"introduces",
"hardware"
],
[
"make",
"text",
"related",
"clinical",
"language",
"community",
"techniques",
"recently",
"word",
"progress"
],
[
"adversarial",
"training",
"attack",
"against",
"models",
"attacks",
"model",
"networks",
"examples",
"neural"
],
[
"that",
"are",
"models",
"to",
"and",
"in",
"we",
"can",
"this",
"data"
],
[
"agent",
"optimization",
"gradient",
"algorithm",
"rl",
"stochastic",
"policy",
"reward",
"regret",
"algorithms"
]
] | 1,002.15203 | all-MiniLM-L6-v2 | 0.83 | -0.062475 | 0.140377 | 0.684834 |