metadata

model-index:
  - name: roberta-large-self-disclosure-sentence-classification
    results: []
language:
  - en
base_model: FacebookAI/roberta-large
license: cc-by-nc-2.0
tags:
  - roberta
  - privacy
  - self-disclosure classification
  - PII

Model Card for roberta-large-self-disclosure-sentence-classification

The model is used to classify whether a given sentence contains disclosure or not. It is a binary sentence-level classification where label 1 means containing self-disclosure, and 0 means not containing.

For more details, please read the paper: Reducing Privacy Risks in Online Self-Disclosures with Language Models .

Accessing this model implies automatic agreement to the following guidelines:

Only use the model for research purposes.
No redistribution without the author's agreement.
Any derivative works created using this model must acknowledge the original author.

Model Description

Model type: A finetuned sentence level classifier that classifies whether a given sentence contains disclosure or not.
Language(s) (NLP): English
License: Creative Commons Attribution-NonCommercial
Finetuned from model: FacebookAI/roberta-large

Example Code

import torch
from transformers import AutoTokenizer, AutoModelForSequenceClassification, AutoConfig

config = AutoConfig.from_pretrained("douy/roberta-large-self-disclosure-sentence-classification")
tokenizer = AutoTokenizer.from_pretrained("douy/roberta-large-self-disclosure-sentence-classification")

model = AutoModelForSequenceClassification.from_pretrained("douy/roberta-large-self-disclosure-sentence-classification", 
            config=config, device_map="cuda:0").eval()

sentences = [
    "I am a 23-year-old who is currently going through the last leg of undergraduate school.",
    "There is a joke in the design industry about that.",
    "My husband and I live in US.",
    "I was messing with advanced voice the other day and I was like, 'Oh, I can do this.'",
]

inputs = tokenizer(sentences, return_tensors="pt", padding=True, truncation=True).to(model.device)

with torch.no_grad():
    logits = model(**inputs).logits

# predicted is the argmax of each row
predicted_class = logits.argmax(dim=-1)

# 1 means the sentence contains self-disclosure
# 0 means the sentence does not contain self-disclosure

# predicted_class: tensor([1, 0, 1, 0], device='cuda:0')

Evaluation

The model achieves 88.6% accuracy.

Citation

@article{dou2023reducing,
  title={Reducing Privacy Risks in Online Self-Disclosures with Language Models},
  author={Dou, Yao and Krsek, Isadora and Naous, Tarek and Kabra, Anubha and Das, Sauvik and Ritter, Alan and Xu, Wei},
  journal={arXiv preprint arXiv:2311.09538},
  year={2023}
}