RicardoRei
commited on
Commit
•
7ed04dd
1
Parent(s):
078ca15
Update README.md
Browse files
README.md
CHANGED
@@ -1,3 +1,65 @@
|
|
1 |
---
|
2 |
license: apache-2.0
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
3 |
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
---
|
2 |
license: apache-2.0
|
3 |
+
size_categories:
|
4 |
+
- 1M<n<10M
|
5 |
+
language:
|
6 |
+
- cs
|
7 |
+
- de
|
8 |
+
- en
|
9 |
+
- hr
|
10 |
+
- ja
|
11 |
+
- liv
|
12 |
+
- ru
|
13 |
+
- sah
|
14 |
+
- uk
|
15 |
+
- zh
|
16 |
+
tags:
|
17 |
+
- mt-evaluation
|
18 |
+
- WMT
|
19 |
+
- 12-lang-pairs
|
20 |
---
|
21 |
+
|
22 |
+
# Dataset Summary
|
23 |
+
|
24 |
+
In 2022, several changes were made to the annotation procedure used in the WMT Translation task. In contrast to the standard DA (sliding scale from 0-100) used in previous years, in 2022 annotators performed DA+SQM (Direct Assessment + Scalar Quality Metric). In DA+SQM, the annotators still provide a raw score between 0 and 100, but also are presented with seven labeled tick marks. DA+SQM helps to stabilize scores across annotators (as compared to DA).
|
25 |
+
|
26 |
+
The data is organised into 8 columns:
|
27 |
+
- lp: language pair
|
28 |
+
- src: input text
|
29 |
+
- mt: translation
|
30 |
+
- ref: reference translation
|
31 |
+
- score: direct assessment
|
32 |
+
- system: MT engine that produced the `mt`
|
33 |
+
- annotators: number of annotators
|
34 |
+
- domain: domain of the input text (e.g. news)
|
35 |
+
- year: collection year
|
36 |
+
|
37 |
+
You can also find the original data [here](https://www.statmt.org/wmt22/results.html)
|
38 |
+
|
39 |
+
## Python usage:
|
40 |
+
|
41 |
+
```python
|
42 |
+
from datasets import load_dataset
|
43 |
+
dataset = load_dataset("RicardoRei/wmt-sqm-human-evaluation", split="train")
|
44 |
+
```
|
45 |
+
|
46 |
+
There is no standard train/test split for this dataset but you can easily split it according to year, language pair or domain. E.g. :
|
47 |
+
|
48 |
+
```python
|
49 |
+
# split by year
|
50 |
+
data = dataset.filter(lambda example: example["year"] == 2022)
|
51 |
+
|
52 |
+
# split by LP
|
53 |
+
data = dataset.filter(lambda example: example["lp"] == "en-de")
|
54 |
+
|
55 |
+
# split by domain
|
56 |
+
data = dataset.filter(lambda example: example["domain"] == "news")
|
57 |
+
```
|
58 |
+
|
59 |
+
Note that, so far, all data is from [2022 General Translation task](https://www.statmt.org/wmt22/translation-task.html)
|
60 |
+
|
61 |
+
## Citation Information
|
62 |
+
|
63 |
+
If you use this data please cite the WMT findings:
|
64 |
+
- [Findings of the 2022 Conference on Machine Translation (WMT22)](https://aclanthology.org/2022.wmt-1.1.pdf)
|
65 |
+
|