Update README.md
Browse files
README.md
CHANGED
@@ -1 +1,16 @@
|
|
1 |
-
This model is used to generate paraphrases. It has been trained on a mix of 3 different paraphrase detection datasets: MSR, Quora, Google-PAWS.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
This model is used to generate paraphrases. It has been trained on a mix of 3 different paraphrase detection datasets: MSR, Quora, Google-PAWS.
|
2 |
+
|
3 |
+
We use this model in our ACL'21 Paper ["PROTAUGMENT: Unsupervised diverse short-texts paraphrasing for intent detection meta-learning"](https://arxiv.org/abs/2105.12995)
|
4 |
+
|
5 |
+
This model, jointly used with generation constraints, allows to generate diverse paraphrases. We use those paraphrases as a data augmentation technique to further boosts a classification model's generalization capability. Feel free to play with the [code](https://github.com/tdopierre/ProtAugment)!
|
6 |
+
|
7 |
+
If you use this model, please consider citing our paper.
|
8 |
+
```bash
|
9 |
+
@article{Dopierre2021ProtAugmentUD,
|
10 |
+
title={ProtAugment: Unsupervised diverse short-texts paraphrasing for intent detection meta-learning},
|
11 |
+
author={Thomas Dopierre and C. Gravier and Wilfried Logerais},
|
12 |
+
journal={ArXiv},
|
13 |
+
year={2021},
|
14 |
+
volume={abs/2105.12995}
|
15 |
+
}
|
16 |
+
```
|