SoftEDA: Rethinking Rule-Based Data Augmentation with Soft Labels
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2024
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866909099047780352 |
|---|---|
| author | Choi, Juhwan Jin, Kyohoon Lee, Junho Song, Sangmin Kim, Youngbin |
| author_facet | Choi, Juhwan Jin, Kyohoon Lee, Junho Song, Sangmin Kim, Youngbin |
| contents | Rule-based text data augmentation is widely used for NLP tasks due to its simplicity. However, this method can potentially damage the original meaning of the text, ultimately hurting the performance of the model. To overcome this limitation, we propose a straightforward technique for applying soft labels to augmented data. We conducted experiments across seven different classification tasks and empirically demonstrated the effectiveness of our proposed approach. We have publicly opened our source code for reproducibility. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2402_05591 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | SoftEDA: Rethinking Rule-Based Data Augmentation with Soft Labels Choi, Juhwan Jin, Kyohoon Lee, Junho Song, Sangmin Kim, Youngbin Computation and Language Artificial Intelligence Rule-based text data augmentation is widely used for NLP tasks due to its simplicity. However, this method can potentially damage the original meaning of the text, ultimately hurting the performance of the model. To overcome this limitation, we propose a straightforward technique for applying soft labels to augmented data. We conducted experiments across seven different classification tasks and empirically demonstrated the effectiveness of our proposed approach. We have publicly opened our source code for reproducibility. |
| title | SoftEDA: Rethinking Rule-Based Data Augmentation with Soft Labels |
| topic | Computation and Language Artificial Intelligence |
| url | https://arxiv.org/abs/2402.05591 |