READ: Improving Relation Extraction from an ADversarial Perspective
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_ | 1866916192123355136 |
|---|---|
| author | Li, Dawei Hogan, William Shang, Jingbo |
| author_facet | Li, Dawei Hogan, William Shang, Jingbo |
| contents | Recent works in relation extraction (RE) have achieved promising benchmark accuracy; however, our adversarial attack experiments show that these works excessively rely on entities, making their generalization capability questionable. To address this issue, we propose an adversarial training method specifically designed for RE. Our approach introduces both sequence- and token-level perturbations to the sample and uses a separate perturbation vocabulary to improve the search for entity and context perturbations. Furthermore, we introduce a probabilistic strategy for leaving clean tokens in the context during adversarial training. This strategy enables a larger attack budget for entities and coaxes the model to leverage relational patterns embedded in the context. Extensive experiments show that compared to various adversarial training methods, our method significantly improves both the accuracy and robustness of the model. Additionally, experiments on different data availability settings highlight the effectiveness of our method in low-resource scenarios. We also perform in-depth analyses of our proposed method and provide further hints. We will release our code at https://github.com/David-Li0406/READ. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2404_02931 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | READ: Improving Relation Extraction from an ADversarial Perspective Li, Dawei Hogan, William Shang, Jingbo Computation and Language Artificial Intelligence Recent works in relation extraction (RE) have achieved promising benchmark accuracy; however, our adversarial attack experiments show that these works excessively rely on entities, making their generalization capability questionable. To address this issue, we propose an adversarial training method specifically designed for RE. Our approach introduces both sequence- and token-level perturbations to the sample and uses a separate perturbation vocabulary to improve the search for entity and context perturbations. Furthermore, we introduce a probabilistic strategy for leaving clean tokens in the context during adversarial training. This strategy enables a larger attack budget for entities and coaxes the model to leverage relational patterns embedded in the context. Extensive experiments show that compared to various adversarial training methods, our method significantly improves both the accuracy and robustness of the model. Additionally, experiments on different data availability settings highlight the effectiveness of our method in low-resource scenarios. We also perform in-depth analyses of our proposed method and provide further hints. We will release our code at https://github.com/David-Li0406/READ. |
| title | READ: Improving Relation Extraction from an ADversarial Perspective |
| topic | Computation and Language Artificial Intelligence |
| url | https://arxiv.org/abs/2404.02931 |