Self-distilled Transitive Instance Weighting for Denoised Distantly Supervised Relation Extraction

Xiangyu Lin, Weijia Jia, Zhiguo Gong


Abstract
The widespread existence of wrongly labeled instances is a challenge to distantly supervised relation extraction. Most of the previous works are trained in a bag-level setting to alleviate such noise. However, sentence-level training better utilizes the information than bag-level training, as long as combined with effective noise alleviation. In this work, we propose a novel Transitive Instance Weighting mechanism integrated with the self-distilled BERT backbone, utilizing information in the intermediate outputs to generate dynamic instance weights for denoised sentence-level training. By down-weighting wrongly labeled instances and discounting the weights of easy-to-fit ones, our method can effectively tackle wrongly labeled instances and prevent overfitting. Experiments on both held-out and manual datasets indicate that our method achieves state-of-the-art performance and consistent improvements over the baselines.
Anthology ID:
2023.findings-emnlp.13
Volume:
Findings of the Association for Computational Linguistics: EMNLP 2023
Month:
December
Year:
2023
Address:
Singapore
Editors:
Houda Bouamor, Juan Pino, Kalika Bali
Venue:
Findings
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
168–180
Language:
URL:
https://aclanthology.org/2023.findings-emnlp.13
DOI:
10.18653/v1/2023.findings-emnlp.13
Bibkey:
Cite (ACL):
Xiangyu Lin, Weijia Jia, and Zhiguo Gong. 2023. Self-distilled Transitive Instance Weighting for Denoised Distantly Supervised Relation Extraction. In Findings of the Association for Computational Linguistics: EMNLP 2023, pages 168–180, Singapore. Association for Computational Linguistics.
Cite (Informal):
Self-distilled Transitive Instance Weighting for Denoised Distantly Supervised Relation Extraction (Lin et al., Findings 2023)
Copy Citation:
PDF:
https://preview.aclanthology.org/dois-2013-emnlp/2023.findings-emnlp.13.pdf