Attention Calibration for Transformer in Neural Machine Translation

Yu Lu, Jiali Zeng, Jiajun Zhang, Shuangzhi Wu, Mu Li


Abstract
Attention mechanisms have achieved substantial improvements in neural machine translation by dynamically selecting relevant inputs for different predictions. However, recent studies have questioned the attention mechanisms’ capability for discovering decisive inputs. In this paper, we propose to calibrate the attention weights by introducing a mask perturbation model that automatically evaluates each input’s contribution to the model outputs. We increase the attention weights assigned to the indispensable tokens, whose removal leads to a dramatic performance decrease. The extensive experiments on the Transformer-based translation have demonstrated the effectiveness of our model. We further find that the calibrated attention weights are more uniform at lower layers to collect multiple information while more concentrated on the specific inputs at higher layers. Detailed analyses also show a great need for calibration in the attention weights with high entropy where the model is unconfident about its decision.
Anthology ID:
2021.acl-long.103
Volume:
Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (Volume 1: Long Papers)
Month:
August
Year:
2021
Address:
Online
Venues:
ACL | IJCNLP
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
1288–1298
Language:
URL:
https://aclanthology.org/2021.acl-long.103
DOI:
10.18653/v1/2021.acl-long.103
Bibkey:
Cite (ACL):
Yu Lu, Jiali Zeng, Jiajun Zhang, Shuangzhi Wu, and Mu Li. 2021. Attention Calibration for Transformer in Neural Machine Translation. In Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (Volume 1: Long Papers), pages 1288–1298, Online. Association for Computational Linguistics.
Cite (Informal):
Attention Calibration for Transformer in Neural Machine Translation (Lu et al., ACL 2021)
Copy Citation:
PDF:
https://aclanthology.org/2021.acl-long.103.pdf