On Sample Based Explanation Methods for NLP: Faithfulness, Efficiency and Semantic Evaluation

Wei Zhang, Ziming Huang, Yada Zhu, Guangnan Ye, Xiaodong Cui, Fan Zhang


Abstract
In the recent advances of natural language processing, the scale of the state-of-the-art models and datasets is usually extensive, which challenges the application of sample-based explanation methods in many aspects, such as explanation interpretability, efficiency, and faithfulness. In this work, for the first time, we can improve the interpretability of explanations by allowing arbitrary text sequences as the explanation unit. On top of this, we implement a hessian-free method with a model faithfulness guarantee. Finally, to compare our method with the others, we propose a semantic-based evaluation metric that can better align with humans’ judgment of explanations than the widely adopted diagnostic or re-training measures. The empirical results on multiple real data sets demonstrate the proposed method’s superior performance to popular explanation techniques such as Influence Function or TracIn on semantic evaluation.
Anthology ID:
2021.acl-long.419
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
Editors:
Chengqing Zong, Fei Xia, Wenjie Li, Roberto Navigli
Venues:
ACL | IJCNLP
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
5399–5411
Language:
URL:
https://aclanthology.org/2021.acl-long.419
DOI:
10.18653/v1/2021.acl-long.419
Bibkey:
Cite (ACL):
Wei Zhang, Ziming Huang, Yada Zhu, Guangnan Ye, Xiaodong Cui, and Fan Zhang. 2021. On Sample Based Explanation Methods for NLP: Faithfulness, Efficiency and Semantic Evaluation. 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 5399–5411, Online. Association for Computational Linguistics.
Cite (Informal):
On Sample Based Explanation Methods for NLP: Faithfulness, Efficiency and Semantic Evaluation (Zhang et al., ACL-IJCNLP 2021)
Copy Citation:
PDF:
https://aclanthology.org/2021.acl-long.419.pdf
Optional supplementary material:
 2021.acl-long.419.OptionalSupplementaryMaterial.zip
Video:
 https://aclanthology.org/2021.acl-long.419.mp4
Data
MAMS