Inserting Information Bottlenecks for Attribution in Transformers

Zhiying Jiang, Raphael Tang, Ji Xin, Jimmy Lin


Abstract
Pretrained transformers achieve the state of the art across tasks in natural language processing, motivating researchers to investigate their inner mechanisms. One common direction is to understand what features are important for prediction. In this paper, we apply information bottlenecks to analyze the attribution of each feature for prediction on a black-box model. We use BERT as the example and evaluate our approach both quantitatively and qualitatively. We show the effectiveness of our method in terms of attribution and the ability to provide insight into how information flows through layers. We demonstrate that our technique outperforms two competitive methods in degradation tests on four datasets. Code is available at https://github.com/bazingagin/IBA.
Anthology ID:
2020.findings-emnlp.343
Volume:
Findings of the Association for Computational Linguistics: EMNLP 2020
Month:
November
Year:
2020
Address:
Online
Venue:
Findings
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
3850–3857
Language:
URL:
https://aclanthology.org/2020.findings-emnlp.343
DOI:
10.18653/v1/2020.findings-emnlp.343
Bibkey:
Cite (ACL):
Zhiying Jiang, Raphael Tang, Ji Xin, and Jimmy Lin. 2020. Inserting Information Bottlenecks for Attribution in Transformers. In Findings of the Association for Computational Linguistics: EMNLP 2020, pages 3850–3857, Online. Association for Computational Linguistics.
Cite (Informal):
Inserting Information Bottlenecks for Attribution in Transformers (Jiang et al., Findings 2020)
Copy Citation:
PDF:
https://preview.aclanthology.org/auto-file-uploads/2020.findings-emnlp.343.pdf
Code
 bazingagin/IBA
Data
GLUEIMDb Movie ReviewsMultiNLI