Towards Robust Extractive Question Answering Models: Rethinking the Training Methodology

Son Quoc Tran, Matt Kretchmar

Abstract
This paper proposes a novel training method to improve the robustness of Extractive Question Answering (EQA) models. Previous research has shown that existing models, when trained on EQA datasets that include unanswerable questions, demonstrate a significant lack of robustness against distribution shifts and adversarial attacks. Despite this, the inclusion of unanswerable questions in EQA training datasets is essential for ensuring real-world reliability. Our proposed training method includes a novel loss function for the EQA problem and challenges an implicit assumption present in numerous EQA datasets. Models trained with our method maintain in-domain performance while achieving a notable improvement on out-of-domain datasets. This results in an overall F1 score improvement of 5.7 across all testing sets. Furthermore, our models exhibit significantly enhanced robustness against two types of adversarial attacks, with a performance decrease of only about one-third compared to the default models.
Anthology ID:
2024.findings-emnlp.121
Volume:
Findings of the Association for Computational Linguistics: EMNLP 2024
Month:
November
Year:
2024
Address:
Miami, Florida, USA
Editors:
Yaser Al-Onaizan, Mohit Bansal, Yun-Nung Chen
Venue:
Findings
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
2222–2236
Language:
URL:
https://preview.aclanthology.org/add_missing_videos/2024.findings-emnlp.121/
DOI:
10.18653/v1/2024.findings-emnlp.121
Bibkey:
Cite (ACL):
Son Quoc Tran and Matt Kretchmar. 2024. Towards Robust Extractive Question Answering Models: Rethinking the Training Methodology. In Findings of the Association for Computational Linguistics: EMNLP 2024, pages 2222–2236, Miami, Florida, USA. Association for Computational Linguistics.
Cite (Informal):
Towards Robust Extractive Question Answering Models: Rethinking the Training Methodology (Tran & Kretchmar, Findings 2024)
Copy Citation:
PDF:
https://preview.aclanthology.org/add_missing_videos/2024.findings-emnlp.121.pdf
PDF Search Fix data