Abstract
While numerous methods have been proposed as defenses against adversarial examples in question answering (QA), these techniques are often model specific, require retraining of the model, and give only marginal improvements in performance over vanilla models. In this work, we present a simple model-agnostic approach to this problem that can be applied directly to any QA model without any retraining. Our method employs an explicit answer candidate reranking mechanism that scores candidate answers on the basis of their content overlap with the question before making the final prediction. Combined with a strong base QAmodel, our method outperforms state-of-the-art defense techniques, calling into question how well these techniques are actually doing and strong these adversarial testbeds are.- Anthology ID:
- 2021.eacl-srw.8
- Volume:
- Proceedings of the 16th Conference of the European Chapter of the Association for Computational Linguistics: Student Research Workshop
- Month:
- April
- Year:
- 2021
- Address:
- Online
- Venue:
- EACL
- SIG:
- Publisher:
- Association for Computational Linguistics
- Note:
- Pages:
- 50–57
- Language:
- URL:
- https://aclanthology.org/2021.eacl-srw.8
- DOI:
- 10.18653/v1/2021.eacl-srw.8
- Cite (ACL):
- Sagnik Majumder, Chinmoy Samant, and Greg Durrett. 2021. Model Agnostic Answer Reranking System for Adversarial Question Answering. In Proceedings of the 16th Conference of the European Chapter of the Association for Computational Linguistics: Student Research Workshop, pages 50–57, Online. Association for Computational Linguistics.
- Cite (Informal):
- Model Agnostic Answer Reranking System for Adversarial Question Answering (Majumder et al., EACL 2021)
- PDF:
- https://preview.aclanthology.org/paclic-22-ingestion/2021.eacl-srw.8.pdf
- Data
- SQuAD