English

Robust Natural Language Understanding with Residual Attention Debiasing

Computation and Language 2023-05-30 v1 Artificial Intelligence Machine Learning

Abstract

Natural language understanding (NLU) models often suffer from unintended dataset biases. Among bias mitigation methods, ensemble-based debiasing methods, especially product-of-experts (PoE), have stood out for their impressive empirical success. However, previous ensemble-based debiasing methods typically apply debiasing on top-level logits without directly addressing biased attention patterns. Attention serves as the main media of feature interaction and aggregation in PLMs and plays a crucial role in providing robust prediction. In this paper, we propose REsidual Attention Debiasing (READ), an end-to-end debiasing method that mitigates unintended biases from attention. Experiments on three NLU tasks show that READ significantly improves the performance of BERT-based models on OOD data with shortcuts removed, including +12.9% accuracy on HANS, +11.0% accuracy on FEVER-Symmetric, and +2.7% F1 on PAWS. Detailed analyses demonstrate the crucial role of unbiased attention in robust NLU models and that READ effectively mitigates biases in attention. Code is available at https://github.com/luka-group/READ.

Keywords

Cite

@article{arxiv.2305.17627,
  title  = {Robust Natural Language Understanding with Residual Attention Debiasing},
  author = {Fei Wang and James Y. Huang and Tianyi Yan and Wenxuan Zhou and Muhao Chen},
  journal= {arXiv preprint arXiv:2305.17627},
  year   = {2023}
}

Comments

ACL 2023 Findings

R2 v1 2026-06-28T10:48:34.121Z