Improving Lexical Embeddings for Robust Question Answering

TL;DR

This paper introduces ESC, a method to enhance lexical embeddings in QA models by applying semantic and contextual constraints, significantly improving robustness against adversarial examples.

Contribution

The paper proposes a novel representation enhancement technique using semantic and context constraints to improve QA model robustness and generalization.

Findings

Significant robustness gains on four adversarial test sets

Improved ability to distinguish context clues for correct answers

Enhanced lexical embeddings lead to better model resilience

Abstract

Recent techniques in Question Answering (QA) have gained remarkable performance improvement with some QA models even surpassed human performance. However, the ability of these models in truly understanding the language still remains dubious and the models are revealing limitations when facing adversarial examples. To strengthen the robustness of QA models and their generalization ability, we propose a representation Enhancement via Semantic and Context constraints (ESC) approach to improve the robustness of lexical embeddings. Specifically, we insert perturbations with semantic constraints and train enhanced contextual representations via a context-constraint loss to better distinguish the context clues for the correct answer. Experimental results show that our approach gains significant robustness improvement on four adversarial test sets.