On the Robustness of Reading Comprehension Models to Entity Renaming

TL;DR

This paper investigates how machine reading comprehension models' accuracy is affected by renaming entities in questions, revealing their reliance on entity names and proposing methods to improve robustness.

Contribution

It introduces a scalable pipeline to test model robustness to entity renaming and compares different models and training strategies for improved resilience.

Findings

Models perform worse when entities are renamed across datasets.

SpanBERT is more robust than RoBERTa despite similar baseline accuracy.

Entity-based masking during pretraining enhances model robustness.

Abstract

We study the robustness of machine reading comprehension (MRC) models to entity renaming -- do models make more wrong predictions when the same questions are asked about an entity whose name has been changed? Such failures imply that models overly rely on entity information to answer questions, and thus may generalize poorly when facts about the world change or questions are asked about novel entities. To systematically audit this issue, we present a pipeline to automatically generate test examples at scale, by replacing entity names in the original test sample with names from a variety of sources, ranging from names in the same test set, to common names in life, to arbitrary strings. Across five datasets and three pretrained model architectures, MRC models consistently perform worse when entities are renamed, with particularly large accuracy drops on datasets constructed via distant supervision. We also find large differences between models: SpanBERT, which is pretrained with span-level masking, is more robust than RoBERTa, despite having similar accuracy on unperturbed test data. We further experiment with different masking strategies as the continual pretraining objective and find that entity-based masking can improve the robustness of MRC models.