Which Shortcut Solution Do Question Answering Models Prefer to Learn?

TL;DR

This paper investigates the learnability of shortcut solutions in question answering models, revealing how certain shortcuts are preferentially learned and how this knowledge can improve training strategies to enhance model robustness.

Contribution

The study analyzes the learnability of shortcuts in QA models, linking it to loss landscape characteristics and proposing a method to optimize training set composition based on shortcut learnability.

Findings

Shortcut solutions exploiting answer positions and word-label correlations are preferentially learned.

More learnable shortcuts correspond to flatter, deeper loss landscapes.

Utilizing shortcut learnability reduces the need for anti-shortcut examples in training.

Abstract

Question answering (QA) models for reading comprehension tend to learn shortcut solutions rather than the solutions intended by QA datasets. QA models that have learned shortcut solutions can achieve human-level performance in shortcut examples where shortcuts are valid, but these same behaviors degrade generalization potential on anti-shortcut examples where shortcuts are invalid. Various methods have been proposed to mitigate this problem, but they do not fully take the characteristics of shortcuts themselves into account. We assume that the learnability of shortcuts, i.e., how easy it is to learn a shortcut, is useful to mitigate the problem. Thus, we first examine the learnability of the representative shortcuts on extractive and multiple-choice QA datasets. Behavioral tests using biased training sets reveal that shortcuts that exploit answer positions and word-label correlations are preferentially learned for extractive and multiple-choice QA, respectively. We find that the more learnable a shortcut is, the flatter and deeper the loss landscape is around the shortcut solution in the parameter space. We also find that the availability of the preferred shortcuts tends to make the task easier to perform from an information-theoretic viewpoint. Lastly, we experimentally show that the learnability of shortcuts can be utilized to construct an effective QA training set; the more learnable a shortcut is, the smaller the proportion of anti-shortcut examples required to achieve comparable performance on shortcut and anti-shortcut examples. We claim that the learnability of shortcuts should be considered when designing mitigation methods.