Single-Turn Debate Does Not Help Humans Answer Hard Reading-Comprehension Questions

TL;DR

This paper investigates whether single-turn debates with explanations help humans answer difficult reading comprehension questions, finding that explanations do not improve accuracy but selected snippets do.

Contribution

The study introduces a debate-style dataset with explanations for both correct and incorrect answers to evaluate their effectiveness in aiding human comprehension.

Findings

Explanations did not improve human accuracy in answering questions.

Providing human-selected text snippets increased accuracy.

The results suggest improvements for future debate-based data collection.

Abstract

Current QA systems can generate reasonable-sounding yet false answers without explanation or evidence for the generated answer, which is especially problematic when humans cannot readily check the model's answers. This presents a challenge for building trust in machine learning systems. We take inspiration from real-world situations where difficult questions are answered by considering opposing sides (see Irving et al., 2018). For multiple-choice QA examples, we build a dataset of single arguments for both a correct and incorrect answer option in a debate-style set-up as an initial step in training models to produce explanations for two candidate answers. We use long contexts -- humans familiar with the context write convincing explanations for pre-selected correct and incorrect answers, and we test if those explanations allow humans who have not read the full context to more accurately determine the correct answer. We do not find that explanations in our set-up improve human accuracy, but a baseline condition shows that providing human-selected text snippets does improve accuracy. We use these findings to suggest ways of improving the debate set up for future data collection efforts.