Think you have Solved Question Answering? Try ARC, the AI2 Reasoning Challenge

TL;DR

The paper introduces the AI2 Reasoning Challenge (ARC), a large dataset of grade-school science questions designed to push AI systems towards advanced reasoning and knowledge understanding, highlighting current models' limitations.

Contribution

The paper presents the ARC dataset, including Challenge and Easy sets, along with a large corpus and baseline models, to foster progress in complex question answering.

Findings

Current neural models perform no better than random on the Challenge Set.

ARC is the largest public dataset of its kind for science questions.

Baseline models struggle with advanced reasoning required by ARC.

Abstract

We present a new question set, text corpus, and baselines assembled to encourage AI research in advanced question answering. Together, these constitute the AI2 Reasoning Challenge (ARC), which requires far more powerful knowledge and reasoning than previous challenges such as SQuAD or SNLI. The ARC question set is partitioned into a Challenge Set and an Easy Set, where the Challenge Set contains only questions answered incorrectly by both a retrieval-based algorithm and a word co-occurence algorithm. The dataset contains only natural, grade-school science questions (authored for human tests), and is the largest public-domain set of this kind (7,787 questions). We test several baselines on the Challenge Set, including leading neural models from the SQuAD and SNLI tasks, and find that none are able to significantly outperform a random baseline, reflecting the difficult nature of this task. We are also releasing the ARC Corpus, a corpus of 14M science sentences relevant to the task, and implementations of the three neural baseline models tested. Can your model perform better? We pose ARC as a challenge to the community.