Break It Down: A Question Understanding Benchmark

TL;DR

This paper introduces QDMR, a structured representation of question decomposition, along with a large dataset, demonstrating its utility in question answering and semantic parsing tasks.

Contribution

It presents a scalable crowdsourcing method for annotating QDMR, releases a large dataset, and shows how QDMR improves question answering and semantic parsing performance.

Findings

QDMR improves open-domain question answering accuracy.

The dataset contains over 83,000 question-QDMR pairs.

A sequence-to-sequence model effectively parses questions into QDMR structures.

Abstract

Understanding natural language questions entails the ability to break down a question into the requisite steps for computing its answer. In this work, we introduce a Question Decomposition Meaning Representation (QDMR) for questions. QDMR constitutes the ordered list of steps, expressed through natural language, that are necessary for answering a question. We develop a crowdsourcing pipeline, showing that quality QDMRs can be annotated at scale, and release the Break dataset, containing over 83K pairs of questions and their QDMRs. We demonstrate the utility of QDMR by showing that (a) it can be used to improve open-domain question answering on the HotpotQA dataset, (b) it can be deterministically converted to a pseudo-SQL formal language, which can alleviate annotation in semantic parsing applications. Last, we use Break to train a sequence-to-sequence model with copying that parses questions into QDMR structures, and show that it substantially outperforms several natural baselines.