Is a Question Decomposition Unit All We Need?

TL;DR

This paper investigates whether human-assisted question decomposition can improve NLP model performance, offering an alternative to building larger models by leveraging human input to simplify questions for models.

Contribution

It introduces Human-in-the-loop Question Decomposition (HQD), demonstrating significant performance improvements across multiple datasets without increasing model size.

Findings

Performance improved by 24% for GPT-3 and 29% for RoBERTa-SQuAD.

Question decomposition enhances model accuracy on reasoning tasks.

Human involvement in question simplification is a viable alternative to larger models.

Abstract

Large Language Models (LMs) have achieved state-of-the-art performance on many Natural Language Processing (NLP) benchmarks. With the growing number of new benchmarks, we build bigger and more complex LMs. However, building new LMs may not be an ideal option owing to the cost, time and environmental impact associated with it. We explore an alternative route: can we modify data by expressing it in terms of the model's strengths, so that a question becomes easier for models to answer? We investigate if humans can decompose a hard question into a set of simpler questions that are relatively easier for models to solve. We analyze a range of datasets involving various forms of reasoning and find that it is indeed possible to significantly improve model performance (24% for GPT3 and 29% for RoBERTa-SQuAD along with a symbolic calculator) via decomposition. Our approach provides a viable option to involve people in NLP research in a meaningful way. Our findings indicate that Human-in-the-loop Question Decomposition (HQD) can potentially provide an alternate path to building large LMs. Code and data is available at https://github.com/Pruthvi98/QuestionDecomposition