Calibrating Trust of Multi-Hop Question Answering Systems with Decompositional Probes

TL;DR

This paper investigates using question decomposition as a probing method to improve explainability and trust calibration in multi-hop QA systems, demonstrating that exposing decompositions helps users predict system correctness.

Contribution

It introduces decompositional probes as a novel approach for explaining multi-hop QA systems and shows their effectiveness in helping users assess system performance.

Findings

Decomposition probes improve user ability to predict QA system correctness.

Exposing decomposition results enhances explainability of multi-hop QA.

Current decomposition systems require further improvements.

Abstract

Multi-hop Question Answering (QA) is a challenging task since it requires an accurate aggregation of information from multiple context paragraphs and a thorough understanding of the underlying reasoning chains. Recent work in multi-hop QA has shown that performance can be boosted by first decomposing the questions into simpler, single-hop questions. In this paper, we explore one additional utility of the multi-hop decomposition from the perspective of explainable NLP: to create explanation by probing a neural QA model with them. We hypothesize that in doing so, users will be better able to predict when the underlying QA system will give the correct answer. Through human participant studies, we verify that exposing the decomposition probes and answers to the probes to users can increase their ability to predict system performance on a question instance basis. We show that decomposition is an effective form of probing QA systems as well as a promising approach to explanation generation. In-depth analyses show the need for improvements in decomposition systems.