Is Sluice Resolution really just Question Answering?

TL;DR

This paper investigates sluice resolution, a task involving identifying antecedents of elided content behind wh-words, framing it as question answering and analyzing what distinguishes it from standard QA, with improved system performance.

Contribution

The paper explores the differences between sluice resolution and question answering, and demonstrates that recent QA systems can significantly improve sluice resolution performance.

Findings

QA systems improve sluice resolution F1 from 86.01 to 90.39.

Sluice resolution shares similarities with QA but has unique challenges.

Further analysis of error gaps reveals specific differences from QA.

Abstract

Sluice resolution is a problem where a system needs to output the corresponding antecedents of wh-ellipses. The antecedents are elided contents behind the wh-words but are implicitly referred to using contexts. Previous work frames sluice resolution as question answering where this setting outperforms all its preceding works by large margins. Ellipsis and questions are referentially dependent expressions (anaphoras) and retrieving the corresponding antecedents are like answering questions to output pieces of clarifying information. However, the task is not fully solved. Therefore, we want to further investigate what makes sluice resolution differ to question answering and fill in the error gaps. We also present some results using recent state-of-the-art question answering systems which improve the previous work (86.01 to 90.39 F1).