A Simple and Effective Model for Answering Multi-span Questions

TL;DR

This paper introduces a straightforward sequence tagging model for multi-span question answering, enabling models to predict multiple non-contiguous answer spans, thus improving performance on relevant datasets.

Contribution

It presents a simple, effective approach to multi-span question answering by framing it as a sequence tagging task, expanding beyond single-span limitations.

Findings

Improved EM scores on DROP and Quoref datasets

Model outperforms previous span extraction methods

Demonstrates effectiveness of sequence tagging for multi-span answers

Abstract

Models for reading comprehension (RC) commonly restrict their output space to the set of all single contiguous spans from the input, in order to alleviate the learning problem and avoid the need for a model that generates text explicitly. However, forcing an answer to be a single span can be restrictive, and some recent datasets also include multi-span questions, i.e., questions whose answer is a set of non-contiguous spans in the text. Naturally, models that return single spans cannot answer these questions. In this work, we propose a simple architecture for answering multi-span questions by casting the task as a sequence tagging problem, namely, predicting for each input token whether it should be part of the output or not. Our model substantially improves performance on span extraction questions from DROP and Quoref by 9.9 and 5.5 EM points respectively.