A Simple and Effective Model for Answering Multi-span Questions
This addresses a limitation in existing models that cannot handle multi-span answers, offering a practical solution for reading comprehension tasks.
The paper tackles the problem of answering multi-span questions in reading comprehension by proposing a simple sequence tagging model, achieving improvements of 9.9 and 5.5 EM points on DROP and Quoref datasets, respectively.
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.