Summarize-then-Answer: Generating Concise Explanations for Multi-hop Reading Comprehension
This addresses the need for minimally sufficient explanations in multi-hop reading comprehension, offering a semi-supervised method that is incremental over existing extractive strategies.
The paper tackles the problem of generating concise explanations for multi-hop reading comprehension by proposing an abstractive summarization approach, which produces more compact explanations than extractive methods while maintaining sufficiency, as demonstrated with only 2k supervised instances.
How can we generate concise explanations for multi-hop Reading Comprehension (RC)? The current strategies of identifying supporting sentences can be seen as an extractive question-focused summarization of the input text. However, these extractive explanations are not necessarily concise i.e. not minimally sufficient for answering a question. Instead, we advocate for an abstractive approach, where we propose to generate a question-focused, abstractive summary of input paragraphs and then feed it to an RC system. Given a limited amount of human-annotated abstractive explanations, we train the abstractive explainer in a semi-supervised manner, where we start from the supervised model and then train it further through trial and error maximizing a conciseness-promoted reward function. Our experiments demonstrate that the proposed abstractive explainer can generate more compact explanations than an extractive explainer with limited supervision (only 2k instances) while maintaining sufficiency.