Retrieve-and-Read: Multi-task Learning of Information Retrieval and Reading Comprehension
Kyosuke Nishida, Itsumi Saito, Atsushi Otsuka, Hisako Asano, and Junji Tomita NTT Media Intelligence Laboratories, NTT Corporation
Abstract
This study considers the task of machine reading at scale (MRS) wherein, given a question, a system first performs the information retrieval (IR) task of finding relevant passages in a knowledge source and then carries out the reading comprehension (RC) task of extracting an answer span from the passages. Previous MRS studies, in which the IR component was trained without considering answer spans, struggled to accurately find a small number of relevant passages from a large set of passages. In this paper, we propose a simple and effective approach that incorporates the IR and RC tasks by using supervised multi-task learning in order that the IR component can be trained by considering answer spans. Experimental results on the standard benchmark, answering SQuAD questions using the full Wikipedia as the knowledge source, showed that our model achieved state-of-the-art performance. Moreover, we thoroughly evaluated the individual contributions of our model components with our new Japanese dataset and SQuAD. The results showed significant improvements in the IR task and provided a new perspective on IR for RC: it is effective to teach which part of the passage answers the question rath
中文速览
大规模开放域问答系统需要先从海量文章中找到相关段落，再从中抽取答案，而以往研究中检索（信息检索，IR）与阅读理解（RC）两个环节是分开训练的，导致检索环节成为整个系统的瓶颈。本文提出了一种名为"Retrieve-and-Read"的神经网络模型，通过有监督的多任务学习（multi-task learning）让IR与RC共享底层隐藏层，并同时优化"段落相关性得分"和"答案片段位置"两个损失，从而让检索模块在训练时能直接参考答案所在的具体位置，而不只是粗略的段落相关标签。在以整个英文维基百科为知识源、回答SQuAD问题的标准测试上，该模型取得了当时的最优结果，在新构建的日语新闻数据集上也验证了检索准确率的显著提升。这项工作揭示了一个重要洞见：告诉检索模型"答案在段落的哪个位置"，比仅给整段文章打一个相关性标签要有效得多，为后续开放域问答研究提供了新的设计思路。
原文 arXiv:1808.10628;中英对照 + 大白话阅读 https://aha.fim.ai/paper/1808.10628v1