Multi-Passage Machine Reading Comprehension with Cross-Passage Answer Verification
Yizhong Wang *This work was done while the first author was doing internship at Baidu Inc. Kai Liu Baidu Inc., Beijing, China Jing Liu Baidu Inc., Beijing, China Wei He Baidu Inc., Beijing, China Yajuan Lyu Baidu Inc., Beijing, China Hua Wu Baidu Inc., Beijing, China Sujian Li Key Laboratory of Computational Linguistics, Peking University, MOE, China Haifeng Wang Baidu Inc., Beijing, China
Abstract
Machine reading comprehension (MRC) on real web data usually requires the machine to answer a question by analyzing multiple passages retrieved by search engine. Compared with MRC on a single passage, multi-passage MRC is more challenging, since we are likely to get multiple confusing answer candidates from different passages. To address this problem, we propose an end-to-end neural model that enables those answer candidates from different passages to verify each other based on their content representations. Specifically, we jointly train three modules that can predict the final answer based on three factors: the answer boundary, the answer content and the cross-passage answer verification. The experimental results show that our method outperforms the baseline by a large margin and achieves the state-of-the-art performance on the English MS-MARCO dataset and the Chinese DuReader dataset, both of which are designed for MRC in real-world settings.
中文速览
真实网页搜索场景下的阅读理解任务要求机器从多个检索段落中找出正确答案，难点在于不同段落往往各自给出措辞相近却真假混杂的候选答案，模型很难辨别孰对孰错。针对这一问题，研究者提出了一个端到端神经网络模型，将答案边界预测、答案内容建模和跨段落答案互验三个模块联合训练：先从每个段落中抽取候选答案，再让各段落的候选答案通过注意力机制相互"对照"，借助彼此的语义相似度来判断谁更可信，最终综合边界分数、内容分数和验证分数共同决定最终答案。实验在英文MS-MARCO和中文DuReader两个真实搜索数据集上均取得了当时最优的成绩，显著超越基线模型。这项工作的意义在于，它首次将"多来源交叉验证"这一人类阅读习惯显式地引入机器阅读理解模型，为解决多段落噪声干扰问题提供了一条切实可行的思路。
原文 arXiv:1805.02220;中英对照 + 大白话阅读 https://aha.fim.ai/paper/1805.02220v2