RikiNet: Reading Wikipedia Pages for Natural Question Answering
Dayiheng Liu , Yeyun Gong, Jie Fu, Yu Yan, Jiusheng Chen, Daxin Jiang, Jiancheng Lv, Nan Duan College of Computer Science, Sichuan University Microsoft Research Asia Mila, Polytechnique Montreal Microsoft AI and Research Microsoft Search Technology Center Asia Work is done during internship at Microsoft Research Asia.
Abstract
Reading long documents to answer open-domain questions remains challenging in natural language understanding. In this paper, we introduce a new model, called RikiNet, which reads Wikipedia pages for natural question answering. RikiNet contains a dynamic paragraph dual-attention reader and a multi-level cascaded answer predictor. The reader dynamically represents the document and question by utilizing a set of complementary attention mechanisms. The representations are then fed into the predictor to obtain the span of the short answer, the paragraph of the long answer, and the answer type in a cascaded manner. On the Natural Questions (NQ) dataset, a single RikiNet achieves 74.3 F1 and 57.9 F1 on long-answer and short-answer tasks. To our best knowledge, it is the first single model that outperforms the single human performance. Furthermore, an ensemble RikiNet obtains 76.1 F1 and 61.3 F1 on long-answer and short-answer tasks, achieving the best performance on the official NQ leaderboard111Till our submission time, 29 Nov. 2019. We refer readers to https://ai.google.com/research/NaturalQuestions/leaderboard for the latest results..
中文速览
用真实用户搜索谷歌时产生的问题去挑战机器阅读理解（Machine Reading Comprehension, MRC），比以往数据集难得多——模型得读完整篇维基百科长文，同时给出"长答案"（包含答案的段落）和"短答案"（具体实体片段），而不只是在一小段话里划线。针对这个挑战，研究者提出了 RikiNet 模型：核心是一个"动态段落双注意力（Dynamic Paragraph Dual-Attention, DPDA）读取器"，通过多组互补的注意力机制在词级和段落级同时表示长文档与问题，再搭配一个"多层级级联答案预测器"，按照"先定位长答案段落→再预测短答案起止位置→最后判断答案类型"的顺序逐级推断，充分利用长短答案之间的包含关系。在 Natural Questions（NQ）官方测试集上，单个 RikiNet 在长答案和短答案任务上分别达到 74.3 和 57.9 的 F1 分数，首次以单一模型超越了人类单人表现；集成版本进一步达到 76.1 和 61.3，刷新了官方排行榜最优成绩。这项工作表明，通过针对性地设计段落级注意力与级联预测结构，机器有望在面向真实搜索场景的长文档问答中向人类水平发起实质性挑战。
原文 arXiv:2004.14560;中英对照 + 大白话阅读 https://aha.fim.ai/paper/2004.14560v1