U-Net: Machine Reading Comprehension with Unanswerable Questions
Fu Sun, Linyang Li, Xipeng Qiu, Yang Liu Shanghai Key Laboratory of Intelligent Information Processing, Fudan University School of Computer Science, Fudan University Liulishuo Silicon Valley AI Lab Corresponding Author.
Abstract
Machine reading comprehension with unanswerable questions is a new challenging task for natural language processing. A key subtask is to reliably predict whether the question is unanswerable. In this paper, we propose a unified model, called U-Net, with three important components: answer pointer, no-answer pointer, and answer verifier. We introduce a universal node and thus process the question and its context passage as a single contiguous sequence of tokens. The universal node encodes the fused information from both the question and passage, and plays an important role to predict whether the question is answerable and also greatly improves the conciseness of the U-Net. Different from the state-of-art pipeline models, U-Net can be learned in an end-to-end fashion. The experimental results on the SQuAD 2.0 dataset show that U-Net can effectively predict the unanswerability of questions and achieves an F1 score of 71.7 on SQuAD 2.0.
中文速览
让机器在阅读理解时不仅能找到答案，还能判断"这道题根本无法从文章中回答"，是当前自然语言处理的核心难题之一。研究者提出了一个名为 U-Net 的统一模型，将"找答案边界""指向无答案位置"和"验证是否可回答"三项子任务融合进同一个端到端框架，并引入一个"通用节点"（universal node）把问题和文章拼成一条连续序列共同编码，让该节点天然汇聚双侧信息，从而既简化了模型结构，又为可答性判断提供了关键的融合表示。在 SQuAD 2.0 数据集上，U-Net 取得了 71.7 的 F1 分数，相比以往需要分步训练的流水线方法更加简洁高效。这项工作的意义在于：它证明了可答性判断与答案抽取可以在一个共享参数的模型里同时学习，为构建"知道自己不知道"的阅读理解系统提供了切实可行的思路。
原文 arXiv:1810.06638;中英对照 + 大白话阅读 https://aha.fim.ai/paper/1810.06638v1