Learning through Dialogue Interactions by Asking Questions
Jiwei Li, Alexander H. Miller, Sumit Chopra, Marc’Aurelio Ranzato, Jason Weston Facebook AI Research, New York, USA
Abstract
A good dialogue agent should have the ability to interact with users by both responding to questions and by asking questions, and importantly to learn from both types of interaction. In this work, we explore this direction by designing a simulator and a set of synthetic tasks in the movie domain that allow such interactions between a learner and a teacher. We investigate how a learner can benefit from asking questions in both offline and online reinforcement learning settings, and demonstrate that the learner improves when asking questions. Finally, real experiments with Mechanical Turk validate the approach. Our work represents a first step in developing such end-to-end learned interactive dialogue agents.
中文速览
对话机器人通常只会被动地回答问题，却不懂得在不确定时主动向用户提问来帮助自己学习，这篇论文就是要填补这一空白。研究者在电影问答领域设计了一套模拟器和合成任务，覆盖三类典型困境——听不懂问题（问题澄清）、找不到推理路径（知识操作）、知识库本身有缺口（知识获取）——让机器人学会在恰当时机向"老师"提问，并利用对方的反馈来提升后续表现。他们同时在离线监督学习和在线强化学习（reinforcement learning）两种设置下验证了这一框架，并通过 Amazon Mechanical Turk 的真实人工实验进一步确认：会提问的机器人在所有任务上的答题准确率都明显高于只被动作答的版本。这项工作意义在于迈出了构建"既能答又能问、还能从交互中持续学习"的端到端对话智能体的第一步，为未来更自然、更自适应的人机对话系统奠定了基础。
原文 arXiv:1612.04936;中英对照 + 大白话阅读 https://aha.fim.ai/paper/1612.04936v4