Learning Language Games through Interaction
Sida I. Wang Percy Liang Christopher D. Manning Computer Science Department Stanford University
Abstract
We introduce a new language learning setting relevant to building adaptive natural language interfaces. It is inspired by Wittgenstein’s language games: a human wishes to accomplish some task (e.g., achieving a certain configuration of blocks), but can only communicate with a computer, who performs the actual actions (e.g., removing all red blocks). The computer initially knows nothing about language and therefore must learn it from scratch through interaction, while the human adapts to the computer’s capabilities. We created a game called SHRDLURN in a blocks world and collected interactions from 100 people playing it. First, we analyze the humans’ strategies, showing that using compositionality and avoiding synonyms correlates positively with task performance. Second, we compare computer strategies, showing that modeling pragmatics on a semantic parsing model accelerates learning for more strategic players.
中文速览
人类和计算机共同完成积木摆放任务时，双方最初没有共同语言，计算机需要从零开始通过互动学习理解人类的指令。研究者设计了一款名为 SHRDLURN 的积木游戏，招募100名玩家与系统实时交互，收集了约1万条对话数据，同时构建了一套基于语义解析（semantic parsing）的在线学习模型，并引入语用推理（pragmatics）机制——让计算机像孩子习得语言时那样运用"互斥原则"，主动排除已有解释、加速学习新词义。实验结果表明，加入语用模型后，对表现最好的10位玩家的在线准确率提升了8%；对人类行为的分析也发现，使用组合性表达、避免同义词的玩家任务完成得更好，且随着游戏推进许多人会自发地调整说话方式以适应计算机。这项研究揭示了构建真正自适应自然语言界面的关键挑战——系统与用户需要相互磨合、共同"协商"出一套可用的语言，对资源匮乏语言和全新应用领域的自然语言处理具有重要的现实意义。
原文 arXiv:1606.02447;中英对照 + 大白话阅读 https://aha.fim.ai/paper/1606.02447v1