Modeling Question Asking Using Neural Program Generation
Ziyun Wang Department of Computer Science New York University、Brenden M. Lake Department of Psychology and Center for Data Science New York University Ziyun is now at Tencent.
Abstract
People ask questions that are far richer, more informative, and more creative than current AI systems. We propose a neuro-symbolic framework for modeling human question asking, which represents questions as formal programs and generates programs with an encoder-decoder based deep neural network. From extensive experiments using an information-search game, we show that our method can predict which questions humans are likely to ask in unconstrained settings. We also propose a novel grammar-based question generation framework trained with reinforcement learning, which is able to generate creative questions without supervised human data.
中文速览
人类提问的方式远比现有AI系统更丰富、更有创造力，如何让机器模拟这种能力一直是计算建模的难题。研究者提出了一种神经符号（neuro-symbolic）框架，将问题表示为可执行的形式化程序，并用卷积编码器加Transformer解码器的深度神经网络来生成这些程序；此外还设计了一套基于文法的强化学习训练方案，使模型无需人类示范数据也能自主生成有价值的问题。在类似"海战棋"的信息搜索游戏上进行大量实验后，模型不仅能准确预测人类在开放场景中倾向于提出哪些问题，还能通过强化学习生成全新的创造性问题，同时保留与人类行为相符的关键特征。这项工作填补了人机提问能力之间的显著鸿沟，为构建更接近人类认知方式的主动学习系统提供了切实可行的框架。
原文 arXiv:1907.09899;中英对照 + 大白话阅读 https://aha.fim.ai/paper/1907.09899v4