Interactive Fiction Games: A Colossal Adventure
Matthew Hausknecht Microsoft Research、Prithviraj Ammanabrolu Georgia Institute of Technology、Marc-Alexandre Côté Microsoft Research Montréal、Xingdi Yuan Microsoft Research Montréal
Abstract
A hallmark of human intelligence is the ability to understand and communicate with language. Interactive Fiction games are fully text-based simulation environments where a player issues text commands to effect change in the environment and progress through the story. We argue that IF games are an excellent testbed for studying language-based autonomous agents. In particular, IF games combine challenges of combinatorial action spaces, language understanding, and commonsense reasoning. To facilitate rapid development of language-based agents, we introduce Jericho, a learning environment for man-made IF games and conduct a comprehensive study of text-agents across a rich set of games, highlighting directions in which agents can improve.
中文速览
人类智能的核心之一是用语言理解和与世界互动，而互动小说（Interactive Fiction，IF）游戏——那种纯文字、靠输入命令推进剧情的老式游戏——恰好是测试这种能力的天然场景，因为玩好它需要同时具备自然语言理解、常识推理和长程序列决策三种能力。为此，研究者推出了一个名为 Jericho 的开源学习框架，专门对接数十款真人创作的 IF 游戏，并设计了一种"模板填空"动作生成机制，将原本高达 2400 亿的动作搜索空间压缩到约 9800 万，让强化学习智能体能够更高效地探索。基于 Jericho，论文系统评测了 DRRN、Template-DQN 和通用游戏智能体 NAIL 三类方法，在覆盖地下城、科幻、悬疑等多种类型的游戏上进行了大规模实验对比。结果表明，现有智能体在真人创作的 IF 游戏中表现仍远不如人类，揭示了语言理解、常识获取和世界知识表示等方面的明显短板，为未来基于语言的自主智能体研究指出了清晰的改进方向。
原文 arXiv:1909.05398;中英对照 + 大白话阅读 https://aha.fim.ai/paper/1909.05398v3