Interactive Fiction Games: A Colossal Adventure

TL;DR

Interactive Fiction games serve as a complex testbed for developing and evaluating language-based autonomous agents, combining challenges of language understanding, reasoning, and action planning.

Contribution

The paper introduces Jericho, a new environment for developing and testing agents in Interactive Fiction games, and provides a comprehensive study of agent performance.

Findings

Identified key challenges in language understanding and reasoning in IF games.

Benchmark results showing current agents' limitations.

Highlighted future directions for improving language-based agents.

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.