Language Understanding for Text-based Games Using Deep Reinforcement Learning

TL;DR

This paper introduces a deep reinforcement learning approach to enable agents to understand and play text-based games by learning semantic representations of game states from textual descriptions.

Contribution

It presents a novel framework that jointly learns state representations and action policies directly from text, improving performance over traditional bag-of-words methods.

Findings

Our method outperforms baselines in two game worlds.

Learning semantic representations enhances game-playing performance.

Joint learning of state and policy is effective for text-based environments.

Abstract

In this paper, we consider the task of learning control policies for text-based games. In these games, all interactions in the virtual world are through text and the underlying state is not observed. The resulting language barrier makes such environments challenging for automatic game players. We employ a deep reinforcement learning framework to jointly learn state representations and action policies using game rewards as feedback. This framework enables us to map text descriptions into vector representations that capture the semantics of the game states. We evaluate our approach on two game worlds, comparing against baselines using bag-of-words and bag-of-bigrams for state representations. Our algorithm outperforms the baselines on both worlds demonstrating the importance of learning expressive representations.