In plain words: A player reads each text description and, learning by trial and error from the game's score, works out what the situation means and which command to type next. On two text game worlds it beat players that just counted words or word pairs, showing meaning matters.
Abstract · Language Understanding for Text-based Games Using Deep Reinforcement Learning
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.
Karthik Narasimhan, Tejas Kulkarni, Regina Barzilay
arXiv:1506.08941 · cs.CL, cs.AI · submitted Jun 30, 2015 · updated Sep 11, 2015
abstract · pdf · html · 11 pages, Appearing at EMNLP, 2015
