Pre-trained Language Models as Prior Knowledge for Playing Text-based Games

Recently, text world games have been proposed to enable artificial agents to understand and reason about real-world scenarios. These text-based games are challenging for artificial agents, as it requires an understanding of and interaction using natural language in a partially observable environment. Past approaches have paid less attention to the language understanding capability of the proposed agents. In this paper, we improve the semantic understanding of the agent by proposing a simple RL with LM framework where we use transformer-based language models with Deep RL models. Overall, our proposed approach outperforms on 4 games out of the 14 textbased games, while performing comparable to the state-of-the-art models on the remaining games.