Transfer in Deep Reinforcement Learning using Knowledge Graphs
This work addresses the challenge of grounding action in language for text adventure games, but it appears incremental as it builds on prior methods for transfer learning.
The paper tackles the problem of training reinforcement learning agents for text adventure games by using knowledge graphs for domain knowledge transfer, resulting in faster learning and higher-quality control policies across multiple games.
Text adventure games, in which players must make sense of the world through text descriptions and declare actions through text descriptions, provide a stepping stone toward grounding action in language. Prior work has demonstrated that using a knowledge graph as a state representation and question-answering to pre-train a deep Q-network facilitates faster control policy transfer. In this paper, we explore the use of knowledge graphs as a representation for domain knowledge transfer for training text-adventure playing reinforcement learning agents. Our methods are tested across multiple computer generated and human authored games, varying in domain and complexity, and demonstrate that our transfer learning methods let us learn a higher-quality control policy faster.