Playing Text-Adventure Games with Graph-Based Deep Reinforcement Learning
This work addresses the challenge of reinforcement learning in text-based games, which is important for AI research in natural language understanding and decision-making, but it is an incremental improvement over existing methods.
The authors tackled the problem of efficient exploration in text-adventure games with large combinatorial action spaces by developing a deep reinforcement learning architecture that uses a learned knowledge graph to prune actions and transfer learning from question-answering. They demonstrated that their method learns control policies faster than baseline alternatives in experiments using the TextWorld framework.
Text-based adventure games provide a platform on which to explore reinforcement learning in the context of a combinatorial action space, such as natural language. We present a deep reinforcement learning architecture that represents the game state as a knowledge graph which is learned during exploration. This graph is used to prune the action space, enabling more efficient exploration. The question of which action to take can be reduced to a question-answering task, a form of transfer learning that pre-trains certain parts of our architecture. In experiments using the TextWorld framework, we show that our proposed technique can learn a control policy faster than baseline alternatives. We have also open-sourced our code at https://github.com/rajammanabrolu/KG-DQN.