Towards Solving Text-based Games by Producing Adaptive Action Spaces
This addresses a key bottleneck in text-based game solving for AI agents, though it is incremental as it builds on existing deep reinforcement learning approaches.
The paper tackles the problem of generating valid text commands for text-based games, proposing a model that produces adaptive action spaces and achieves high F1 scores on unseen commands.
To solve a text-based game, an agent needs to formulate valid text commands for a given context and find the ones that lead to success. Recent attempts at solving text-based games with deep reinforcement learning have focused on the latter, i.e., learning to act optimally when valid actions are known in advance. In this work, we propose to tackle the first task and train a model that generates the set of all valid commands for a given context. We try three generative models on a dataset generated with Textworld. The best model can generate valid commands which were unseen at training and achieve high $F_1$ score on the test set.