Towards Solving Text-based Games by Producing Adaptive Action Spaces

TL;DR

This paper introduces a generative model that predicts valid commands in text-based games, enabling agents to adaptively produce action spaces and improve game-solving capabilities.

Contribution

It presents a novel approach to generate valid commands dynamically, addressing a key challenge in text-based game AI that was previously overlooked.

Findings

The best model generates unseen valid commands with high accuracy.

Achieves high F1 score on test set.

Demonstrates potential for adaptive action space generation.

Abstract

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.