Behavior Cloned Transformers are Neurosymbolic Reasoners
This work addresses the challenge of improving reasoning abilities for game agents in language-based environments, though it appears incremental as it builds on existing behavior cloning and symbolic augmentation techniques.
The paper tackled the problem of enhancing interactive agents' multi-step reasoning in text games by injecting symbolic module actions into a behavior cloned transformer's action space, resulting in an average performance increase of 22% across four benchmarks.
In this work, we explore techniques for augmenting interactive agents with information from symbolic modules, much like humans use tools like calculators and GPS systems to assist with arithmetic and navigation. We test our agent's abilities in text games -- challenging benchmarks for evaluating the multi-step reasoning abilities of game agents in grounded, language-based environments. Our experimental study indicates that injecting the actions from these symbolic modules into the action space of a behavior cloned transformer agent increases performance on four text game benchmarks that test arithmetic, navigation, sorting, and common sense reasoning by an average of 22%, allowing an agent to reach the highest possible performance on unseen games. This action injection technique is easily extended to new agents, environments, and symbolic modules.