Safe Reinforcement Learning with Free-form Natural Language Constraints and Pre-Trained Language Models

arXiv:2401.07553v2 Announce Type: replace
Abstract: Safe reinforcement learning (RL) agents accomplish given tasks while adhering to specific constraints. Employing constraints expressed via easily-understandable human language offers considerable potential for real-world applications due to its accessibility and non-reliance on domain expertise. Previous safe RL methods with natural language constraints typically adopt a recurrent neural network, which leads to limited capabilities when dealing with various forms of human language input. Furthermore, these methods often require a ground-truth cost function, necessitating domain expertise …

abstract accessibility agents applications arxiv constraints cs.cl cs.lg domain expertise form free human language language models natural natural language reinforcement reinforcement learning reliance safe tasks type via world