用不确定性触发语言模型，提升强化学习在未知场景下的表现。
When to ASK: Uncertainty-Gated Language Assistance for Reinforcement Learning
- 根据不确定性阈值决定是否调用语言模型获取建议
- 在迁移任务中实现0.95的奖励，显著优于基线
- 适合需要鲁棒性与实时性的强化学习应用
强化学习（RL）代理在分布外（OOD）场景下常因高不确定性而表现随机。尽管语言模型（LM）蕴含丰富世界知识，但其规模大导致计算开销高，难以实时使用，且在自主规划方面存在局限。我们提出自适应安全知识（ASK），将小型语言模型与训练好的RL策略结合，在不重新训练的前提下增强OOD泛化能力。ASK采用蒙特卡洛丢弃法评估不确定性，仅当不确定性超过阈值时才向语言模型请求动作建议。这种选择性调用在保持原有策略效率的同时，利用语言模型在不确定情境下的推理能力。在FrozenLake环境中，ASK在原域内无性能提升，但在迁移任务中表现出色，获得0.95的奖励。结果表明，有效的神经符号融合需精细协调而非简单拼接，强调模型规模与有效混合机制对成功OOD泛化的重要性。
原文摘要 · Abstract (English)
Reinforcement learning (RL) agents often struggle with out-of-distribution (OOD) scenarios, leading to high uncertainty and random behavior. While language models (LMs) contain valuable world knowledge, larger ones incur high computational costs, hindering real-time use, and exhibit limitations in autonomous planning. We introduce Adaptive Safety through Knowledge (ASK), which combines smaller LMs with trained RL policies to enhance OOD generalization without retraining. ASK employs Monte Carlo Dropout to assess uncertainty and queries the LM for action suggestions only when uncertainty exceeds a set threshold. This selective use preserves the efficiency of existing policies while leveraging the language model's reasoning in uncertain situations. In experiments on the FrozenLake environment, ASK shows no improvement in-domain, but demonstrates robust navigation in transfer tasks, achieving a reward of 0.95. Our findings indicate that effective neuro-symbolic integration requires careful orchestration rather than simple combination, highlighting the need for sufficient model scale and effective hybridization mechanisms for successful OOD generalization.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。