能完成多步目标任务的智能体必然具备环境预测模型。
General agents contain world models
- 从策略中提取智能体的环境预测模型
- 性能越强或目标越复杂，模型精度越高
- 为安全智能体设计与能力评估提供新思路
通用智能体是否必须包含世界模型？我们给出形式化回答：任何能够泛化到多步目标导向任务的智能体，都必须学习环境的预测模型。该模型可从智能体策略中提取，且其准确性随智能体性能提升或目标复杂度增加而提高。这一发现对构建安全通用智能体、在复杂环境中约束智能体能力，以及从智能体中提取世界模型的新算法具有重要意义。
原文摘要 · Abstract (English)
Are world models a necessary ingredient for flexible, goal-directed behaviour, or is model-free learning sufficient? We provide a formal answer to this question, showing that any agent capable of generalizing to multi-step goal-directed tasks must have learned a predictive model of its environment. We show that this model can be extracted from the agent's policy, and that increasing the agents performance or the complexity of the goals it can achieve requires learning increasingly accurate world models. This has a number of consequences: from developing safe and general agents, to bounding agent capabilities in complex environments, and providing new algorithms for eliciting world models from agents.
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。