World Model Learning from Demonstrations with Active Inference: Application to Driving Behavior
World Model Learning from Demonstrations with Active Inference: Application to Driving Behavior
复制标题
DOI：
10.1007/978-3-031-28719-0_9
复制
发表时间：
2022
期刊：
影响因子：
--
通讯作者：
Ran Wei;Alfredo Garcia;Anthony D. McDonald;G. Markkula;J. Engström;Isaac Supeene;Matthew O'Kelly
中科院分区：
文献类型：
--
作者：
Ran Wei;Alfredo Garcia;Anthony D. McDonald;G. Markkula;J. Engström;Isaac Supeene;Matthew O'Kelly
主动推理为感知和行动提出了一个统一的原则，即共同最小化主体内部世界模型的自由能。在主动推理文献中，世界模型通常是预先指定的，或者通过与环境的交互来学习。本文探讨了学习世界模型的主动推理代理记录演示的可能性，与人类驾驶行为建模的应用。实验结果表明，该方法可以生成类人驾驶行为的模型，但对输入特征敏感。
Active inference proposes a unifying principle for perception and action as jointly minimizing the free energy of an agent’s internal world model. In the active inference literature, world models are typically pre-specified or learned through interacting with an environment. This paper explores the possibility of learning world models of active inference agents from recorded demonstrations, with an application to human driving behavior modeling. The results show that the presented method can create models that generate human-like driving behavior but the approach is sensitive to input features.