Online Reinforcement Learning by Bayesian Inference

CASIA OpenIR > 多模态人工智能系统全国重点实验室 > 复杂系统智能机理与平行控制团队

	Online Reinforcement Learning by Bayesian Inference
	Xia ZP(夏中谱); Dongbin Zhao
	2015-07
会议名称	International Joint Conference on Neural Networks
会议录名称	Proceedings of International Joint Conference on Neural Networks 2015
会议日期	2015年7月
会议地点	Ireland
摘要	Policy evaluation has long been one of the core issues of the online reinforcement learning, especially in the continuous state domain. In this paper, the issue is addressed by employing Gaussian processes to represent the action value function from the probability perspective. By modeling the return as a stochastic variable, the action value function can sequentially update according to observed variables such as state and reward by Bayesian inference during the policy evaluation. The update rule shows that it is a temporal difference learning method with the learning rate determined by the uncertainty of a collected sample. Incorporating the policy evaluation method with the E-greedy action selection method, we propose an online reinforcement learning algorithm referred as to Bayesian-SARSA. It is tested on some benchmark problems and the empirical results verifies its effectiveness.
关键词	Reinforcement Learning Bayesian Inference Gaussian Processes
收录类别	EI
语种	英语
文献类型	会议论文
条目标识符	http://ir.ia.ac.cn/handle/173211/11434
专题	多模态人工智能系统全国重点实验室_复杂系统智能机理与平行控制团队
通讯作者	Dongbin Zhao
推荐引用方式 GB/T 7714	Xia ZP,Dongbin Zhao. Online Reinforcement Learning by Bayesian Inference[C],2015.

条目包含的文件		下载所有文件
文件名称/大小	文献类型	版本类型	开放类型	使用许可
2015. IJCNN_XiaZhao.（751KB）	会议论文		开放获取	CC BY-NC-SA	浏览下载