Bayesian RL and PGMRL

Posted huangshiyu13

tags:

篇首语:本文由小常识网(cha138.com)小编为大家整理,主要介绍了Bayesian RL and PGMRL相关的知识,希望对你有一定的参考价值。

简介:

PGMRL: PGMRL就是把RL问题建模成一个概率图模型,如下图所示:

 技术图片

然后通过variational inference的方法进行学习:

技术图片

PGMRL给RL问题的表示给了一个范例,对解决很多RL新问题提供了一种思路和工具。

 

Bayesian RL:

 

思考:为什么PGMRL推导过程中没有Beyesian RL的exploration-exploitation trade-off的问题。

 

thinking: what things does the Beyesian RL not consider?

以上是关于Bayesian RL and PGMRL的主要内容,如果未能解决你的问题,请参考以下文章

UPenn - Robotics 5:Robotics: Estimation and Learning - week 2:Bayesian Estimation - Target Tracking(

论文导读Learning Bayesian Networks: The Combination of Knowledge and Statistical Data

论文导读Learning Bayesian Networks: The Combination of Knowledge and Statistical Data

Deep RL Bootcamp Lecture 7: SVG, DDPG, and Stochastic Computation Graphs

Deep RL Bootcamp Lecture 2: Sampling-based Approximations and Function Fitting

lec-1-Deep Reinforcement Learning, Decision Making, and Control