MPI für biologische Kybernetik / Biologische Kybernetik |
|Derivatives of Logarithmic Stationary Distributions for Policy Gradient Reinforcement Learning|
|Authors:||Morimura, T.; Uchibe, E.; Yoshimoto, J.; Peters, J.; Doya, K.|
|Date of Publication (YYYY-MM-DD):||2010-02|
|Title of Journal:||Neural Computation|
|Issue / Number:||2|
|Intended Educational Use:||No|
|Abstract / Description:||Most conventional policy gradient reinforcement learning (PGRL) algorithms neglect (or do not explicitly make use of) a term in the average reward gradient with respect to the policy parameter. That term involves the derivative of the stationary state distribution that corresponds to the sensitivity of its distribution to changes in the policy parameter. Although the bias introduced by this omission can be reduced by setting the forgetting rate &#947; for the value functions close to 1, these algorithms do not permit &#947; to be set exactly at &#947; = 1. In this article, we propose a method for estimating the log stationary state distribution derivative (LSD) as a useful form of the derivative of the stationary state distribution through backward Markov chain formulation and a temporal difference learning framework. A new policy gradient (PG) framework with an LSD is also proposed, in which the average reward gradient can be estimated by setting //!--|
MFG_und--//amp;#947; = 0, so it becomes unnecessary to learn the value functions. We also test the performance of the proposed algorithms using simple benchmark tasks and show that these can improve the performances of existing PG methods.
|External Publication Status:||published|
|Communicated by:||Holger Fischer|
|Affiliations:||MPI für biologische Kybernetik/Empirical Inference (Dept. Schölkopf)|
The scope and number of records on eDoc is subject
to the collection policies defined by each institute
- see "info" button in the collection browse view.