Input normalization using a reinforcement learning DQN agent
8 views (last 30 days)
Show older comments
Margarita Cabrera
on 26 Mar 2021
Commented: H. M.
on 3 Dec 2022
Hi all
I have built a DQN agent to solv a custom reinforcement problem.
Following mathworks examples I have no used any kind of normalization applied to the critic input.
In fact, as far as I could check, all examples of RL that use a DNN to create an actor or a critic especify 'Normalization', 'none' at the input layers of the Actor and Critic.
My question is, is it possible to use a normalization as for instance "zscore" at the input layers of a critic or of an actor when these are based on a DNN?
I'have tried to applied zscore normalization, but then, the agent does not work.
thanks
0 Comments
Accepted Answer
Emmanouil Tzorakoleftherakis
on 26 Mar 2021
Hello,
Normalization through the input layers is not supported for RL training. As a workaround, you can scale the observations rewards on the environment side.
2 Comments
H. M.
on 3 Dec 2022
@Emmanouil Tzorakoleftherakis
Could you explain more, the way you mentioned about normalization. I want to do it, but I can't figure it out.
Regards
More Answers (0)
See Also
Products
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!