Reinforcement Learning: Gaussian noise #3515
Merged
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This commit adds GaussianNoise class which is just a wrapper around arma::randi that is built to interact with reinforcement learning agents that expect an object of type Noise. As of now, there is only one reinforcement learning agent that expect a Noise instance which is DDPG Signed-off-by: Tarek <[email protected]>
This commit adds 2 new tests for the GaussianNoise class: 1- GaussianNoiseTest which makes sure that the returned noise sample has the right mean and standard deviation. 2- PendulumWithGaussianDDPG which makes sure that the DDPG agent converges when supplied with a GaussianNoise object Signed-off-by: Tarek <[email protected]>
This commit separates the reinforcement learning policy gradient algorithms tests from other reinforcement learning agents' tests Signed-off-by: Tarek <[email protected]>
This allows for reusability in /tests/q_learning_test.cpp, /tests/policy_gradient_test.cpp and any future reinforcement learning testing file Signed-off-by: Tarek <[email protected]>
Signed-off-by: Tarek <[email protected]>
zoq
reviewed
Jul 28, 2023
Signed-off-by: Tarek <[email protected]>
zoq
approved these changes
Jul 31, 2023
Member
zoq
left a comment
zoq
left a comment
There was a problem hiding this comment.
Thanks for putting this together, no further comments from my side.
![mlpack-bot[bot]](https://avatars.githubusercontent.com/in/23934?s=60&v=4){kind=small}
Member
|
Thanks for another great contribution. |
Merged
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This pull request introduces the
GaussianNoiseclass and reorganizes the reinforcement learning test suite.Implementation Details:
GaussianNoiseClass:GaussianNoiseclass provides the necessary noise object required for the DDPG algorithm.GaussianNoiseclass allows for potential usage in other RL algorithms, where a noise object might be necessary for training and convergence.policy_gradient_test.cppfile includes the tests for DDPG, TD3, and SAC.How Has This Been Tested?:
GaussianNoiseclass for exploration.