[DIA] Bayesian optimisation #1155
Closed
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
* Adding script bayesianOptimisation.dml which will contain the concrete implementation. * Adding BuiltingBazyesianOptimisationTest.java used for tests. * Adding bayesianOptimisationLM, bootstraping tests for the implementation. * Adding the script bayesianOptimisation.dml to the Builtin function, such that the parser understands the call to bayesianOptimisation().
files into the staging directory. * Moving new implementation to the correct directory. * Removing the uneeded bayesianOptimisationLM.dml file, which would call the bayesianOptimisation.dml script.
* Working in implementing a gaussian_kernel. * Calling aquisition/kernel functions.
* Creating intermediate test file bayesianOptimisationSinTest.dml * Moving acquisition/kernel function to the intermediate test file.
| kernel_gaus = function(Matrix[Double] X1, Matrix[Double] X2, Double variance) | ||
| return (Matrix[Double] result) | ||
| { | ||
| #TODO: What to do about multiple dimensional matrix? |
Contributor
There was a problem hiding this comment.
Hi,
Just quick feedback, in DML we use two-space indentation.
Contributor
Author
There was a problem hiding this comment.
Hi and thanks for the Feedback, I applied it in the last commit.
Contributor
|
Currently the tests are failing because of missing licenses, these can be simply added at the top of the files. |
* The algorithm now initialized with random samples. * With each iteration the algorithm will try to pick better hyperparameter sets. * There is still a bug which prevent accurate calculation of the means and variances used in the acquisition function. * Improved documentation. * Added further print messages for verbose mode. * Using the upper bound confidence instead of the probability for improvement as acquisition function.
Contributor
Author
|
@Baunsgaard Thanks, I've added them now. |
DodoTheDeveloper
changed the title
means/variances. Rewrote the updating of the model.
* The comparison test with a lm function using the default parameters, accidently used the TEST and not the TRAINING dataset. Making the lm using the searched parameters losing always. * Fixing license headers.
DodoTheDeveloper
changed the title
| @@ -1,3 +1,22 @@ | |||
| /* | |||
Contributor
There was a problem hiding this comment.
Please reformat the test file. Add some test cases for different parameters of the optimizer such as number of iterations, minimize, and ExecType (CP or SPARK).
* Kernel function now are element wise instead of vector vise. The previous implementation would hand some complexity for the user to handle, when choosing/designing the kernel function. * Fixing a offset error of the idxScoreEntry. * Acquisition function would receive the variance instead of the standard deviation. * Remove some debugging prints. * Logs printed in debug mode now show up to 10 decimals. * Added a test for maximisation.
Contributor
|
LGTM, |
Contributor
Author
Thanks, I didn't see your reply in time. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 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 implement bayesian optimisation for hyperparameters.