Fix spaopt #39
Merged
Fix spaopt #39
Conversation
This file contains 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
Spa is not optimized for this basis any longer and needs some manual adjustments.
Seanny123
approved these changes
Jun 10, 2017
The previous choice of intercept and eval point distribution was not sufficiently validated. The choice in this commit performs better and over a wider set of parameters.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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Outdated suggestions cannot be applied.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.
Motivation and context:
Benchmarking is hard and I screwed up. The circular convolution wasn't nearly as accurate as it was supposed to be because I only validated the choice of distributions on 50 neurons per dimensions. The choice proposed in this PR is validated on 50 and 200 neurons per dimensions and also for convolution of random + unitary vector, random + random vector, unitary + unitary vector.
I updated the corresponding section in the intercept distribution notebook.
@rworr, this might be relevant to you.
Interactions with other PRs:
none
How has this been tested?
Validation done elsewhere, the linked notebook gives the main conclusions.
Unit tests still pass, though one test needed adjustment because it violated the assumption of high-dimensional random vectors.
How long should this take to review?
Types of changes:
Checklist: