FIX Make dict learning loss match the paper #19210
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
|
@jeremiedbb this PR can now be rebased on top of the main branch now that #19198 has been merged. |
|
How does this impact qualitatively the existing examples? Apparently no tests were impacted by this change. I am not sure this is a good sign ;) |
ogrisel
approved these changes
Jun 3, 2021
…-learn into fix-dict-learning-loss
|
Here's the result of the plot_faces_decomposition.py example. there are some variations but nothing to worry I guess. |
|
Thanks @jeremiedbb |
samronsin
pushed a commit
to samronsin/scikit-learn
that referenced
this pull request
Nov 30, 2021
Co-authored-by: Olivier Grisel <[email protected]>
glemaitre
pushed a commit
to glemaitre/scikit-learn
that referenced
this pull request
Dec 24, 2021
Co-authored-by: Olivier Grisel <[email protected]>
glemaitre
pushed a commit
that referenced
this pull request
Dec 25, 2021
Co-authored-by: Olivier Grisel <[email protected]>
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Outdated suggestions cannot be applied.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.
Built on top of #19198 for convenience.
In the paper
the constraint is that the norm of each atom is less or equal to 1. The current doc and implementation make it equal to 1. I'm not sure what it implies on the optimization problem but it might be considered a bug.
SparsePCA relies on dict learning and has the same issue. In the paper
the constraint on the components is also less or equal to 1.
This change was also proposed in #9036