-
Notifications
You must be signed in to change notification settings - Fork 1
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Fanfan #7
Merged
Merged
Fanfan #7
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
Sync Forked Repo
Sync from master
This can be dangerous because it is a "hidden" behavior, which can be unexpected by the user. It could default to False, but for the moment I put no default value.
This is not commutative!
The result is not in the positive orthant.
max_angular_dilatation_factor (add) Embeddings.dilated_aux (add) Embeddings.dilated (fix documentation) Embeddings.dilated_new (add)
* Correct a bug in __new__ / __array_finalize__ : when norm=False, n_voters and n_dim were switched. * get_center and dilated_aux now work also for non-normalized embeddings. * Correct bugs in the degerate cases (like when all embeddings are aligned). * Rewrite recentered using a matrix operation. * Add `recentered_and_dilated`. * No need to override the default implementation of `copy`. * max_angular_dilatation_factor handles better the numerical glitches.
* Add `Embeddings.mixed_with` * Rename `EmbeddingsGeneratorRandom` to `EmbeddingsGeneratorUniform`. * Add `EmbeddingsGeneratorFullyPolarized` * Simplify `EmbeddingsGeneratorPolarized`, based on the above.
Useless since it is now a synonym of self.truth_generator.
* Before, it relied on the ratings themselves, and the covariance was computed inside the class. Advantage of the new version: if you know the true covariance matrix of the noises, you can give it to the rule. * Add `EmbeddingsFromRatingsCovariance`.
Singular values are not sorted. Unfortunately, this behavior is implicitly assumed in `RuleSVD`, because of the instruction `s[:matrix_rank]`.
* Singular values are now sorted by decreasing order. * Indirectly, this fixes a bug in `RuleSVD` for the case where `use_rank` is True.
Minor editing
The new code behaves as if this parameter was always True.
Demonstrated in Aggregator: the embeddings are a concatenation of a correlation matrix and a history of scores, which is certainly not intended!
It was the sqrt of the singular value (probably by mistake).
* embeddings.copy() or Embeddings(embeddings) also copies the attributes of `embeddings`. This is useful typically for embeddings obtained with `EmbeddingsFromRatingsCorrelation`, which have the attribute `n_sing_val_`. * `RuleFast` takes as embeddings the correlation matrix (and not the ratings themselves). Hence the default `embeddings_from_ratings` is `EmbeddingsFromRatingsCorrelation()`.
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.
No description provided.