Rank nearest neighbors in embedding space #211
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
ziw-liu
added
enhancement
Collaborator
Author
|
Using standardized features (949df3c) changes the angular distance, but doesn't really change the shape of the rankings curve:
|
edyoshikun
approved these changes
Nov 22, 2024
Member
edyoshikun
left a comment
edyoshikun
left a comment
There was a problem hiding this comment.
Works like a charm! Also super helpful and thanks for this elegant implementation.
edyoshikun
pushed a commit
that referenced
this pull request
Dec 18, 2024
* methods to rank nearest neighbors in embeddings * example script to plot state change of a single track * test using scaled features
edyoshikun
pushed a commit
that referenced
this pull request
Dec 19, 2024
* methods to rank nearest neighbors in embeddings * example script to plot state change of a single track * test using scaled features
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Outdated suggestions cannot be applied.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.
An ideal cell state representation has the following properties:
We observed these properties for time-regularized models visually in UMAPs. However, Euclidean distances in UMAP space carry no global meaning, so the observations cannot be directly quantified by comparing UMAP values. This PR provides methods to describe the embedding similarities close to how UMAP is computed: each sample is ranked as the$k$ -th nearest neighbor for each sample, and the displacement in this neighborhood ($k$ at $t_{i+1}$ for each $t_i$ ) can then be used to describe the fluctuation of embeddings in a way that preserves latent space topology.
Example from ALFI dataset: