Use one formula to calculate cosine similarity #2357
Merged
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
VijayanB
force-pushed
the
update-score
branch
3 times, most recently
from
f14147b to
6e7c88e
Compare
VijayanB
changed the title
VijayanB
force-pushed
the
update-score
branch
from
6e7c88e to
7d19cac
Compare
VijayanB
requested review from
heemin32,
navneet1v,
vamshin,
jmazanec15,
naveentatikonda,
junqiu-lei,
martin-gaievski,
ryanbogan,
luyuncheng and
shatejas
as code owners
VijayanB
force-pushed
the
update-score
branch
2 times, most recently
from
37d132a to
cc49de3
Compare
|
Looks like flaky test failure since the same test is successful in Linux and Mac. Create GH issue |
jmazanec15
reviewed
Dec 30, 2024
There was a problem hiding this comment.
thanks @VijayanB - can you summarize impact of change in description. For instance, this only impacts nmslib, correct?
VijayanB
force-pushed
the
update-score
branch
2 times, most recently
from
a4eb621 to
424765e
Compare
VijayanB
force-pushed
the
update-score
branch
2 times, most recently
from
d4cbc14 to
6597eb1
Compare
navneet1v
reviewed
Dec 30, 2024
src/main/java/org/opensearch/knn/plugin/script/KNNScoringSpace.java
Outdated
Show resolved
Hide resolved
navneet1v
reviewed
Dec 30, 2024
navneet1v
previously approved these changes
Jan 3, 2025
@navneet1v Yes. Will add it as next PR |
Currently we have different score calculation for cosine similarity, for ex: script score, approximate search, exact search has diffent formula to convert distance to cosine similarity that is aligned with OpenSearch score. To keep it consistent, we will be using one defintion which is used by Lucene as standard definition for cosine similarity for all search types. Signed-off-by: Vijayan Balasubramanian <[email protected]>
Signed-off-by: Vijayan Balasubramanian <[email protected]>
Signed-off-by: Vijayan Balasubramanian <[email protected]>
|
Rebased to resolve CHANGE.MD file conflict |
navneet1v
approved these changes
Jan 3, 2025
jmazanec15
approved these changes
Jan 6, 2025
|
The backport to To backport manually, run these commands in your terminal: # Fetch latest updates from GitHub
git fetch
# Create a new working tree
git worktree add .worktrees/backport-2.x 2.x
# Navigate to the new working tree
cd .worktrees/backport-2.x
# Create a new branch
git switch --create backport/backport-2357-to-2.x
# Cherry-pick the merged commit of this pull request and resolve the conflicts
git cherry-pick -x --mainline 1 84cfa8e00ac04197cc98a5f32379637730cdb979
# Push it to GitHub
git push --set-upstream origin backport/backport-2357-to-2.x
# Go back to the original working tree
cd ../..
# Delete the working tree
git worktree remove .worktrees/backport-2.xThen, create a pull request where the |
Gankris96
pushed a commit
to Gankris96/k-NN
that referenced
this pull request
Jan 8, 2025
* Have one score definition for cosinesimilarity Currently we have different score calculation for cosine similarity, for ex: script score, approximate search, exact search has diffent formula to convert distance to cosine similarity that is aligned with OpenSearch score. To keep it consistent, we will be using one defintion which is used by Lucene as standard definition for cosine similarity for all search types. Signed-off-by: Vijayan Balasubramanian <[email protected]> * update test Signed-off-by: Vijayan Balasubramanian <[email protected]> * add version check Signed-off-by: Vijayan Balasubramanian <[email protected]> --------- Signed-off-by: Vijayan Balasubramanian <[email protected]>
VijayanB
added a commit
to VijayanB/k-NN-2
that referenced
this pull request
Jan 9, 2025
* Have one score definition for cosinesimilarity Currently we have different score calculation for cosine similarity, for ex: script score, approximate search, exact search has diffent formula to convert distance to cosine similarity that is aligned with OpenSearch score. To keep it consistent, we will be using one defintion which is used by Lucene as standard definition for cosine similarity for all search types. Signed-off-by: Vijayan Balasubramanian <[email protected]> * update test Signed-off-by: Vijayan Balasubramanian <[email protected]> * add version check Signed-off-by: Vijayan Balasubramanian <[email protected]> --------- Signed-off-by: Vijayan Balasubramanian <[email protected]>
VijayanB
added a commit
to VijayanB/k-NN-2
that referenced
this pull request
Jan 9, 2025
* Have one score definition for cosinesimilarity Currently we have different score calculation for cosine similarity, for ex: script score, approximate search, exact search has diffent formula to convert distance to cosine similarity that is aligned with OpenSearch score. To keep it consistent, we will be using one defintion which is used by Lucene as standard definition for cosine similarity for all search types. Signed-off-by: Vijayan Balasubramanian <[email protected]> * update test Signed-off-by: Vijayan Balasubramanian <[email protected]> * add version check Signed-off-by: Vijayan Balasubramanian <[email protected]> --------- Signed-off-by: Vijayan Balasubramanian <[email protected]>
opensearch-trigger-bot bot
pushed a commit
that referenced
this pull request
Jan 9, 2025
* Have one score definition for cosinesimilarity Currently we have different score calculation for cosine similarity, for ex: script score, approximate search, exact search has diffent formula to convert distance to cosine similarity that is aligned with OpenSearch score. To keep it consistent, we will be using one defintion which is used by Lucene as standard definition for cosine similarity for all search types. Signed-off-by: Vijayan Balasubramanian <[email protected]> * update test Signed-off-by: Vijayan Balasubramanian <[email protected]> * add version check Signed-off-by: Vijayan Balasubramanian <[email protected]> --------- Signed-off-by: Vijayan Balasubramanian <[email protected]> (cherry picked from commit 84cfa8e)
owenhalpert
pushed a commit
to owenhalpert/k-NN
that referenced
this pull request
Jan 9, 2025
* Have one score definition for cosinesimilarity Currently we have different score calculation for cosine similarity, for ex: script score, approximate search, exact search has diffent formula to convert distance to cosine similarity that is aligned with OpenSearch score. To keep it consistent, we will be using one defintion which is used by Lucene as standard definition for cosine similarity for all search types. Signed-off-by: Vijayan Balasubramanian <[email protected]> * update test Signed-off-by: Vijayan Balasubramanian <[email protected]> * add version check Signed-off-by: Vijayan Balasubramanian <[email protected]> --------- Signed-off-by: Vijayan Balasubramanian <[email protected]>
VijayanB
added a commit
that referenced
this pull request
Jan 12, 2025
* Have one score definition for cosinesimilarity Currently we have different score calculation for cosine similarity, for ex: script score, approximate search, exact search has diffent formula to convert distance to cosine similarity that is aligned with OpenSearch score. To keep it consistent, we will be using one defintion which is used by Lucene as standard definition for cosine similarity for all search types. Signed-off-by: Vijayan Balasubramanian <[email protected]> (cherry picked from commit 84cfa8e)
VijayanB
added a commit
that referenced
this pull request
Jan 15, 2025
* Have one score definition for cosinesimilarity Currently we have different score calculation for cosine similarity, for ex: script score, approximate search, exact search has diffent formula to convert distance to cosine similarity that is aligned with OpenSearch score. To keep it consistent, we will be using one defintion which is used by Lucene as standard definition for cosine similarity for all search types. Signed-off-by: Vijayan Balasubramanian <[email protected]> (cherry picked from commit 84cfa8e)
VijayanB
added a commit
that referenced
this pull request
Jan 15, 2025
* Have one score definition for cosinesimilarity Currently we have different score calculation for cosine similarity, for ex: script score, approximate search, exact search has diffent formula to convert distance to cosine similarity that is aligned with OpenSearch score. To keep it consistent, we will be using one defintion which is used by Lucene as standard definition for cosine similarity for all search types. Signed-off-by: Vijayan Balasubramanian <[email protected]> (cherry picked from commit 84cfa8e)
VijayanB
added a commit
that referenced
this pull request
Jan 15, 2025
* Have one score definition for cosinesimilarity Currently we have different score calculation for cosine similarity, for ex: script score, approximate search, exact search has diffent formula to convert distance to cosine similarity that is aligned with OpenSearch score. To keep it consistent, we will be using one defintion which is used by Lucene as standard definition for cosine similarity for all search types. Signed-off-by: Vijayan Balasubramanian <[email protected]> (cherry picked from commit 84cfa8e) Co-authored-by: Vijayan Balasubramanian <[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.
Description
Currently we have different score calculation for cosine similarity,
for ex: script score, approximate search, exact search has diffent formula to convert distance to cosine similarity in order to align with OpenSearch score definition. To keep it consistent, we will be using one definition which is used
by Lucene as standard definition for cosine similarity for all search types.
What changed?
Going forward, nmslib engine with cosine similarity as space type will use new formula to define similarity ( same as how exact search and lucene calculates cosine similarity between query and input vector).
During script scoring ( painless and knn script score ), cosinesimil method will also new formula to calculate similarity between query and input vector.
Other engines, are not affected with this change.
Related Issues
#2319
Check List
--signoff.By submitting this pull request, I confirm that my contribution is made under the terms of the Apache 2.0 license.
For more information on following Developer Certificate of Origin and signing off your commits, please check here.