Improve activeSelect pair weighting with power transform#101
Merged
kronosapiens merged 1 commit intomainfrom Mar 13, 2026
Merged
Improve activeSelect pair weighting with power transform#101kronosapiens merged 1 commit intomainfrom
kronosapiens merged 1 commit intomainfrom
Conversation
…overage Changed regularization from linear blend (r*w + (1-r)) to power transform (w^r), which preserves meaningful signal ratios at intermediate r values instead of collapsing everything toward uniform. Also softened coverage curve from 1/(1+n) to 1/√(1+n) to reduce the penalty for moderately-observed entries. These changes help activeSelect steer toward more balanced sampling, particularly important for avoiding hub bias where one entry dominates comparisons. Co-Authored-By: Claude Haiku 4.5 <[email protected]>
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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Summary
Improved the regularization mechanism in activeSelect to produce meaningful signal differentiation at moderate r values. Changed from linear blend to power transform and softened the coverage penalty curve.
Changes:
r*w + (1-r)→w^r(power transform). At r=0.5, linear blend collapsed ratios to ~1.001:1; power transform preserves ~100:1 ratios.1/(1+n)→1/√(1+n). Moderately-observed entries (5-10 observations) were penalized too heavily and now get more selection weight.These changes help activeSelect steer sampling toward the ladder/tournament graph structure needed for balanced rankings, rather than defaulting to hub-biased selection.
🤖 Generated with Claude Code