Implement closed form strategy#86
Merged
Merged
Conversation
Codecov Report✅ All modified and coverable lines are covered by tests. Additional details and impacted files@@ Coverage Diff @@
## main #86 +/- ##
==========================================
+ Coverage 99.38% 99.41% +0.03%
==========================================
Files 12 14 +2
Lines 486 517 +31
==========================================
+ Hits 483 514 +31
Misses 3 3 ☔ View full report in Codecov by Sentry. 🚀 New features to boost your workflow:
|
Member
|
Tests are failing, but LGTM overall |
albertpod
approved these changes
Nov 24, 2025
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
2 participants
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Feature:
ClosedFormStrategyfor Exact Gradient ProjectionsOverview
This PR introduces a new projection strategy,
ClosedFormStrategy, which leveragesClosedFormExpectations.jlto compute exact, analytic gradients for KL divergence minimization. By replacing Monte Carlo sampling with symbolic-numeric integration, this strategy offers "Zero-Variance" gradients, improving both accuracy and convergence speed for supported distributions.Key Changes
1. New Strategy
ClosedFormStrategystruct insrc/strategies/closed_form.jl.compute_gradient!logic within a new package extension:ext/ClosedFormExpectationsExt.ExponentialFamilyProjection.jlmaintains a weak dependency onClosedFormExpectations.jl(only loaded when the user explicitly uses it).2. Analytic Gradient Computation
ClosedWilliamsProductfromClosedFormExpectations.jl.Design Choices
[extensions]to isolate heavy symbolic dependencies. The core package remains lightweight.