## [2.0.0-alpha.11](https://github.com/astroautomata/SymbolicRegression.jl/compare/v2.0.0-alpha.10...v2.0.0-alpha.11) (2026-05-17) ### Features * add backsolve mutation ([de86fef](https://github.com/astroautomata/SymbolicRegression.jl/commit/de86fef0dc6e99cee05d2e80969ab0d0e0016907)) ### Bug Fixes * harden sparse regression library ([46a1ad2](https://github.com/astroautomata/SymbolicRegression.jl/commit/46a1ad2972c11f00969139c1b687a26ff976907a))

## SymbolicRegression v0.2.0 This release has been identified as a backport. Automated changelogs for backports tend to be wildly incorrect. Therefore, the list of issues and pull requests is hidden. <!-- **Merged pull requests:** - Clarify the installation instructions (#2) (@sheevy) - Generic types (#3) (@MilesCranmer) **Closed issues:** - Cannot run the example from the documentation (#1) -->

## What's Changed * BREAKING: Full support for arbitrary degree operators by @MilesCranmer in https://github.com/MilesCranmer/SymbolicRegression.jl/pull/471 * Create `mutate_feature` for mutating a feature node exclusively by @MilesCranmer in https://github.com/MilesCranmer/SymbolicRegression.jl/pull/475 * Clean up output of errormonitor by @MilesCranmer in https://github.com/MilesCranmer/SymbolicRegression.jl/pull/479 * Support arbitrary degree operators for dimensional analysis by @MilesCranmer in https://github.com/MilesCranmer/SymbolicRegression.jl/pull/472 **Full Changelog**: https://github.com/MilesCranmer/SymbolicRegression.jl/compare/v2.0.0-alpha.3...v2.0.0-alpha.4

## What's Changed * full Mooncake integration by @MilesCranmer in https://github.com/MilesCranmer/SymbolicRegression.jl/pull/468 * CompatHelper: bump compat for ProgressMeter to 1, (keep existing compat) by @github-actions[bot] in https://github.com/MilesCranmer/SymbolicRegression.jl/pull/437 * CompatHelper: bump compat for MLJModelInterface to 1, (keep existing compat) by @github-actions[bot] in https://github.com/MilesCranmer/SymbolicRegression.jl/pull/474 **Full Changelog**: https://github.com/MilesCranmer/SymbolicRegression.jl/compare/v2.0.0-alpha.1...v2.0.0-alpha.2

## What's Changed * Minimal compatibility with DynamicExpressions.jl v2 by @MilesCranmer in https://github.com/MilesCranmer/SymbolicRegression.jl/pull/464 * Note that the current support does not add ternary operators to SymbolicRegression yet **Full Changelog**: https://github.com/MilesCranmer/SymbolicRegression.jl/compare/v1.12.0...v2.0.0-alpha.1

## SymbolicRegression v1.12.0 [Diff since v1.11.3](https://github.com/MilesCranmer/SymbolicRegression.jl/compare/v1.11.3...v1.12.0) **Merged pull requests:** - fix: issue where minimizer is not loaded to tree (#467) (@MilesCranmer) **Closed issues:** - (Non-)negative custom loss functions (#81)

## SymbolicRegression v1.11.3 [Diff since v1.11.2](https://github.com/MilesCranmer/SymbolicRegression.jl/compare/v1.11.2...v1.11.3) **Merged pull requests:** - CompatHelper: bump compat for Optim to 1, (keep existing compat) (#461) (@github-actions[bot]) - ci: switch to codecov (#462) (@MilesCranmer)

## SymbolicRegression v1.11.0 [Diff since v1.10.0](https://github.com/MilesCranmer/SymbolicRegression.jl/compare/v1.10.0...v1.11.0) **Merged pull requests:** - Unused `TemplateExpression` in template_expression_complex.jl (#441) (@Moelf) - feat: allow negative losses (#447) (@MilesCranmer) - fix: avoid printing heap size hint if already created (#448) (@MilesCranmer) - fix: per task cache can be immutable (#450) (@MilesCranmer) - feat: allow `@template_spec` to declare explicit num_features (#453) (@MilesCranmer) - Deprecate ParametricExpression (#454) (@MilesCranmer) - Move enzyme tests outside of main suite (#455) (@MilesCranmer)