Add Ablation Guided Shapley Value Sampling attribution method - #1897
Add Ablation Guided Shapley Value Sampling attribution method#1897AniketPatel148 wants to merge 1 commit into
Conversation
New perturbation based attribution method that lowers Shapley Value Sampling cost by only sampling the features that interact. Implemented as a subclass of ShapleyValueSampling that composes FeatureAblation for two cheap ablation screens (leave one out and inclusion). Additive features keep their exact leave one out attribution, while interacting features are sampled with the parent estimator over the active set, preserving the standard Captum output shape. Adds the method, its export in captum/attr/__init__.py, and unit tests mirroring tests/attr/test_shapley.py. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
|
Hi @AniketPatel148! Thank you for your pull request and welcome to our community. Action RequiredIn order to merge any pull request (code, docs, etc.), we require contributors to sign our Contributor License Agreement, and we don't seem to have one on file for you. ProcessIn order for us to review and merge your suggested changes, please sign at https://code.facebook.com/cla. If you are contributing on behalf of someone else (eg your employer), the individual CLA may not be sufficient and your employer may need to sign the corporate CLA. Once the CLA is signed, our tooling will perform checks and validations. Afterwards, the pull request will be tagged with If you have received this in error or have any questions, please contact us at cla@meta.com. Thanks! |
|
Thank you for signing our Contributor License Agreement. We can now accept your code for this (and any) Meta Open Source project. Thanks! |
New perturbation based attribution method that lowers Shapley Value Sampling cost by only sampling the features that interact. Implemented as a subclass of ShapleyValueSampling that composes FeatureAblation for two cheap ablation screens (leave one out and inclusion). Additive features keep their exact leave one out attribution, while interacting features are sampled with the parent estimator over the active set, preserving the standard Captum output shape.
Adds the method, its export in captum/attr/init.py, and unit tests mirroring tests/attr/test_shapley.py.