Methods
=======

RecourseBench integrates a broad set of algorithmic recourse methods behind a
single ``MethodObject`` interface. The links below point to the original method
papers identified from the RecourseBench paper and its per-method reproduction
notes.

.. list-table::
   :header-rows: 1
   :widths: 22 28 50

   * - Registry name
     - Method
     - Paper
   * - ``apas``
     - APAS
     - `Probabilistically Robust Counterfactual Explanations under Model Changes <https://doi.org/10.1016/j.artint.2025.104459>`__
   * - ``arg_ensembling``
     - Argumentative Ensembles
     - `Argumentative Ensembling for Robust Recourse under Model Multiplicity <https://arxiv.org/abs/2506.20260>`__
   * - ``cchvae``
     - C-CHVAE
     - `Learning Model-Agnostic Counterfactual Explanations for Tabular Data <https://arxiv.org/abs/1910.09398>`__
   * - ``cemsp``
     - CEMSP
     - `Flexible and Robust Counterfactual Explanations with Minimal Satisfiable Perturbations <https://arxiv.org/abs/2309.04676>`__
   * - ``cfrl``
     - CFRL
     - `Model-Agnostic and Scalable Counterfactual Explanations via Reinforcement Learning <https://arxiv.org/abs/2106.02597>`__
   * - ``cfvae``
     - CF-VAE
     - `Preserving Causal Constraints in Counterfactual Explanations for Machine Learning Classifiers <https://arxiv.org/abs/1912.03277>`__
   * - ``claproar``
     - CLAPROAR
     - `Endogenous Macrodynamics in Algorithmic Recourse <https://arxiv.org/abs/2308.08187>`__
   * - ``clue``
     - CLUE
     - `Getting a CLUE: A Method for Explaining Uncertainty Estimates <https://arxiv.org/abs/2006.06848>`__
   * - ``cogs``
     - COGS
     - `Robustness of Recourse Methods <https://arxiv.org/search/?query=virgolin+robustness+recourse&searchtype=all>`__
   * - ``cols``
     - COLS
     - `Low-Cost Algorithmic Recourse for Users With Uncertain Cost Functions <https://arxiv.org/abs/2111.01235>`__
   * - ``cruds``
     - CRUDS
     - `CRUDS: Counterfactual Recourse Using Disentangled Subspaces <https://arxiv.org/search/?query=Counterfactual+Recourse+Using+Disentangled+Subspaces&searchtype=all>`__
   * - ``cvas_proj``
     - CVAS-PROJ
     - `Coverage-Validity-Aware Algorithmic Recourse <https://arxiv.org/abs/2311.11349>`__
   * - ``dice``
     - DiCE
     - `Explaining Machine Learning Classifiers through Diverse Counterfactual Explanations <https://arxiv.org/abs/1905.07697>`__
   * - ``dice_repro``
     - DiCE Reproduction
     - `Explaining Machine Learning Classifiers through Diverse Counterfactual Explanations <https://arxiv.org/abs/1905.07697>`__
   * - ``diverse_dist``
     - Diverse Dist
     - `Promoting Counterfactual Robustness through Diversity <https://arxiv.org/abs/2312.06564>`__
   * - ``face``
     - FACE
     - `FACE: Feasible and Actionable Counterfactual Explanations <https://arxiv.org/abs/1909.09369>`__
   * - ``feature_tweak``
     - Feature Tweak
     - `Interpretable Predictions of Tree-based Ensembles via Actionable Feature Tweaking <https://arxiv.org/abs/1706.06691>`__
   * - ``gravitational``
     - Gravitational
     - `Endogenous Macrodynamics in Algorithmic Recourse <https://arxiv.org/abs/2308.08187>`__
   * - ``gs``
     - Growing Spheres
     - `Inverse Classification for Comparison-based Interpretability in Machine Learning <https://arxiv.org/abs/1712.08443>`__
   * - ``larr``
     - LARR
     - `Learning-Augmented Robust Algorithmic Recourse <https://arxiv.org/abs/2410.01580>`__
   * - ``mace``
     - MACE
     - `Model-Agnostic Counterfactual Explanations for Consequential Decisions <https://arxiv.org/abs/1905.11190>`__
   * - ``probe``
     - PROBE
     - `Probabilistically Robust Recourse: Navigating the Trade-offs between Costs and Robustness in Algorithmic Recourse <https://arxiv.org/abs/2203.06768>`__
   * - ``proplace``
     - PROPLACE
     - `Provably Robust and Plausible Counterfactual Explanations for Neural Networks via Robust Optimisation <https://arxiv.org/abs/2309.12545>`__
   * - ``rbr``
     - RBR
     - `Robust Bayesian Recourse <https://arxiv.org/abs/2206.10833>`__
   * - ``revise``
     - REVISE
     - `Towards Realistic Individual Recourse and Actionable Explanations in Black-Box Decision Making Systems <https://arxiv.org/abs/1907.09615>`__
   * - ``roar``
     - ROAR
     - `Towards Robust and Reliable Algorithmic Recourse <https://arxiv.org/abs/2102.13620>`__
   * - ``sns``
     - SNS
     - `Consistent Counterfactuals for Deep Models <https://arxiv.org/abs/2110.03109>`__
   * - ``toy``
     - Toy
     - `RecourseBench: A Modular Framework for Reproducible Algorithmic Recourse Evaluation <https://arxiv.org/abs/2606.16113>`__
   * - ``trex``
     - T-REX
     - `Robust Counterfactual Explanations for Neural Networks With Probabilistic Guarantees <https://arxiv.org/abs/2305.11997>`__
   * - ``wachter``
     - Wachter
     - `Counterfactual Explanations without Opening the Black Box: Automated Decisions and the GDPR <https://arxiv.org/abs/1711.00399>`__

Source
------

The mappings above follow the method names and citations used in the
RecourseBench paper:

* `RecourseBench: A Modular Framework for Reproducible Algorithmic Recourse Evaluation <https://arxiv.org/abs/2606.16113>`__