Research map / Learning theory and algorithms / Prediction and decision losses
CARLA: A Python Library to Benchmark Algorithmic Recourse and Counterfactual Explanation Algorithms
Martin Pawelczyk, Sascha Bielawski, Johan Van den Heuvel, Tobias Richter, Gjergji Kasneci
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Use AR, AR-LIME, or DICE rather than other independence-based methods | Independence-based recourse methods · Adult, GMC, COMPAS | redundancy | Improved |
| Use CEM rather than other independence-based methods | Independence-based recourse methods · Adult, GMC, COMPAS | cost (l1) | Improved |
| Use GS rather than other independence-based methods | Independence-based recourse methods · Adult, GMC, COMPAS | time | Improved |
| Use dependence-based rather than independence-based recourse methods | Counterfactual recourse methods · Adult, GMC, COMPAS | yNN | Improved |
| Use FACE-EPS rather than other dependence-based methods | Dependence-based recourse methods · Adult, GMC, COMPAS | yNN | Improved |
| Use CLUE or REVISE rather than FACE for l1 recourse cost | Dependence-based recourse methods · Adult, GMC, COMPAS | cost (l1) | Improved |
| Use CLUE or REVISE rather than FACE | Dependence-based recourse methods · Adult, GMC, COMPAS | cost (l0) | Hurt |
| Use generative methods CEM-VAE, CLUE, or REVISE | Dependence-based recourse methods · Adult, GMC, COMPAS | time | Improved |
| Tune CLUE hyperparameters rather than use the defaults (ablation) | CLUE · Recourse cost at fixed success rate | distance | No clear effect |
| Use smaller GS sphere-growth steps (ablation) | GS · Counterfactual generation | time | Hurt |
| Use larger GS sphere-growth steps (ablation) | GS · Counterfactual generation | performance | Hurt |
| Set GS sphere-growth step size to 0.02 (ablation) | GS · Counterfactual generation | performance | Improved |
Closest papers by meaning
- Counterfactual Metarules for Local and Global Recourse (ICML 2024 · similarity 0.89)
- LeapFactual: Reliable Visual Counterfactual Explanation Using Conditional Flow Matching (NeurIPS 2025 · similarity 0.88)
- Towards Unifying Evaluation of Counterfactual Explanations: Leveraging Large Language Models for Human-Centric Assessments (AAAI 2025 · similarity 0.87)
- A New Approach to Backtracking Counterfactual Explanations: A Unified Causal Framework for Efficient Model Interpretability (ICML 2025 · similarity 0.87)
- Faithful Model Explanations through Energy-Constrained Conformal Counterfactuals (AAAI 2024 · similarity 0.87)
- Generating Likely Counterfactuals Using Sum-Product Networks (ICLR 2025 · similarity 0.87)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
