Research map / Label-efficient and robust learning / Recognition and long-tail
Resolving Predictive Multiplicity for the Rashomon Set
Parian Haghighat, Hadis Anahideh, Cynthia Rudin
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 |
|---|---|---|---|
| add pairwise reconciliation to the Rashomon set predictions (ablation) | Rashomon set ensemble · Adult | Disagreement rate: +0.00% -> -99.67% | Improved |
| add pairwise reconciliation (ablation) | Rashomon set ensemble · COMPAS | Disagreement rate: +0.00% -> -99.77% | Improved |
| add local patching alone (ablation) | Rashomon set ensemble · Adult, COMPAS, Folk Mobility, Folk Travel | Accuracy: -0.49%, -1.82%, -0.80%, -4.19% | Hurt |
| add local patching alone (ablation) | Rashomon set ensemble · Adult | Disagreement rate: +0.00% -> -39.95% | Improved |
| add local patching to pairwise reconciliation (ablation) | PR · Adult | LCAE: -5.08% -> -13.96% | Improved |
| add local patching to pairwise reconciliation (ablation) | PR · Folk Travel | Accuracy: +2.58% -> -2.34% | Hurt |
| add outlier correction to pairwise reconciliation (ablation) | PR · Adult | Disagreement rate: -99.67% -> -67.20% | Hurt |
| add outlier correction alone (ablation) | Rashomon set ensemble · four datasets | Ensemble-wide disagreement | No clear effect |
| add outlier correction to local patching (ablation) | LP · Folk Mobility | Variance: +5.26% -> +7.89% | Hurt |
| add all three methods (OC+PR+LP) vs PR+LP (ablation) | PR + LP · Adult | Accuracy: -0.35% -> +0.25% | Improved |
| add pairwise reconciliation to local patching (ablation) | LP · Folk Travel | Accuracy: -4.19% -> -2.34% | Improved |
| use pairwise reconciliation alone for local agreement (ablation) | PR · Adult | LCAE: -5.08% | No clear effect |
Closest papers by meaning
- Dropout-Based Rashomon Set Exploration for Efficient Predictive Multiplicity Estimation (ICLR 2024 · similarity 0.91)
- RashomonGB: Analyzing the Rashomon Effect and Mitigating Predictive Multiplicity in Gradient Boosting (NeurIPS 2024 · similarity 0.89)
- Position: Amazing Things Come From Having Many Good Models (ICML 2024 · similarity 0.87)
- Quantifying Prediction Consistency Under Fine-tuning Multiplicity in Tabular LLMs (ICML 2025 · similarity 0.86)
- ElliCE: Efficient and Provably Robust Algorithmic Recourse via the Rashomon Sets (NeurIPS 2025 · similarity 0.86)
- Reconciling Model Multiplicity for Downstream Decision Making (ICLR 2025 · similarity 0.85)
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.
