Research map / Learning theory and algorithms / Prediction and decision losses
Softmax is not Enough (for Adaptive Conformal Classification)
Navid Akhavan Attar, Hesam Asadollahzadeh, Ling Luo, Uwe Aickelin
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 energy-based reweighting of the score | APS · CIFAR-100 (ResNet-56, alpha=0.05) | average set size: 6.91 -> 6.49 | Improved |
| add energy-based reweighting | RAPS · CIFAR-100 (ResNet-56, alpha=0.05) | average set size: 8.17 -> 6.18 | Improved |
| add energy-based reweighting | SAPS · CIFAR-100 (ResNet-56, alpha=0.05) | average set size: 7.47 -> 5.94 | Improved |
| add energy-based reweighting | APS · ImageNet (ResNet-50, alpha=0.1) | average set size: 1.60 -> 1.66 | Hurt |
| add energy-based reweighting | APS · ImageNet (ResNet-50, alpha=0.01) | average set size: 39.08 -> 32.93 | Improved |
| add energy-based reweighting | SAPS · ImageNet (ResNet-50, alpha=0.1) | average set size: 1.67 -> 1.66 | No clear effect |
| add energy-based reweighting | APS · Places365 (ResNet-50, alpha=0.05) | average set size: 14.28 -> 12.98 | Improved |
| add energy-based reweighting | APS · CIFAR-100-LT lambda=0.005 (alpha=0.05) | average set size: 17.22 -> 13.30 | Improved |
| add energy-based reweighting | SAPS · CIFAR-100-LT lambda=0.005 (alpha=0.01) | average set size: 50.30 -> 35.18 | Improved |
| add energy-based reweighting | APS · CIFAR-100 to Places365 OOD (alpha=0.05) | OOD average set size: 14.91 -> 93.40 | Improved |
| add energy-based reweighting | RAPS · CIFAR-100 to Places365 OOD (alpha=0.1) | OOD average set size: 3.70 -> 5.53 | Improved |
| add energy-based reweighting | RAPS · CIFAR-100 to Places365 OOD (alpha=0.05) | OOD average set size: 8.95 -> 9.05 | No clear effect |
Closest papers by meaning
- Direct Prediction Set Minimization via Bilevel Conformal Classifier Training (ICML 2025 · similarity 0.90)
- Hierarchical Conformal Classification (NeurIPS 2026 · similarity 0.89)
- Conformal Prediction for Deep Classifier via Label Ranking (ICML 2024 · similarity 0.89)
- An Information Theoretic Perspective on Conformal Prediction (NeurIPS 2024 · similarity 0.89)
- On Temperature Scaling and Conformal Prediction of Deep Classifiers (ICML 2025 · similarity 0.89)
- Generative Conformal Prediction with Optimized Coverage Allocation (NeurIPS 2026 · similarity 0.89)
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.
