Research map / Label-efficient and robust learning / Recognition and long-tail
Conformal Prediction Sets for Ordinal Classification
PRASENJIT DEY, Srujana Merugu, Sivaramakrishnan R Kaveri
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 unimodal construction COPOC instead of vanilla CE | V-CE · HCI | |PS| size: 3.28 -> 2.66 | Improved |
| add unimodal construction COPOC instead of vanilla CE | V-CE · HCI | Acc@3: 88.9 -> 91.7 | Improved |
| enforce unimodality via loss or soft labels (SORD, AVDL, Uni-loss, POE) instead of by construction | baselines · image datasets | contiguity violations CV%: 24.4 -> 2.1 (V-CE vs AVDL), COPOC 0 | Hurt |
| use COPOC instead of SORD on equi-spaced Double-Exponential data | COPOC · Synthetic D1 | KL divergence: 0.03 -> 0.04 | No clear effect |
| use COPOC instead of V-CE on non-equi-spaced and complex data | COPOC · Synthetic D4 | KL divergence: 0.14 -> 0.08 | Improved |
| use COPOC instead of V-CE with smaller training sets | COPOC · Synthetic D4 varying N | KL divergence | Improved |
| use LAC instead of APS conformal inference | COPOC · HCI | |PS| size: 2.66 -> 2.61 | Improved |
| use LAC instead of APS conformal inference | COPOC · Retina-MNIST | SSCV: 0.09 -> 0.24 | Hurt |
| use COPOC instead of Ordinal-APS | Ordinal-APS · Adience | |PS| size: 2.67 -> 2.26 | Improved |
| use Ordinal-APS instead of APS on V-CE | V-CE · Adience | |PS| size: 4.82 -> 2.67 | Improved |
| use COPOC instead of V-CE for likelihood fit | COPOC · Aesthetic | NLL: 1.49 -> 0.71 | Improved |
| use COPOC instead of V-CE for likelihood fit | COPOC · Retina MNIST | NLL: 1.24 -> 1.23 | No clear effect |
Closest papers by meaning
- Conformal Prediction for Class-wise Coverage via Augmented Label Rank Calibration (NeurIPS 2024 · similarity 0.86)
- Conformal Structured Prediction (ICLR 2025 · similarity 0.85)
- Conformal Prediction for Deep Classifier via Label Ranking (ICML 2024 · similarity 0.85)
- Class conditional conformal prediction for multiple inputs by p-value aggregation (NeurIPS 2025 · similarity 0.84)
- Conformity Score Averaging for Classification (ICML 2025 · similarity 0.84)
- Ensuring Class-Conditional Coverage for Pathological Workflows (Student Abstract) (AAAI 2025 · similarity 0.84)
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.
