Research map / Label-efficient and robust learning / Recognition and long-tail

Ensemble Selective Classification

NeurIPS 2026 · Label-efficient and robust learning › Recognition and long-tail · approach: Group-robust classification

Read the paperCheck your idea against 108,967 papers

AI-extracted summary. Written by AI from the paper's abstract only. Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Proposes ensembling multiple post-hoc confidence scores for selective classification so a single robust score is not needed across models, datasets, and shifts.

Problem

Post-hoc selective classification with a fixed, possibly black-box classifier: no single confidence score is consistently best across models, datasets, and deployment shifts.

Approach

Combine multiple existing post-hoc confidence scores into an ensemble score for selective classification, rather than selecting one score per setting (based on the abstract; the abstract is truncated, so details are unknown).

Closest papers by meaning

• What Does It Take to Build a Performant Selective Classifier? (NeurIPS 2025 · similarity 0.89)
• Overcoming Common Flaws in the Evaluation of Selective Classification Systems (NeurIPS 2024 · similarity 0.86)
• Confidence-aware Contrastive Learning for Selective Classification (ICML 2024 · similarity 0.86)
• Structure-informed Risk Minimization for Robust Ensemble Learning (ICML 2025 · similarity 0.84)
• Hierarchical Selective Classification (NeurIPS 2024 · similarity 0.84)
• Perturb and Correct: Post-Hoc Ensembles using Affine Redundancy (NeurIPS 2026 · similarity 0.84)