Research map / Label-efficient and robust learning / Recognition and long-tail
Identity-Free Deferral For Unseen Experts
Joshua Strong, Pramit Saha, Yasin Ibrahim, Cheng Ouyang, Alison Noble
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 |
|---|---|---|---|
| replace L2D-Pop (QC) with Identity-Free Deferral | L2D rejector for unseen experts · Liver tumours (variable specialists, OOD experts) | AURSAC: .68 -> .78 | Improved |
| replace L2D-Pop (QC) with Identity-Free Deferral | L2D rejector for unseen experts · Blood Cells (variable specialists, OOD experts) | AURSAC: .74 -> .80 | Improved |
| replace L2D-Pop (QC) with Identity-Free Deferral | L2D rejector for unseen experts · HAM10000 (variable specialists, OOD experts) | AURSAC: .73 -> .77 | Improved |
| replace L2D-Pop with Identity-Free Deferral | L2D rejector for unseen experts · HAM10000 (stable specialists, ID experts) | AURSAC: .86 -> .86 | No clear effect |
| replace L2D-Pop (QC) with Identity-Free Deferral | L2D rejector for unseen experts · Blood Cells (stable specialists, ID experts) | AURSAC: .88 -> .89 | No clear effect |
| replace L2D-Pop (QC) with Identity-Free Deferral | L2D rejector for unseen experts · ImageNet-16H (noise 125, OOD experts) | AURSAC: .61 -> .66 | Improved |
| replace L2D-Pop (QC) with Identity-Free Deferral | L2D rejector for unseen experts · ImageNet-16H (noise 95, ID experts) | AURSAC: .72 -> .76 | Improved |
| replace L2D-Pop (QC) with Identity-Free Deferral | L2D rejector for unseen experts · Liver tumours (variable specialists, OOD experts) | AURDAC: .41 -> .62 | Improved |
| use class-level Bayesian profile and context-only loss instead of query-time expert labels | L2D rejector for unseen experts · L2D-Pop | annotation count: E x N query annotations -> sum of N_C^E context labels | Improved |
Closest papers by meaning
- Exploiting Human-AI Dependence for Learning to Defer (ICML 2024 · similarity 0.87)
- A Unifying Post-Processing Framework for Multi-Objective Learn-to-Defer Problems (NeurIPS 2024 · similarity 0.87)
- Density-Ratio Losses for Post-Hoc Learning to Defer (NeurIPS 2026 · similarity 0.86)
- When to Act and When to Ask: Policy Learning With Deferral Under Hidden Confounding (NeurIPS 2024 · similarity 0.86)
- Behavior Cloning is Not All You Need: The Optimality of On-Policy Distillation for Noisy Expert Feedback (NeurIPS 2026 · similarity 0.85)
- Autonomy-of-Experts Models (ICML 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.
