Research map / Probabilistic and scientific modeling / Uncertainty and Gaussian processes
DEGRE: Dynamic Gating Ensembles for Trust-Aware Rejection in Medical Image Diagnostics
Hai Nguyen Hong, Duong Bach, Nam Phan, Cuong V. Nguyen, Cuong Do
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 |
|---|---|---|---|
| use rejection gating instead of re-weighting gating (ablation) | DEGRE · COVID-QU-Ex (M=3) | AURC: 0.0025 -> 0.0018 | Improved |
| use rejection gating instead of re-weighting gating (ablation) | DEGRE · Brain Tumor (M=3) | AURC: 0.0003 -> 0.0002 | Improved |
| use rejection gating instead of re-weighting gating (ablation) | DEGRE · COVID-QU-Ex (M=3) | rejection rate: 11.19% -> 8.10% | Improved |
| add learned gating over confidence and disagreement instead of fixed heuristic (ablation) | DEGRE · Head CT (M=3) | AURC: 0.0282 -> 0.0089 | Improved |
| add learned gating over confidence and disagreement instead of fixed heuristic (ablation) | DEGRE · SARS-COV-2 (M=3) | AURC: 0.0081 -> 0.0020 | Improved |
| add disagreement to confidence with a fixed heuristic (ablation) | Baseline 0 · Head CT (M=3) | AURC: 0.0334 -> 0.0282 | Improved |
| add disagreement to confidence with a fixed heuristic (ablation) | Baseline 0 · SARS-COV-2 (M=3) | AURC: 0.0058 -> 0.0081 | Hurt |
| use DEGRE with 3 members instead of baseline with 5 members | DEGRE · Intracranial Hemorrhage | AURC: 0.0221 -> 0.0031 | Improved |
| use DEGRE with 3 members instead of baseline with 5 members | DEGRE · Head CT | AURC: 0.0334 -> 0.0182 | Improved |
| use DEGRE with 3 members instead of baseline with 5 members | DEGRE · Brain Tumor | AURC: 0.0003 -> 0.0002 | Improved |
| replace Training Dynamics with DEGRE gating | DEGRE · 12 datasets average (M=5) | AURC: 0.0071 -> 0.0027 | Improved |
| replace Energy Score with DEGRE | DEGRE · Head CT (M=5) | AURC: 0.2174 -> 0.0089 | Improved |
Closest papers by meaning
- The Power of Ensembles for Active Learning in Image Classification (CVPR 2018 · similarity 0.84)
- Deep Ensembling with No Overhead for either Training or Testing: The All-Round Blessings of Dynamic Sparsity (ICLR 2022 · similarity 0.83)
- The Boundaries of Fair AI in Medical Image Prognosis: A Causal Perspective (NeurIPS 2025 · similarity 0.83)
- Rate-In: Information-Driven Adaptive Dropout Rates for Improved Inference-Time Uncertainty Estimation (CVPR 2025 · similarity 0.83)
- Entropy Minimization without Model Collapse: Mitigating Prediction Bias in Medical Imaging (NeurIPS 2026 · similarity 0.82)
- DDx-TRACE: A Benchmark for Medical Diagnostic Trajectories in VLMs (NeurIPS 2026 · similarity 0.81)
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.
