Research map / Label-efficient and robust learning / Domain adaptation and generalization
Debiased Curriculum Adaptation for Safe Transfer Learning in Chest X-ray Classification
Mingyang Liu, Xinyang Chen, Yang Shu, Xiucheng Li, Weili Guan, Liqiang Nie
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 curriculum adaptation (Lst) (ablation) | DCA · Open-i to NIH-CXR avg of 3 diseases | AUROC: 62.074 -> 67.834 | Improved |
| add class-wise spectral debiasing (Lcsd) (ablation) | DCA · Open-i to NIH-CXR avg of 3 diseases | AUROC: 62.074 -> 65.541 | Improved |
| add domain-wise spectral debiasing (Ldsd) (ablation) | DCA · Open-i to NIH-CXR avg of 3 diseases | AUROC: 62.074 -> 67.364 | Improved |
| add Lcsd to curriculum adaptation (ablation) | DCA · Open-i to NIH-CXR avg of 3 diseases | AUROC: 67.834 -> 69.350 | Improved |
| add Ldsd to curriculum adaptation (ablation) | DCA · Open-i to NIH-CXR avg of 3 diseases | AUROC: 67.834 -> 69.054 | Improved |
| add Ldsd to Lst+Lcsd (ablation) | DCA · Open-i to NIH-CXR avg of 3 diseases | AUROC: 69.350 -> 69.887 | Improved |
| add DCA vs Source-only on Atelectasis | DCA · Atelectasis 12 tasks avg | AUROC: 74.434 -> 75.994 | Improved |
| add DCA vs Source-only on Pneumonia | DCA · Pneumonia 12 tasks avg | AUROC: 70.098 -> 74.238 | Improved |
| add DCA vs Source-only on Cardiomegaly | DCA · Cardiomegaly 12 tasks avg | AUROC: 79.077 -> 81.471 | Improved |
| apply feature adaptation or self-training baselines | DANN, CDAN, MCD, DALN, ST, CST · 72 chest X-ray transfer tasks | avg improvement over Source-only: -0.590% to -3.854% | Hurt |
| DCA vs Source-only across 72 tasks | DCA · 72 chest X-ray transfer tasks | avg improvement: +3.196% | Improved |
| add DCA to reduce calibration error | DCA · CheXpert to NIH-CXR Atelectasis confidence 0.8-1.0 | ECE: 0.89 -> 0.20 | Improved |
Closest papers by meaning
- Bootstrapping Chest CT Image Understanding by Distilling Knowledge from X-ray Expert Models (CVPR 2024 · similarity 0.85)
- Novel Class Discovery in Chest X-rays via Paired Images and Text (AAAI 2024 · similarity 0.85)
- BenchX: A Unified Benchmark Framework for Medical Vision-Language Pretraining on Chest X-Rays (NeurIPS 2024 · similarity 0.84)
- CXReasonBench: A Benchmark for Evaluating Structured Diagnostic Reasoning in Chest X-rays (NeurIPS 2025 · similarity 0.83)
- Knowledge Transfer Scaling Laws for 3D Medical Imaging (NeurIPS 2026 · similarity 0.83)
- A Textbook Remedy for Domain Shifts: Knowledge Priors for Medical Image Analysis (NeurIPS 2024 · similarity 0.83)
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.
