Research map / Label-efficient and robust learning / Domain adaptation and generalization
CoFiDA-M: Concept-Aware Feature Modulation for Cross-Domain Adaptation with Image-Only Inference
Nurjahan Sultana, Moi Hoon Yap, Xinqi Fan, Wenqi Lu
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 confidence gating to the MONET embedding (ablation) | CoFiDA-M teacher · Clinical avg melanoma recall | Melanoma recall (%): 43.20 -> 86.36 | Improved |
| add confidence gating to the MONET embedding (ablation) | CoFiDA-M teacher · Clinical avg AUROC | AUROC (%): 75.24 -> 83.81 | Improved |
| use FiLM modulation instead of concatenating concepts (ablation) | CoFiDA-M teacher · Clinical avg melanoma recall | Melanoma recall (%): 73.86 -> 86.36 | Improved |
| use FiLM modulation instead of concatenating concepts (ablation) | CoFiDA-M teacher · Clinical avg AUROC | AUROC (%): 81.51 -> 83.81 | Improved |
| add feature alignment to post-edit teacher features on top of logit KD (ablation) | CoFiDA-M student · Clinical avg AUROC | AUROC (%): 64.10 -> 67.50 | Improved |
| add feature alignment to post-edit teacher features on top of logit KD (ablation) | CoFiDA-M student · Clinical avg melanoma recall | Melanoma recall (%): 64.53 -> 77.89 | Improved |
| align to post-edit u_T rather than pre-edit v_T (ablation) | CoFiDA-M student · Clinical avg AUROC | AUROC (%): 66.45 -> 67.50 | Improved |
| align to post-edit u_T rather than pre-edit v_T (ablation) | CoFiDA-M student · Dermoscopic avg AUROC | AUROC (%): 70.68 -> 76.50 | Improved |
| distill concept-guided teacher into image-only student vs standard UDA without MONET (ablation) | CoFiDA-M · Dermoscopic avg melanoma recall | Melanoma recall (%): 69.50 -> 84.92 | Improved |
| distill concept-guided teacher into image-only student vs standard UDA without MONET (ablation) | CoFiDA-M · Clinical avg AUROC | AUROC (%): 62.32 -> 67.50 | Improved |
| distill from random-weight or zero-weight teacher instead of trained teacher (ablation) | CoFiDA-M student · MILK Clinical | feature separation | Hurt |
| apply CoFiDA-M on Fitzpatrick vs source only | CoFiDA-M · Fitzpatrick (unseen) | Melanoma recall (%): 35.45 -> 59.46 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- TENT — Tent: Fully Test-Time Adaptation by Entropy Minimization (ICLR 2021) · better on 1 of 1 shared result
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Doctor Approved: Generating Medically Accurate Skin Disease Images through AI-Expert Feedback (NeurIPS 2025 · similarity 0.86)
- MICA: Towards Explainable Skin Lesion Diagnosis via Multi-Level Image-Concept Alignment (AAAI 2024 · similarity 0.86)
- A Textbook Remedy for Domain Shifts: Knowledge Priors for Medical Image Analysis (NeurIPS 2024 · similarity 0.85)
- DermaCon-IN: A Multiconcept-Annotated Dermatological Image Dataset of Indian Skin Disorders for Clinical AI Research (NeurIPS 2025 · similarity 0.85)
- Spider: A Unified Framework for Context-dependent Concept Segmentation (ICML 2024 · similarity 0.85)
- FOCUS: Knowledge-enhanced Adaptive Visual Compression for Few-shot Whole Slide Image Classification (CVPR 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.
