Research map / Label-efficient and robust learning / Domain adaptation and generalization
Semi-supervised Deep Transfer for Regression without Domain Alignment
Mainak Biswas, Ambedkar Dukkipati, Devarajan Sridharan
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 CRAFT unsupervised contradistinguisher loss to supervised transfer | EEGNet-LSTM TL · Large Grid saccade (1% labels) | RMSE: 92.26 -> 84.17 | Improved |
| add CRAFT unsupervised loss to supervised transfer | SFCN TL · TLSA brain age (20% labels) | RMSE: 7.41 -> 7.14 | Improved |
| add CRAFT loss under biased label sampling with true marginal prior | SFCN TL · TLSA brain age (biased sampling) | RMSE: 8.40 -> 7.98 | Improved |
| add CRAFT unsupervised loss to supervised transfer | Attn-Resnet TL · JHU-crowd people counting (2% labels) | RMSE: 459.05 -> 550.77 | Hurt |
| add CRAFT unsupervised loss to supervised transfer | Resnet TL · Camelyon-17 tumor size (1% labels) | RMSE: 0.260 -> 0.243 | Improved |
| use CRAFT instead of DataFree | EEGNet-LSTM · Large Grid saccade (1% labels) | RMSE: 87.64 -> 84.17 | Improved |
| use CRAFT instead of Mixup | EEGNet-LSTM · Large Grid saccade (1% labels) | RMSE: 135.70 -> 84.17 | Improved |
| use CRAFT instead of DataFree | SFCN · TLSA brain age (20% labels) | RMSE: 7.36 -> 7.14 | Improved |
| replace Pyramidal CNN with EEGNet-LSTM decoder | source/target model · Large Grid saccade (100% labels) | RMSE: 63.84 -> 51.47 | Improved |
| increase unsupervised loss weight alpha from 0.1 to 1.0 (ablation) | CRAFT · TLSA brain age (20% labels) | RMSE: 7.14 -> 8.64 | Hurt |
| decrease unsupervised loss weight alpha from 0.1 to 0.01 (ablation) | CRAFT · TLSA brain age (20% labels) | RMSE: 7.14 -> 7.31 | Mixed |
| increase alpha from 0.1 to 1.0 (ablation) | DataFree · TLSA brain age (20% labels) | RMSE: 7.35 -> 8.70 | Hurt |
Closest papers by meaning
- Pseudo-Labeling for Unsupervised Domain Adaptation with Kernel GLMs (NeurIPS 2026 · similarity 0.86)
- Revisiting Source-Free Domain Adaptation: Insights into Representativeness, Generalization, and Variety (CVPR 2025 · similarity 0.86)
- A General Representation-Based Approach to Multi-Source Domain Adaptation (ICML 2025 · similarity 0.85)
- Diffusion-Driven Progressive Target Manipulation for Source-Free Domain Adaptation (NeurIPS 2025 · similarity 0.85)
- Universal Semi-Supervised Domain Adaptation by Mitigating Common-Class Bias (CVPR 2024 · similarity 0.85)
- Perturbating, Tuning, and Collaborating: Harnessing Vision Foundation Models for Single Domain Generalization on Medical Imaging (AAAI 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.
