Research map / Label-efficient and robust learning / Domain adaptation and generalization
Improving Domain Adaptation Through Class Aware Frequency Transformation
Vikash Kumar, Himanshu Patil, Rohit Lal, Anirban Chakraborty
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 CAFT++ (class aware frequency transformation with ADT2P) | DSAN · Office-Home | Accuracy (%): 67.6 -> 72.1 | Improved |
| add CAFT++ | MDD · Office-31 | Accuracy (%): 89.3 -> 92.1 | Improved |
| add CAFT++ | DANN · VisDA | Accuracy (%): 78.8 -> 80.2 | Improved |
| add CAFT (without ADT2P/improvements) | DANN · VisDA | Accuracy (%): 78.8 -> 73.4 | Hurt |
| add CAFT | MCC · Office-Home | Accuracy (%): 72.3 -> 69.0 | Hurt |
| add CAFT | DANN · Office-31 | Accuracy (%): 86.1 -> 84.3 | Hurt |
| upgrade CAFT to CAFT++ | DSAN · Office-Home | Accuracy (%): 68.4 -> 72.1 | Improved |
| add CAFT++ to DANN under partial-set UDA | DANN · Office-Home Partial | Accuracy (%): 56.6 -> 58.4 | Improved |
| add CAFT++ to DANN under open-set UDA | DANN · Office-Home Open set (H-score) | H-Score (%): 64.3 -> 64.6 | No clear effect |
| use ADT2P instead of entropy for pseudo label filtering (ablation) | DSAN + CAFT++ · Office-Home | Accuracy (%): 71.1 -> 72.1 | Improved |
| add ADT2P pseudo label filtering versus threshold based labelling (ablation) | CAFT++ · Office-Home | Accuracy (%): average gain 2.1% | Improved |
| replace frequency swapping with CycleGAN style transfer (ablation) | CAFT++ + DANN · Office-Home | Accuracy (%): 66.0 -> 66.9 | Improved |
Closest papers by meaning
- Maximum Structural Generation Discrepancy for Unsupervised Domain Adaptation (TPAMI 2022 · similarity 0.89)
- Prior Knowledge Guided Unsupervised Domain Adaptation (ECCV 2022 · similarity 0.89)
- Can We Evaluate Domain Adaptation Models Without Target-Domain Labels? (ICLR 2024 · similarity 0.89)
- Incorporating Pre-Training Data Matters in Unsupervised Domain Adaptation (TPAMI 2025 · similarity 0.88)
- Partially-Shared Variational Auto-encoders for Unsupervised Domain Adaptation with Target Shift (ECCV 2020 · similarity 0.88)
- Make the U in UDA Matter: Invariant Consistency Learning for Unsupervised Domain Adaptation (NeurIPS 2023 · similarity 0.88)
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.
