Research map / Label-efficient and robust learning / Domain adaptation and generalization
Channel-masked Asymmetric Distribution Matching for Cross-Domain Generalized Dataset Distillation
Qi Liu, Chenghao Xu, Jiexi Yan, Guangtao Lyu, Erkun Yang, Guihai Chen, Yanhua Yang
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 asymmetric distribution matching (channel masking of real features) (ablation) | MMD baseline · PACS ACS->P, IPC 1 | accuracy (%): 16.4 -> 23.6 | Improved |
| add asymmetric distribution matching (channel masking of real features) (ablation) | MMD baseline · PACS ACS->P, IPC 10 | accuracy (%): 39.9 -> 45.9 | Improved |
| add masked classification regularization and dual consistency loss on top of Ladm (ablation) | CADM · PACS ACS->P, IPC 10 | accuracy (%): 45.9 -> 48.4 | Improved |
| add Lreg and Lcons without Ladm (ablation) | MMD baseline · PACS ACS->P, IPC 1 | accuracy (%): 16.4 -> 23.1 | Improved |
| replace MMD with CADM distillation | ResNet-18 distilled images · PACS ACS->P cross-architecture, IPC 10 | accuracy (%): MMD 38.8 -> CADM 44.4 (AlexNet) | Improved |
| add Fourier augmentation (FACT) to MMD | MMD · PACS ACP->S, IPC 1 | accuracy (%): 4.8 -> 5.7 (S target); VLCS/others vary | Mixed |
| add IRM to MMD pretraining | MMD · PACS ACS->P, IPC 10 | accuracy (%): 40.8 -> 38.4 | Hurt |
| use CADM instead of MMD+DAM | distillation · PACS in-domain, IPC 10 | accuracy (%): 52.2 -> 58.4 | Improved |
Closest papers by meaning
- Large Scale Dataset Distillation with Domain Shift (ICML 2024 · similarity 0.87)
- D^4: Dataset Distillation via Disentangled Diffusion Model (CVPR 2024 · similarity 0.86)
- Efficient Dataset Distillation for Pre-Trained Self-Supervised Models via Statistical Flow Matching (NeurIPS 2026 · similarity 0.86)
- Real-Fake: Effective Training Data Synthesis Through Distribution Matching (ICLR 2024 · similarity 0.86)
- Optimizing Distributional Geometry Alignment with Optimal Transport for Generative Dataset Distillation (NeurIPS 2025 · similarity 0.86)
- Dataset Distillation with Neural Characteristic Function: A Minmax Perspective (CVPR 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.
