Research map / Image generation / Fast diffusion sampling
CaO2: Rectifying Inconsistencies in Diffusion-Based Dataset Distillation
Haoxuan Wang, Zhenghao Zhao, Junyi Wu, Yuzhang Shang, Gaowen Liu, Yan Yan
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 objective-guided sample selection (ablation) | CaO2 · ImageWoof IPC=10 | accuracy: 38.7 -> 42.6 | Improved |
| add condition-aware latent optimization (ablation) | CaO2 · ImageWoof IPC=10 | accuracy: 38.7 -> 42.1 | Improved |
| combine selection and latent optimization (ablation) | CaO2 · ImageWoof IPC=10 | accuracy: 38.7 -> 45.6 | Improved |
| add both components (ablation) | CaO2 · ImageNette IPC=50 | accuracy: 82.0 -> 84.5 | Improved |
| add L-infinity regularization (ablation) | CaO2 · ImageNette IPC=10 | accuracy: 62.2 -> 63.5 | Improved |
| add L1 regularization (ablation) | CaO2 · ImageNette IPC=10 | accuracy: 62.2 -> 61.3 | Hurt |
| run selection first then optimization (vs reverse order) (ablation) | CaO2 · ImageWoof IPC=10 | accuracy: 37.3 -> 45.6 | Improved |
| use stronger selector models than ResNet-18 (ablation) | CaO2 · ImageNet subsets | accuracy | No clear effect |
| enlarge selection pool beyond 2x or 4x IPC (ablation) | CaO2 · ImageNet subsets | accuracy | No clear effect |
| plug CaO2 into MAR | MAR · ImageWoof IPC=10 | accuracy: 38.6 -> 43.9 | Improved |
| plug CaO2 into Minimax | Minimax · ImageNette IPC=10 | accuracy: 61.4 -> 64.1 | Improved |
| use soft labels instead of hard labels (ablation) | Minimax images · ImageNette IPC=50 | accuracy: 74.2 -> 83.9 | Improved |
Later papers that compared with this method
2 later accepted papers report CaO2 as a baseline in their results tables; 2 of them report a better number on most of the results they share with it.
- Set-Coupled Guidance: Set-Level Coordination in Diffusion-Based Dataset Distillation (ICML 2026 · better on 4 of 4 shared results)
- ProtoVAR: Efficient Dataset Distillation via Prototype-Guided Visual Autoregressive Modeling (ICML 2026 · better on 3 of 3 shared results)
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
- Unlocking Dataset Distillation with Diffusion Models (NeurIPS 2025 · similarity 0.90)
- MGD$^3$ : Mode-Guided Dataset Distillation using Diffusion Models (ICML 2025 · similarity 0.89)
- D^4: Dataset Distillation via Disentangled Diffusion Model (CVPR 2024 · similarity 0.89)
- Efficient Dataset Distillation via Minimax Diffusion (CVPR 2024 · similarity 0.89)
- Taming Diffusion for Dataset Distillation with High Representativeness (ICML 2025 · similarity 0.89)
- CADS: Unleashing the Diversity of Diffusion Models through Condition-Annealed Sampling (ICLR 2024 · 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.
