Research map / Image generation / Fast diffusion sampling
Diffusion Reconstruction-based Data Likelihood Estimation for Core-Set Selection
Mingyang Chen, Jiawei Du, Bo Huang, Yi Wang, Xiaobo Zhang, Wei Wang
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 |
|---|---|---|---|
| replace Forgetting score with DRD score under the same BWS selection (ablation) | BWS selection · ImageWoof (ResNet-18), 10% selection | test accuracy: 70.1 -> 74.5 | Improved |
| replace Forgetting score with DRD score under the same BWS selection (ablation) | BWS selection · ImageWoof (ResNet-18), 75% selection | test accuracy: 87.3 -> 88.7 | Improved |
| use BWS instead of CCS with DRD score (ablation) | DRD · ImageWoof (ResNet-18), 30% selection | test accuracy: 80.1 -> 83.3 | Improved |
| use BWS instead of CCS with Forgetting score (ablation) | Forgetting score · ImageWoof (ResNet-18), 10% selection | test accuracy: 68.4 -> 70.1 | Improved |
| use DRD instead of Forgetting under CCS selection (ablation) | CCS selection · ImageWoof (ResNet-18), 30% selection | test accuracy: 80.9 -> 80.1 | Hurt |
| use IB-informed class-wise timestep instead of fixed timesteps (15, 20, 25, 30) (ablation) | DRD · ImageNette and ImageWoof | test accuracy | Improved |
| increase noise samples from 5 to 20 in diffusion classifier (ablation) | DRD timestep selection · ImageWoof (ResNet-18), 10% selection | test accuracy: 67.8 -> 74.5 | Improved |
| increase noise samples and B from 20 to 40 (ablation) | DRD timestep selection · ImageWoof (ResNet-18) | test accuracy: 74.5 -> 74.6 at 10% | No clear effect |
| select a mid-range window starting at 20%-40% of the DRD-sorted list instead of extremes (ablation) | DRD with BWS · ImageNette and ImageWoof | test accuracy | Improved |
| use DRD score instead of Forgetting score in BWS at ratios up to 40% (ablation) | BWS · ImageNet-1K (ResNet-50) | test accuracy: average margin 2.0% | Improved |
Closest papers by meaning
- Provable Posterior Sampling with Denoising Oracles via Tilted Transport (NeurIPS 2024 · similarity 0.89)
- Diffusion Models With Learned Adaptive Noise (NeurIPS 2024 · similarity 0.89)
- Adapt and Diffuse: Sample-adaptive Reconstruction via Latent Diffusion Models (ICML 2024 · similarity 0.88)
- Provably Robust Score-Based Diffusion Posterior Sampling for Plug-and-Play Image Reconstruction (NeurIPS 2024 · similarity 0.88)
- Analyzing and Guiding Zero-Shot Posterior Sampling in Diffusion Models (NeurIPS 2026 · similarity 0.88)
- O(d/T) Convergence Theory for Diffusion Probabilistic Models under Minimal Assumptions (ICLR 2025 · 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.
