Research map / Image generation / Fast diffusion sampling
Star-Shaped Denoising Diffusion Probabilistic Models
Andrey Okhotin, Dmitry Molchanov, Arkhipkin Sergeevich Vladimir, Grigory Bartosh, Viktor Ohanesian, Aibek Alanov, Dmitry P. Vetrov
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 |
|---|---|---|---|
| use Dirichlet star-shaped diffusion instead of Gaussian DDPM | DDPM · Dirichlet simplex data | KL: 0.200 -> 0.011 | Improved |
| use Wishart star-shaped diffusion instead of Gaussian DDPM | DDPM · Wishart p.d. matrices | KL: 0.096 -> 0.037 | Improved |
| use Categorical SS-DDPM instead of Multinomial Text Diffusion | MTD · text8 | NLL (bits/char): 1.72 -> 1.69 | Improved |
| use Beta noise instead of Gaussian noise | DDPM · CIFAR-10 | FID: 3.17 -> 3.17 | No clear effect |
| use a Markov reverse process instead of conditioning on the whole tail (ablation) | SS-DDPM · CIFAR-10 samples | sample realism | Hurt |
| fewer sampling steps, Beta SS-DDPM vs DDPM | Beta SS-DDPM · CIFAR-10 | FID | Improved |
| compare DDIM with SS-DDPM and DDPM at few steps | DDIM · CIFAR-10 | FID | Mixed |
Closest papers by meaning
- UDPM: Upsampling Diffusion Probabilistic Models (NeurIPS 2024 · similarity 0.91)
- Assessing the quality of denoising diffusion models in Wasserstein distance: noisy score and optimal bounds (NeurIPS 2025 · similarity 0.90)
- Denoising Diffusion Step-aware Models (ICLR 2024 · similarity 0.90)
- SPD-DDPM: Denoising Diffusion Probabilistic Models in the Symmetric Positive Definite Space (AAAI 2024 · similarity 0.90)
- Adapting to Unknown Low-Dimensional Structures in Score-Based Diffusion Models (NeurIPS 2024 · similarity 0.90)
- O(d/T) Convergence Theory for Diffusion Probabilistic Models under Minimal Assumptions (ICLR 2025 · similarity 0.89)
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.
