Research map / Image generation / Fast diffusion sampling
Forward-Learned Discrete Diffusion: Learning how to noise to denoise faster
Grigory Bartosh, Teodora Pandeva, Sushrut Karmalkar, Javier Zazo
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 |
|---|---|---|---|
| reduce sampling steps from T=100 to T=10 | FLDD · ROCStories | MAUVE: 0.538 -> 0.511 | Mixed |
| reduce sampling steps from T=100 to T=10 | FLDD · ROCStories | PPL: 55.2 -> 60.5 | Hurt |
| reduce sampling steps from T=100 to T=10 | FLDD · QM9 | FCD: 0.328 -> 0.385 | Hurt |
| reduce sampling steps from T=100 to T=10 | FLDD · ZINC250k | FCD: 8.487 -> 10.414 | Hurt |
| reduce sampling steps from T=100 to T=10 | FLDD · QM9 | Validity: 99.67 -> 99.08 | Hurt |
| learn masking schedule conditioned on data instead of uniform unmasking (ablation) | masked diffusion model · binarized MNIST (4 steps) | sample quality | Improved |
| use FLDD instead of conventional discrete diffusion with few steps | FLDD · ROCStories | MAUVE: 0.511 (T=10) vs 0.598 (SEDD, step count not stated) | Mixed |
Closest papers by meaning
- Towards Non-Asymptotic Convergence for Diffusion-Based Generative Models (ICLR 2024 · similarity 0.91)
- Scalable Discrete Diffusion Samplers: Combinatorial Optimization and Statistical Physics (ICLR 2025 · similarity 0.91)
- Neural Flow Diffusion Models: Learnable Forward Process for Improved Diffusion Modelling (NeurIPS 2024 · similarity 0.91)
- DisCo-Diff: Enhancing Continuous Diffusion Models with Discrete Latents (ICML 2024 · similarity 0.91)
- O(d/T) Convergence Theory for Diffusion Probabilistic Models under Minimal Assumptions (ICLR 2025 · similarity 0.90)
- Why Masking Diffusion Works: Condition on the Jump Schedule for Improved Discrete Diffusion (NeurIPS 2025 · similarity 0.90)
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.
