Paper: DMD turns diffusion models into looped models iterating in data space
CSProfKGD · x · 2026-10-12
A new paper, Few-Step Generation via Data-Space Iteration, proposes a paradigm shift for few-step generation: instead of iterating in flow space, it uses distribution matching distillation (DMD) to turn diffusion/flow models into looped models that iteratively refine samples directly in data space. Commenters note strong ties to self-conditioning and fixed-point iteration.
Related event: ByteDance Seed Proposes Data-Space Iteration for Few-Step Generation(5 posts)→
More from Research
- Reward model shows near-perfect progress prediction on bimanual precision robot task — DominiqueCAPaul · 2026-10-13
- Matthew Green uses AI to mine 16,000 papers and quantify global cryptanalytic effort — matthew_d_green · 2026-10-13
- Muse touts new coding model on ProgramBench, where top models barely hit 4.5% — jyangballin · 2026-10-13
- 127,833 real clinician queries show LLM benchmarks miss actual clinical use — NYU-OLAB · 2026-10-13
- HKU's JoHan recovers smooth 3D hand motion from video via generative 2D-3D joint modeling — HKUCDS · 2026-10-13
- Meta Releases AutoformBench: 341-Task Lean Formalization Benchmark With Cost Curves — ChrSzegedy · 2026-10-13