Simplex Diffusion Models paper talk by EPFL/DeepMind's Justin Deschenaux on Oct 19
ArnaudDoucet1 · x · 2026-10-16
A community session announcement: on Monday, Oct 19, Justin Deschenaux (EPFL, Google DeepMind) will present his paper Simplex Diffusion Models. The slides outline a four-step forward process on the simplex — multiplicative thinning, Dirichlet innovation noise, Beta-distributed mixing weights, and reverse-state update — with visualizations at κ=0.65 and mixing-weight means of 0.49 and 0.20.
Related event: EPFL/DeepMind Researcher to Present Simplex Diffusion Models Paper Online(2 posts)→
More from Research
- A beginner-friendly introduction to causal inference kicks off, Part 1 — Nasereliver · 2026-10-16
- FBI method trains for Pass@K with one sample per update at 12.3% of the cost — EmtiyazKhan · 2026-10-16
- Berkeley researchers sample humanoid robot policies directly from weight space with diffusion — berkeley_ai · 2026-10-16
- OneWM: unified causal training shows world and action modeling improve each other — IAAI233 · 2026-10-16
- Baidu's Virtual Head Attention tames the KV cache, FLOPs, quality triangle in long-context LLMs — baidu · 2026-10-16
- 5.8M-param Speech Event Detector adds laughs, breaths and背景 sounds to Whisper transcripts — Ayush Kumar Bar · 2026-10-16