JapanTech

Diffusion Models: From Foundations to a Working Application

Three talks on diffusion and flow matching: foundations, diffusion LLMs and genome models, and building your own TTS model.

When
Where
KERNEL HONGO, 3F, 4-1-4 Hongo, Bunkyo City, Tokyo·In person
Region
Kanto (Tokyo)
Organizer
Tokyo AI
Language
EN
Source
Luma
Diffusion Models: From Foundations to a Working Application

Summary

A Tokyo AI technical session on diffusion models, from flow matching and core design choices to diffusion language models and applications in biology and speech synthesis. It is aimed at ML practitioners, AI engineers, researchers and research students who want intuition and concrete examples. Three talks with Q&A: Xuan Zhang (SmartNews) on understanding diffusion and flow matching, Federico Berto (Radical Numerics) on diffusion language models such as RND1 and genome foundation models such as Omnii, and Adam Lensenmayer (Shisa AI) on flow matching for audio, including a from-scratch TTS model for hobbyist hardware. Networking follows at KERNEL HONGO in Bunkyo.

About the community

Tokyo AI is an English-speaking community for AI engineers, researchers and practitioners in Tokyo. It runs regular evening technical sessions with talks, Q&A and networking, often co-hosted with companies building AI products.

#diffusion-models#flow-matching#generative-ai#text-to-speech#bioinformatics#machine-learning#llm#tokyo-ai#diffusion-llm#tech-talk