Diffusion Model Study Group Debut - Intro Session with MIT Curriculum

​🧠 TLDR

​With breakthroughs in image/video generation and diffusion-based LLMs, diffusion models have solidified their place as a core architecture in generative AI.

​Even if we don’t have huge chunks of time, just a few focused hours a week can go a long way toward understanding and building with this powerful framework.

​So, we're starting a new 5-month study group (2-4 hrs/week, max 12 ppl), for ourselves and our friends, on Diffusion Models & Flow Matching - based on MIT's curriculum. Starting from August 2nd and 9th, our 2 first sessions of the group, open to non-members. Inviting you as well!

🗓 Aug 2nd’s Agenda:

• ​[10min] Members introduction

• ​[15min] What are Flow Matching and Diffusion Models (with intuitive explanations) Real-world applications: from GenAI art to molecule generation

• ​[15min] What’s this 12-person peer-led study group? + The plan for our full 5-month hands-on course?

​Want to catch up with the latest advancements in diffusion models — while learning in a tight-knit group and building a model from scratch?

​

​Join this intro session for our upcoming Diffusion Model Study Group, starting August 2. We’ll be following MIT’s lecture notes and working with real engineers, researchers, and artists in GenAI.

​

​🤨 Why a 12-person study group?

​We’ve been running peer-led AI study groups for over a year — and based on the NTL Learning Pyramid, here’s what actually works:

​✅ Teaching others / immediate use – 90% retention

​✅ Building real projects – 75%

​✅ Group discussion – 50%

​Compare that to solo learning:

​📉 Lecture – 5%

​📉 Reading – 10%

​📉 Audio-visual – 20%

​

​✨ About the Difussion Model Study Group

• ​Peer-led, project-based sessions

• ​Rotate teaching & learn by doing

• ​TA Q&A + mentor support

• ​~2 hrs/week live + ~2 hrs/week self-paced

​

​🚀 What You’ll Learn

​MIT-based curriculum on diffusion models → View Lecture Notes PDF

​Topics include:

• ​Diffusion processes

• ​Training & denoising models

• ​Generative sampling

• ​Image/video generation

• ​Final outcome: Train your own model + build a GenAI app

​

​👥 Who’s Joining / Organizer

• ​Overall, this is a group of friends who have spent past one year learning through an LLM course. We are expanding our group as we are starting to learn the diffusion model together.

• ​Current group include

  • ​CTO of an AI film tool

  • ​AI art instructors

  • ​Full-time AI researchers

  • ​LLM instructors

• ​The group is organized by Ti Guo (Gen AI Data Scientist and organizer of 30+ study groups) and Colby Wang (LLM Researcher)

​📅 What’s Next?

​🌀 Aug 9: Second free intro session

• ​Learn the fundamentals of PDEs, SDEs & ODEs (math used in diffusion models)

• ​History and evolution of diffusion models

​🎯 After Aug 9, the full 5-month course will begin (paid members only)

​

​💸 Cost & Trial

• ​🆓 Free for first 2 weeks on Aug 2nd and Aug 9th

• ​💵 $50/month early sign-up with limited spots (goes up to $100 later, so sign up early if you want to secure the lower price)

• ​The money is used for paying the TAs for teaching and Q&A, as well as for our assistant for helping with coordination