Schedule

This workshop will be held in-person at ICML 2026 on Friday, July 10 in HALL D1, Seoul. The program covers invited talks, contributed talks, posters, and a panel discussion.

Friday, July 10, 2026 · HALL D1, Seoul. Official ICML page: 4th Structured Probabilistic Inference & Generative Modeling.


Seoul Time (KST) Type Title & Speakers
9:00 - 9:20 Coffee Break
9:20 - 9:30 Opening Remarks
9:30 - 10:00 Invited Talk Mingyuan Zhou (Microsoft AI Superintelligence, UT Austin)
10:00 - 10:10 Contributed Talk Reward Score Matching: Unifying Reward-based Fine-tuning for Flow and Diffusion Models
10:10 - 10:20 Contributed Talk Exact Posterior Score Estimation for Solving Linear Inverse Problems
10:20 - 10:50 Invited Talk Jona Ballé (NYU)
10:50 - 11:00 Contributed Talk When are likely answers right? On Sequence Probability and Correctness in LLMs
11:00 - 11:10 Contributed Talk Structured Coupling for Flow Matching
11:10 - 11:40 Invited Talk Jiatao Gu (UPenn CIS / Apple MLR)
11:40 - 13:30 Poster Session & Lunch Break
13:30 - 14:30 Panel Session Diffusion Language Models vs Autoregressive Language Models?
Subham Sekhar Sahoo (MBZUAI - IFM)
Stefano Ermon (Stanford University)
Jiaxin Shi (Meta SuperIntelligence Labs)
14:30 - 15:00 Invited Talk Pilar Cossio (Flatiron Institute)
15:00 - 15:10 Contributed Talk ABC: Any-Subset Autoregression via Non-Markovian Diffusion Bridges in Continuous Time and Space
15:10 - 15:30 Contributed Talk Enhanced Diffusion Sampling: Efficient Rare Event Sampling and Free Energy Calculation with Diffusion Models
15:30 - 16:00 Invited Talk Soojung Yang
16:00 - 16:10 Closing Remarks
16:10 - 17:00 Poster Session