AI Talk: SyncSDE - A Probabilistic Framework for Diffusion Synchronization

​Sookwan Han is an undergraduate student at Seoul National University, with an already impressive track record of AI publications.

​He will be calling in to give a talk about his recent CVPR paper, which improves generative image (diffusion) models for broader applications such as:

• ​Text driven image editing

• ​Wide image generation

• ​3D mesh texturing

• ​Long horizon motion simulation

​The paper is called SyncSDE, and you can read more here: https://hjl1013.github.io/SyncSDE/.

​This is a technical talk, but we have Sookwan to design his presentation for a more applications focussed group. It should be interesting and all are welcome.

​Learn more about Sookwan's work here: https://jellyheadandrew.github.io/.