Ulugbek S. Kamilov
504 posts

Ulugbek S. Kamilov
@ukmlv
Associate Prof at UW-Madison. Ex WashU, ENS Paris, Google, MERL, Stanford, MIT, CMU, EPFL. Tennis player. Computational Imaging + AI/ML.

Check the SPS Computational Imaging webinar "Computational Imaging: Stochastic Generative Plug-and-play Priors" by Prof. Ulugbek Kamilov @ukmlv on June 3rd, 9:30am EST. Click the link to view details and register! signalprocessingsociety.org/events/sps-ci-…

I’ll be presenting our work, “Measurement Score-Based Diffusion Model,” at #ICLR2026 in Rio 🇧🇷! I am happy to chat about diffusion/flow models, inverse problems, and generative modeling. (Poster session: Friday, P3-#705, 3:15–5:45 PM)



Congratulations to Dr. Yuyang Hu for successfully defending his PhD at WashU ESE. We are very impressed with all the research you have done over the past few years and wishing you even greater achievements in the years to come! 🎉


Excited to share our new work that bridges optimization and generative models for imaging. Our SGPnP is not a sampler (generative model), it is an optimization method (plug-and-play method) that can compete with samplers on inverse problems. Amazing work by @chicagoypark!


Countries where oranges are called “Portugal”

We introduce SGPnP, an optimization-based framework for highly ill-posed inverse problems. By injecting controlled noise into denoisers, SGPnP smooths the objective landscape, avoids strict saddles, and improves convergence—without diffusion sampling. > uw-cig.github.io/sgpnp_page






Congratulations to the former CIG undergraduate researcher – Eugene Joo – for accepting the offer to join the @Penn GRASP Lab as a PhD student. We are looking forward to hear more about his future achievements.







