From Guided Sampling to Learned Solvers: Diffusion Models for Inverse Problems

Diffusion models have recently achieved remarkable success across a wide range of domains, establishing themselves as powerful generative priors for solving complex inverse problems. This work focuses on developing principled and efficient algorithms that leverage diffusion models for both sampling-based and learning-based reconstruction. In the first part, we will present two complementary strategies for guiding the sampling process toward solutions that satisfy measurement constraints. The first approach leverages distillation-based methods, a variational framework that enables fast and accurate inference with latent diffusion models such as Stable Diffusion. The second approach addresses scenarios where constraints are non-differentiable or implicit, enabling inference in settings where gradients are unavailable or unreliable, such as graph-structured data. In the second part, we will move beyond the use of fixed pre-trained diffusion models and instead learn models that directly circumvent the limitations of guidance-based sampling. Specifically, we will introduce a graph reconstruction framework that learns a flow model interpolating from initial predictions to clean graph data, providing a principled alternative to guidance techniques.

Dr. Santiago Segarra

Associate Professor of Electrical and Computer Engineering, Rice University on January 23, 2026 at 10:15 AM in EB2 1231
Join Zoom Webinar

Santiago Segarra received the B.Sc. degree in Industrial Engineering with highest honors (Valedictorian) from the Instituto Tecnológico de Buenos Aires (ITBA), Argentina, in 2011, the M.Sc. in Electrical Engineering from the University of Pennsylvania (Penn), Philadelphia, in 2014 and the Ph.D. degree in Electrical and Systems Engineering from Penn in 2016. From September 2016 to June 2018, he was a postdoctoral research associate with the Institute for Data, Systems, and Society at the Massachusetts Institute of Technology. He joined Rice University in 2018 as an Assistant Professor and, since July 2024, Dr. Segarra is a W. M. Rice Trustee Associate Professor in the Department of Electrical and Computer Engineering at Rice University. He also holds courtesy appointments in the Departments of Computer Science and Statistics. His research interests include network theory, data analysis, machine learning, and graph signal processing. Dr. Segarra received the 2011 Outstanding Graduate Award granted by the National Academy of Engineering of Argentina, the 2017 Penn’s Joseph and Rosaline Wolf Award for Best Doctoral Dissertation in Electrical and Systems Engineering, the 2020 IEEE Signal Processing Society Young Author Best Paper Award, the 2021 Rice’s School of Engineering Research + Teaching Excellence Award, three early career awards (NSF CAREER, ARO ECP, and ARI Early Career), and five best conference paper awards.

This lecture series features exciting and dynamic visiting and virtual speakers from across the range of ECE disciplines. Take some time every Friday morning to be inspired by these great scientists and engineers before heading into the weekend!