
PhD candidate · MIT LIDS
Cole Becker
I’m a PhD candidate at MIT’s LIDS, advised by Pablo Parrilo and supported by an NSF Graduate Research Fellowship.
I work on making generative models faster and easier to adapt, using ideas from sampling, optimization, and control. I also collaborate with Asu Ozdaglar and Russ Tedrake.
Before MIT, I completed my undergraduate and master’s degrees at Princeton, working with Bartolomeo Stellato and Amir Ali Ahmadi.

CB
Selected publications
All publications →

Training and Evaluating Diffusion Policies with Long Context Lengths
arXiv preprint, 2026
Recent writing
All writing →Smoothed Score Queries and Complexity of Sampling
How smoothed-score queries improve the condition-number dependence of Gaussian sampling.
Diffusion for Bayesian inverse problems
Notes on diffusion models, Bayesian inverse problems, posterior sampling, and generative priors.