Mark Goldstein

mark goldstein

goldstein [at] nyu [dot] edu
mgoldstein [at] flatironinstitute [dot] org
google scholar / papers
anonymous feedback
cv

I am a Research Fellow in the Center for Computational Mathematics at the Flatiron Institute. I focus on deep generative models and machine learning in the sciences.

My ongoing work is a mix of theory and methods for diffusion- and flow-based models and applications to chemistry and to nuclear fusion energy ⚛️. See here for previous work on conditional generative models for fluids/PDEs, design choices impacting generative models built on dynamical transport, coupling base and target densities in flow matching, estimating sequential causal effects, modeling the risk of cardiogenic shock, drug-induced ECG abnormalities, score matching with nonlinear diffusions + application to active matter systems, multimodal representation learning, variational learning of the diffusion noise schedules, and understanding generative models' likelihoods.


Previously, I completed my PhD at NYU's Courant Institute of Mathematical Sciences, advised by Rajesh Ranganath. I also worked closely with Eric Vanden-Eijnden. Before that, I studied music composition.


Я свободно говорю по-русски, но с американским акцентом придётся смириться. .أحكي عربي شوي


samarqand.