Marco Cusumano-Towner

Marco Cusumano-Towner, PhD

I am a computer scientist working on probabilistic programming, probabilistic inference, cognitively inspired AI, and human-computer interaction.

I recently finished my PhD in EECS at MIT, advised by Vikash Mansinghka and Josh Tenenbaum. During my PhD I created the Gen probabilistic programming system. Before MIT, I was an early member of a cancer diagnostics startup, a Master’s student at Stanford University, and an undergraduate student in computer science at UC Berkeley, where I worked with Pieter Abbeel on probabilistic robotics. My graduate school research have been funded by the NSF graduate research fellowship and the NDSEG graduate fellowship program, among other sources.

The best way to contact me is via email at imarcoam at gmail dot com.

August 2020	I submitted the final version of my PhD thesis!
July 2020	I have new a paper on arXiv describing how Gen automates the low-level implementation of a flexible class of inference algorithms called involutive MCMC.
May 2020	I succesfully defended my thesis!
December 2019	Presenting our work on Structured differentiable models of 3D scenes via generative scene graphs at the NeurIPS Perception as generative reasoning workshop.
October 2019	Our paper Trace types and denotational semantics for sound programmable inference in probabilistic languages was accepted to POPL 2020 (conference talk by Alex Lew here).
October 2019	I gave an invited talk at the Workshop on Dependable and Secure Software Systems at ETH Zurich.
October 2019	Program commitee for 2nd Symposium on Advances in Approximate Bayesian Inference.
September 2019	I gave a talk on probabilistic programming at the Strange Loop conference in St. Louis.
July 2019	Alex Lew gave a talk on Gen in my place at JuliaCon.
June 2019	I presented our paper on Gen at PLDI (short video abstract here).
April 2019	Reviewed for UAI conference.
February 2019	I completed my EECS Research Qualifying Exam
January 2019	I co-taught an MIT IAP class on probabilistic programming in Gen.
January 2019	Feras Saad presented our paper on Bayesian program synthesis at POPL.
October 2018	Gave a talk on Gen at the first PROBPROG conference (video).
September 2018	First release of Gen.