Greg Izatt

Robotics PhD Student, MIT CSAIL

Contacts and pages

Email: gizatt at csail dot mit dot edu

Github: github.com/gizatt

Scholar: Google Scholar profile


About me

I’m a graduate student with Russ Tedrake at the Robot Locomotion Group at MIT CSAIL. I’ve tackled robotic control, perception, and simulation problems from a variety of angles, and have an eclectic toolbox spanning from bare-metal embedded electronics hacking to mixed-integer formulations of perception and inference problems (with a lot of neat stuff in between).

My research has broadly focused on the interplay of perception and simulation in robot manipulation. I’ve used simulation for many things, like testing Atlas’ walking behaviors during the DRC; prototyping perception and control systems for the MIT Hyperloop vehicle before its test flight; or putting robot arm code used by my lab under CI. But simulation can be so much more than that – in my PhD, I’m wrangling with some fundamental questions about how we can better align simulation to reality in a high-level way – how we can capture the distribution of real worlds our robots will face, and reproduce that distribution in sim.


Projects

Not all publications are listed here; these are just some recent highlights, including personal projects not reflected as publications. See below for my full list of publications.


Publications

See my Google Scholar profile for a fully up-to-date list.

    Gregory Izatt and Russ Tedrake. Capturing Distributions over Worlds for Robotics with Spatial Scene Grammars. PhD Thesis. Final revisions pending. video



    Gregory Izatt and Russ Tedrake. Scene Understanding and Distribution Modeling with Mixed-Integer Scene Parsing. In submission. pdf



    Gregory Izatt and Russ Tedrake. Generative modeling of environments with scene grammars and variational inference. At ICRA 2020. pdf video presentation



    Chris Sweeney, Greg Izatt, and Russ Tedrake. A supervised approach to predicting noise in depth images. At ICRA 2019. pdf video



    Gregory Izatt and Russ Tedrake. Robust object pose estimation with point clouds from vision and touch. Master’s thesis, MIT, Jun 2017. pdf



    Hongkai Dai, Gregory Izatt, and Russ Tedrake. Global inverse kinematics via mixed-integer convex optimization. At ISRR 2017 and in IJRR 2020. pdf



    Gregory Izatt, Hongkai Dai, Russ Tedrake. Globally Optimal Object Pose Estimation in Point Clouds with Mixed-Integer Programming. At ISRR 2017 and in IJRR 2020. pdf



    Gregory Izatt, Geronimo Mirano, Edward Adelson, Russ Tedrake. Tracking objects with point clouds from vision and touch. At ICRA 2017. pdf video



    Pat Marion, Maurice Fallon, Robin Deits, Andrés Valenzuela, Claudia Pérez D’Arpino, Greg Izatt, Lucas Manuelli, Matt Antone, Hongkai Dai, Twan Koolen, John Carter, Scott Kuindersma, Russ Tedrake. Director: A user interface designed for robot operation with shared autonomy. In JFR 2017. pdf



    Gregory Izatt, Julien Dubois, Nathan Faivre, Christof Koch. A direct comparison of unconscious face processing under masking and interocular suppression. In Frontiers in psychology, 2014. html



    Anupama J Thubagere, Wei Li, Robert F Johnson, Zibo Chen, Shayan Doroudi, Yae Lim Lee, Gregory Izatt, Sarah Wittman, Niranjan Srinivas, Damien Woods, Erik Winfree, Lulu Qian. A cargo-sorting DNA robot. In Science, 2017. pdf