Ryan Hoque

I am a first year PhD student in the UC Berkeley EECS department studying Robotics and Artificial Intelligence. I am advised by Ken Goldberg and am associated with Berkeley Artificial Intelligence Research (BAIR).

Before starting my PhD program, I spent some time in industry working on autonomous driving at Uber's Advanced Technologies Group (ATG) and earned my B.S. and M.S. summa cum laude in Electrical Engineering and Computer Science, also from UC Berkeley (Go Bears!). Outside of research, I enjoy rapping (both written and freestyle), playing the piano, reading and writing about philosophy, and parkour.

Email  /  CV  /  Google Scholar  /  Github  /  LinkedIn  /  Social

My master's and undergraduate research was focused on robotic manipulation of deformable objects. We have explored several deep reinforcement learning and imitation learning techniques in simulation ranging from visual foresight to dense object nets to achieve state-of-the-art results on fabric manipulation with real robotic systems. See our BAIR Blog post (in collaboration with Pieter Abbeel's lab) for a summary of our results and my M.S. thesis for a more thorough treatment.

By applying a model-based reinforcement learning technique we call VisuoSpatial Foresight, we train a visual dynamics model and use it to sequentially manipulate fabric toward a variety of goal images, entirely from random interaction RGBD data in simulation.

Here we learn correspondences between images of fabric to capture visual structure (as opposed to dynamical structure) and specify policies by demonstrating actions on a standardized configuration of the fabric (e.g. smooth).

In this paper we learn to smooth real fabric from arbitrarily complex starting states by leveraging our access to ground truth state in simulation.

Various other things you may find interesting.

Website template from Jon Barron.