Aaron Walsman

aaronwalsman (at) gmail.com     google scholar     github

I am a researcher pursuing a PhD in robotics, computer vision and artificial intelligence at the University of Washington advised by Dieter Fox and Ali Farhadi.

My recent work focusses on immitation learning from experts with asymmetric observation spaces (in submission, preprint available upon request) and training agents to solve complex assembly problems using LEGO bricks [1] github. I have also worked on embodied AI [2] [3], 3D computer vision [4] and deformable human pose tracking [5] [6].

Prior to beginning my PhD at the University of Washington, I completed a Masters in Robotics at Carnegie Mellon University advised by Siddhartha Srinivasa, where I was fortunate to work on enabling the robot HERB to perform in a play [7] and contributed to the YCB dataset [8].

Before working on robots and AI, I spent several years working as a technical artist on films such as Rio, Ice Age III and Horton Hears a Who.