Adam Stooke


Hometown: O'Fallon, Illinois
UC Berkeley

Student Page:

Research:  Adam's research goal is to accelerate deep reinforcement learning.  A foundational element of this work is to adapt algorithms to better utilize modern computing hardware--chiefly, by developing parallelized techniques using GPU acceleration--to dramatically reduce experiment turnaround times.  This will enable further research into improved algorithms and neural network architectures, which in turn may enable learning ever more challenging tasks.

Adam is a PhD candidate in computer science at UC Berkeley, where he is advised by Professor Pieter Abbeel.  Adam received B.S. and M.S. degrees in physics from the U.S. Air Force Academy (Class of 2008) and U.C. Berkeley, respectively.  Subsequently, he developed space communication technologies at the Air Force Research Lab in Albuquerque, NM and served as a liaison at the Advanced Research Projects Agency--Energy in Washington, D.C.  Adam returned to graduate school and joined the deep learning community in 2015.