Research Interest
I am broadly interested in developing theory and algorithms that democratize performant deep reinforcement learning and enhance its sustainability when applied to robotic systems. Specifically, I believe this can be accomplished by integrating the three Rs of waste management (reduce, reuse, recycle) into the lifecycle of agent training, validation, and deployment. There are a number of open directions for this work, but I am particularly excited about the promise of this concept in generating generalizable AI at reduced computational cost.
|