Full-day Workshop


Pieter Abbeel

Assistant Professor
Department of Electrical Engineering and Computer Science, UC Berkeley
Berkeley, California USA

Title: Machine Learning and Optimization for Robotics

In this talk I will describe two main ideas. First, I will describe apprenticeship learning, a new approach to equip robots with skills through learning from ensembles of expert human demonstrations. Our initial work in apprenticeship learning enabled the most advanced helicopter aerobatics to-date, including maneuvers such as chaos, tic-tocs, and auto-rotation landings which only exceptional expert human pilots can fly. In our current work we are studying how a robot could learn to perform challenging robotic manipulation tasks, such as knot-tying. Second, I will describe inverse optimal control, which considers the problem of observing a (potentially noisy) optimal controller and recovering the underlying cost function that is being optimized. This is important in human robot interaction, to enable robots to gauge their collaborators’ objectives, as well as in teaching robots how to perform complicated tasks. Third, I will describe advances in belief space planning, where, rather than planning in the original state space, we plan in the space of probability distributions over states. Optimal plans in belief space do not only plan for actions that affect the state of the system, but also for information gathering actions, which can be essential in the presence of significant uncertainty. While in general such problems are intractable, I will present approximate solutions obtained through Gaussian belief space planning that perform well in practice.

Pieter Abbeel received his Ph.D. degree in Computer Science from Stanford University in 2008 and is currently on the faculty at UC Berkeley in the Department of Electrical Engineering and Computer Sciences. He has won various awards, including best paper awards at ICML and ICRA, the Sloan Fellowship, the Air Force Office of Scientific Research Young Investigator Program (AFOSR-YIP) award, the Office of Naval Research Young Investigator Program (ONR-YIP) award, the DARPA Young Faculty Award (DARPA-YFA), the Okawa Foundation award, the TR35, the IEEE Robotics and Automation Society (RAS) Early Career Award, and the Dick Volz Best U.S. Ph.D. Thesis in Robotics and Automation Award. He has developed apprenticeship learning algorithms which have enabled advanced helicopter aerobatics, including maneuvers such as tic-tocs, chaos and auto-rotation, which only exceptional human pilots can perform. His group has also enabled the first end-to-end completion of reliably picking up a crumpled laundry article and folding it. His work has been featured in many popular press outlets, including BBC, New York Times, MIT Technology Review, Discovery Channel, SmartPlanet and Wired. His current research focuses on robotics and machine learning with a particular emphasis on challenges in personal robotics and surgical robotics.