University of Pennsylvania’s Eric Eaton presents “Efficient Lifelong Machine Learning” as part of the IRIM Robotics Seminar Series. The event will be held in the TSRB Banquet Hall from 12-1 p.m. and is open to the public.
Lifelong learning is a key characteristic of human intelligence, largely responsible for the variety and complexity of our behavior. This process allows us to rapidly learn new skills by building upon and continually refining our learned knowledge over a lifetime of experience. Incorporating these abilities into machine learning algorithms remains a mostly unsolved problem, but one that is essential for the development of versatile autonomous systems.
In this talk, I will present our recent progress in developing algorithms for lifelong machine learning. These algorithms acquire knowledge incrementally over consecutive learning tasks, and then transfer that knowledge to rapidly learn to solve new problems. Our approach is highly efficient, scaling to large numbers of tasks and amounts of data, and provides a variety of theoretical guarantees on performance and convergence. I will show that our lifelong learning system achieves state-of-the-art results in multi-task learning for classification and regression on a variety of domains, including facial expression recognition, land mine detection, and student examination score prediction. I will also describe how lifelong learning can be applied to sequential decision making for robotics, demonstrating accelerated learning for optimal control on several dynamical systems, including an application to quadrotor control. Finally, I will discuss our work toward autonomous cross-domain transfer, enabling knowledge to be automatically transferred between different task domains.
Eric Eaton is a faculty member in the Department of Computer and Information Science at the University of Pennsylvania and a member of the GRASP (General Robotics Automation, Sensing, Perception) lab. Prior to joining Penn, he was a visiting assistant professor in the computer science department at Bryn Mawr College. His primary research interests lie in the fields of machine learning, artificial intelligence, and data mining with applications to robotics, search & rescue, environmental sustainability, and medicine. In particular, his research focuses on developing versatile AI systems that can learn multiple tasks over a lifetime of experience in complex environments, transfer learned knowledge to rapidly acquire new abilities, and collaborate effectively with humans and other agents through interaction.
Before moving into academia, Eaton spent two years as a senior research scientist at Lockheed Martin Advanced Technology Laboratories working in applied research. At Lockheed Martin Atlanta, he led a number of machine learning research projects in the Artificial Intelligence Lab with a focus on their application for a variety of DoD organizations. While at Lockheed Martin, he was also a part-time faculty member in computer science at Swarthmore College.
Eaton received his Ph.D. in computer science from the University of Maryland, Baltimore County (UMBC), focusing on artificial intelligence and machine learning. His dissertation developed methods for selective knowledge transfer between learning tasks and was advised by Marie desJardins. At UMBC, he was a member of the Multi-Agent Planning and LEarning (MAPLE) research group and also a part-time instructor.