Foundations and Frontiers of Robot Learning
Office Hours: Tentatively Tuesdays 11.15 a.m. - 12 noon in my office at AGH 342B.
Mondays, 3 hours (1.45–4.45 PM at AGH 214)
CIS 7000 – Ed Discussion (temporarily open for self-signups to Penn students)
This seminar will use structured debates to interrogate major research questions in the study of "physical intelligence". We will be primarily interested in computational implementations of learning and intelligence in robots, but will draw inspiration from adjacent fields, including psychology and cognitive science, computer vision, and control theory.
Each week, students will adopt roles — debaters, primer presenters, scribes, etc. — to ensure that every participant contributes meaningfully. The goal is to cultivate rigorous argumentation, deep reading, and collaborative scholarly output. Students will also work on a course project over the length of the semester.
There will be relatively little overlap with past iterations of CIS 7000 / ESE 6800 real-world robot learning or ESE 6510 physical intelligence. Instead, this class is intended as a potential follow-up class for students who have previously taken those other classes and are looking to delve deeper. The course is also targeted at students who are pursuing research on robot learning, control, or adjacent topics, even if they have not taken any of the aforementioned prior offerings.