Summary: While we have seen tremendous progress in machine learning, a critical shortcoming of current methods lies in dealing with the distributional shift between training and deployment. Distributional shifts are ubiquitous in real-world problems, ranging from natural variation in distribution across locations or domains, to shifts in distribution arising from different decision-making policies, to shifts over time as the world changes. In this talk, I discuss three general principles for addressing these forms of distributional shift: pessimism, adaptation, and anticipation. I will present the most general form of each principle before giving concrete examples of its use in practice. This includes a simple method for substantially improving robustness against spurious correlations, a framework for quickly adapting a model to a new user or domain using only unlabeled data, and an algorithm that enables robots to anticipate and adapt to to adapt to shifts caused by other agents.

Biography: Chelsea Finn is an assistant professor of computer science and electrical engineering at Stanford University. Finn's research interests lie in the ability of robots and other agents to develop broad intelligent behavior through learning and interaction. To this end, her work has included deep learning algorithms for simultaneously learning visual perception and control in robot manipulation skills, inverse reinforcement methods for scalable acquisition of nonlinear reward functions, and meta-learning algorithms that enable rapid adaptation in a few steps. visual perception and deep reinforcement learning. Finn received her bachelor's degree in electrical engineering and computer science from MIT and her doctorate in computer science from UC Berkeley. Her research has been recognized by the ACM Doctoral Dissertation Award, the Microsoft Research Faculty Fellowship, the CV Ramamoorthy Distinguished Research Award, and the MIT Technology Review 35 under 35 Award, and her work has been discussed in several media outlets, including the New York Times, Wired, and Bloomberg. Throughout her career, she has sought to increase the representation of underrepresented minorities within CS and AI by developing an AI outreach camp in Berkeley for underprivileged high school students, a mentorship program for underrepresented students at four universities, and leading efforts within the WiML and Berkeley WiCSE communities of women researchers.

Slides: https://www.dropbox.com/s/hq1cgqvvdvoqq12/20210219_Chelsea_Finn_Principles_For_Tackling_Distribution_Shift.pdf?dl0

Please take the opportunity to connect and share this video with your friends and family if you find it useful.