Seminar Series Archive

Recent work has shown deep learning can significantly improve time series prediction and dynamics learning. However, the inability to generalize under distributional shift limits its applicability to the real world. In this talk, I will demonstrate how to principally incorporate inductive biases such as graphs and symmetry to improve generalization. I will showcase their applications to challenging tasks including robotic manipulation, turbulence modeling, and AV trajectory prediction.

Rose Yu is an assistant professor at UCSD in CSE and a primary faculty in the AI Group. She received her Ph.D. from USC and was a postdoc at CalTech. Her research focuses on advancing machine learning techniques for large-scale spatiotemporal data analysis, with applications to sustainability, health, and physical sciences. Among her awards, she has won multiple industry faculty research awards, several best paper awards, and was nominated as one of the ’MIT Rising Stars in EECS’.