Abstract: Gradient-based data attribution methods, such as influence functions, are critical for understanding the impact of individual training samples without repeated model retraining. However, their scalability is often limited by the high computational and memory costs associated with per-sample gradient computation, especially for large-scale models and datasets. In this talk, I will present our recent work on scalable influence function computation through sparse gradient compression and projection techniques with provable guarantees. I will also discuss how these methods can be applied to real-world scenarios, such as online reinforcement learning where data filtering interacts with policy learning.

Bio: Dr. Han Zhao is an Assistant Professor of Computer Science at the University of Illinois Urbana-Champaign (UIUC). He is also an Amazon Scholar at Amazon AI. Dr. Zhao earned his Ph.D. degree in machine learning from Carnegie Mellon University. His research interest is centered around trustworthy machine learning, with a focus on algorithmic fairness, robust generalization and data interpretability. He has been named a Kavli Fellow of the National Academy of Sciences. His research has been recognized through an NSF CAREER Award, a Google Research Scholar Award, an Amazon Research Award, and a Meta Research Award.

For more information, or for CODA guest access, please contact shatcher8@gatech.edu at least 2 business days prior to the event.

Join Via Zoom: https://gatech.zoom.us/j/98188267850?pwd=SOTPAZaZm0qkaiGxezfwMFaIbP1eeI.1
Meeting ID: 981 8826 7850 
Passcode: 520805