Urbana Campus Research Calendar (OVCRI)

Back to Listing

NFI - Faculty Seminar Series: Bridging Control Theory and Robust Deep Learning (Hu) / Calibration and Image Formation in Interferometry (Kemball)

Event Type
New Frontiers Initiative
Oct 18, 2022   10:00 - 11:00 am  
Scott Lathrop
Originating Calendar
New Frontiers Initiative

Abstract for Bin Hu's talk

Recent years have witnessed phenomenal accomplishments of deep learning in many application domains such as computer vision, natural language processing, Go, personalized healthcare, scientific discovery, and robotic manipulation. For real-world applications, deep learning models that lack robustness against adversarial attacks can run into unexpected catastrophic failures. To address such security/safety risks, Lipschitz network training has emerged as a principled approach for enhancing the robustness of deep learning models in the presence of adversarial attacks, and various techniques that can induce Lipschitz bounds on deep neural networks have been developed in a case-by-case manner.  This talk focuses on the connections between control theory and robust deep learning.  Specifically, we will tailor control-theoretic tools to develop a principled unified framework for enforcing Lipschtiz bounds and inducing robustness properties of deep neural networks in the presence of adversarial attacks. We will discuss how to leverage control-theoretic conditions in the form of semidefinite programs to develop various Lipschitz network structures which can be very useful for robust learning tasks. Several new insights on how to mitigate adversarial attacks in the deep network regime are also presented. 


Abstract for Athol Kemball's talk 

Image formation for interferometric astronomical arrays requires solution of a challenging inverse problem that includes a joint solution for calibration and propagation effects as well as the unknown source brightness distribution. We consider this problem within the broader domain of adaptive optics and consider new approaches facilitated by large-scale computing. Interferometric image formation has many different domain applications and we also briefly discuss some overlapping concerns in these adjacent disciplines.


link for robots only