Schedule

Dyadic Reinforcement Learning
Sequential decision making in digital health aims to enhance health outcomes by delivering interventions to individuals as they go about their daily life. The involvement of care partners and social support networks often proves crucial in helping individuals manage burdensome medical conditions. This presents opportunities in digital health to design interventions that target the dyadic relationship—the relationship between a target person and their care partner—with the aim of enhancing social support. In this paper, we develop dyadic RL, an online reinforcement learning algorithm designed to personalize intervention delivery based on contextual factors and past responses of a target person and their care partner. Here, multiple sets of interventions impact the dyad across multiple time intervals. We formally introduce the problem setup, develop dyadic RL and establish a regret bound. We demonstrate dyadic RL’s empirical performance through simulation studies on both toy scenarios and on a realistic test bed constructed from data collected in a mobile health study.

Complexity of proximal augmented Lagrangian methods and ADMM for solving constrained smooth nonconvex composite optimization problems
This talk discusses proximal augmented Lagrangian (PAL) methods and variants of the alternating minimization method of multipliers (ADMM) for solving constrained smooth non-convex composite optimization problems. Its purpose is to survey latest development and present our latest results related to the above topic. Special attention will be given to adaptive methods which do not require any specific knowledge about the problem instances (e.g., Lipschitz constants or lower curvature parameters for the objective and constraints functions) and hence are easy to run. We will also present numerical results demonstrating the significantly superior performance of adaptive methods compared to its non-adaptive versions. (joint work with William Kong and Arnesh Sujanani).

Random matrix theory and relationships with some interests from the 75 years of the STOR community
We will take a tour of some pioneering breakthroughs in the theory of random matrices and point out connections to some areas of interest of the STOR community at some point in its 75 years, such as statistical communication theory, extreme values, stochastic simulation, and complex data, among others.

Stochastic Modeling: Some STORies
I’ll present two war stories and, if time, describe some current work from my own experience, emphasizing themes and principles embodied by the stochastic modeling faculty in STOR at UNC. The first war story involves a court case on ambulance deployment. The second war story involves our (mostly) successful efforts to keep Cornell open for in-person instruction throughout the COVID 19 epidemic. The current work involves the design of volunteer schemes to decrease response times to the most urgent calls for medical assistance.
This is joint work with too many co-authors to list here.