September 2025 Newsletter

Upcoming Workshops

October 6 - 10, 2025: Data Assimilation and Inverse Problems for Digital Twins (Workshop 2 in the Fall 2025 Long Program on Digital Twins)

October 27 - 31, 2025: Optimal Control and Decision Making Under Uncertainty for Digital Twins (Workshop 3 in the Fall 2025 Long Program on Digital Twins)

November 10 - 14, 2025: Reduced Order and Surrogate Modeling for Digital Twins (Workshop 4 in the Fall 2025 Long Program on Digital Twins)

December 1 - 5, 2025: Applications of Digital Twins to Large-Scale Complex Systems (Workshop 5 in the Fall 2025 Long Program on Digital Twins)

January 12 - 16, 2026: Recent Advances in Random Networks: Theory and Applications

February 2 - 6, 2026: Advances in Quantitative Medical Care

Upcoming Long Programs

Accepting Applications for the Spring 2026 Long Program: Theoretical Advances in Reinforcement Learning and Control

The Long Program on Theoretical Advances in Reinforcement Learning and Control Spring 2026 Long Program (March 9 - May 29, 2026) is accepting applications. Reinforcement learning (RL) and control theory are concerned with training intelligent agents to make sequential decisions by interacting with an environment. In both formulations, an agent learns to navigate its surroundings through a process of trial and error, receiving feedback in the form of rewards or penalties based on the actions it takes. The agent’s goal is to learn a policy, a mapping from states to actions, that maximizes the cumulative reward over time. In the past few years, there has been a notable increase in enthusiasm for RL and the interplay between learning and control. The surge of interest is driven by the compelling application of RL and control methods to diverse challenges in artificial intelligence, robotics, and the natural sciences. Numerous breakthroughs owe their success to large-scale computational resources and creative deployment of adaptable neural network structures and training approaches, as well as both modern and traditional decision-making algorithms. Nevertheless, there remains a significant gap in our understanding regarding the conditions, reasons, and the degree to which these algorithms effectively operate. Such a challenge has drawn significant attention from various communities including computer science, numerical analysis, artificial intelligence, control theory, operations research, and statistics. This program aims to advance the theoretical foundations of reinforcement learning (RL) and control, and foster new collaborations between these researchers.

This Long Program is organized by Xinyi Chen (Princeton University), Elad Hazan (Princeton University), Cong Ma (University of Chicago), Nati Srebro (Toyota Technological Institute at Chicago), and Andrea Zanette (Carnegie Mellon University).

Apply here for Theoretical Advances in Reinforcement Learning and Control

Accepting applications for the Fall 2026 Long Program: Connectomics: Non-Euclidean Data Analysis for Brain Structure and Function

The Long Program on Connectomics: Non-Euclidean Data Analysis for Brain Structure and Function Fall 2026 Long Program (September 14 - December 11, 2026) is accepting applications. Brain connectomics offers a transformative approach to understanding the brain’s complex network of neural connections. By mapping how different brain areas interact and collaborate to support cognition, connectomics provides a comprehensive perspective on both normal and impaired brain function. This integrative approach is essential for studying brain development in children and aging-related physiological changes. Moreover, disruptions in neural connectivity play a central role in neurological and psychiatric conditions, including ADHD, autism and neurodegenerative diseases such as Alzheimer’s disease. The Connectomics: Non-Euclidean Data Analysis for Brain Structure and Function long program will focus on cutting-edge methodologies for analyzing brain structure and function. These include the application of Ordinary Differential Equations (ODEs) to model dynamic interactions between different brain regions and temporal changes in brain activity and system stability; the study of time-varying networks under various longitudinal designs and sampling scenarios; uncertainty quantification, inference and conformal prediction regions for networks and other brain characteristics; and assessing association between brain BOLD signals, aiming at a mathematical framework that leads to a better understanding of both normal development and pathological changes. More generally, the program will address the challenges of analyzing increasingly complex brain data, which often exist in non-Euclidean spaces such as networks, graphs, and high-dimensional data objects. Traditional statistical methods fall short when handling these types of data, making the advancement of non-Euclidean statistics crucial. By developing new tools and methodologies for metric statistics, this program will enable more accurate and interpretable analysis of brain connectivity and its impact on cognition and behavior. Aligning with the goals of the Brain Initiative, which seeks to accelerate neurotechnology development and enhance data science research, this program will explore the latest advancements in brain imaging, data integration, and personalized medicine. By bringing together experts from mathematics, statistics, neuroscience, and engineering, the program aims to foster interdisciplinary collaborations and push the boundaries of brain research. Through a series of lectures, workshops, and collaborative research projects, the program will provide participants with the tools and knowledge necessary to advance the understanding of brain function, improve disease diagnosis, and develop more targeted therapeutic interventions.

This Long Program is organized by Hans-Georg Muller (UC Davis), Alex Peterson (Brigham Young University), Yichao Wu (University of Illinois, Chicago), and Liang Zhan (University of Pittsburgh).

Apply here for Connectomics: Non-Euclidean Data Analysis for Brain Structure and Function

Accepting applications for the Spring 2027 Long Program: Modeling and Control of Vehicular Traffic and Transportation Systems

The Long Program on Modeling and Control of Vehicular Traffic and Transportation Systems Spring 2027 Long Program (March 8 - May 28, 2027) is accepting applications. Transportation science is a broad area attracting researchers from many different disciplines, such as applied mathematics, engineering, operations research, robotics physics, etc. Transportation in broad terms has been part of human activities since early history (see studies on traffic in ancient Rome). The birth of the modern era for vehicular traffic can be dated to 1920-30 with the advent of large-scale use of cars. It was precisely in the 1930s that Greenshield collected data using a camera and generated the first fundamental diagram (plot of flow versus density). The first partial differential equation models appear as early as the 1950s with the seminal work of Lighthill, Whitham, and Richards. Recently, the advent of innovative technology and the availability of large data sets created a demand and opportunity for new research using mathematical and computational models,including in areas such as robotics, autonomy, artificial intelligence, cognitive science,economics, and others. On the other side, there is an increasing interest in estimating the impact of traffic on society, from congestion to pollutant emissions, and the need to design efficient and equitable transportation systems aligned with the needs of different groups,while reducing the negative impacts of congestion and other traffic inefficiencies. This aspect attracted the interest of social scientists and other groups. The long program will explore new avenues of research by bridging different research communities and considering the impact of traffic flows at the societal level. The issues addressed will include but are not limited to, equitable transportation systems, technology impact on efficient transportation, energy and environmental footprint of vehicular traffic, automation and robotics and its social acceptance, large-scale data for traffic monitoring, and others. This program is timely, as human society is facing the challenges brought by increasing urbanization and pollution. The program will consist of an introductory tutorial workshop, which will bring up to speed young researchers or those starting to work in the area for the first time. Then four specific workshops will cover the main goals of the long program.

This Long Program is organized by Maria Chiri (Queen’s University), Qiang Du (Columbia University), Antonella Ferrara (Universita di Pavia), Paolo Goatin (INRIA), and Benedetto Piccoli (Rutgers University).

Apply here for Modeling and Control of Vehicular Traffic and Transportation Systems

IMSI Seeks Proposals for Scientific Activity

IMSI is currently seeking proposals for long programs, workshops, interdisciplinary research clusters, and other scientific activity. Information about how to submit proposals can be found on the proposal overview page and the resources linked therein. There are currently openings for long programs in 2027-28 and beyond, and openings for workshops in August 2027 through July 2028. IMSI holds two proposal cycles per year, with deadlines of March 15 and September 15.

Copyright © 2025. All rights reserved.

IMSI acknowledges support from the U.S. National Science Foundation
(Grant No. DMS-2425650)

Institute for Mathematical and Statistical Innovation
1155 E. 60th Street, Chicago, IL 60637