University of Utah electrical and computer engineering assistant professor Yi Zhou has received the National Science Foundation’s prestigious Faculty Early Career Development (CAREER) Award for his research into multi-agent reinforcement learning as a framework for managing multi-agent systems that involve heterogeneous agents in complex and structured environments. Zhou is the first John and Marcia Price College of Engineering faculty member to receive an NSF career award for 2023.
Reinforcement learning is a popular framework for learning optimal decision-making in complex environments, and many RL algorithms have been developed to improve decision-making of a single agent in normal environments. However, there are many large-scale RL applications that involve multiple heterogeneous agents interacting with complex environments such as navigating multiple drones in an open area. These cases make the optimal decision-making increasingly challenging to learn.
This $541,016 career award aims to support Zhou in developing a resilient RL framework that can manage multi-agent systems in complex environments. To accomplish this, he will be systematically designing efficient and resilient multi-agent RL algorithms and developing comprehensive convergence and complexity analysis in the perspectives of computation and communication. This project will produce RL algorithm software packages that will be fully accessible to the public.
“My earlier Ph.D. research focused on designing convergent and scalable nonconvex optimization algorithms for parallel and distributed computing, where the optimal decision variable is learned in a stationary environment,” explains Zhou. “After joining the University of Utah, my research has extended to developing algorithms for complex multi-agent RL systems in dynamic environments. This is a fundamentally more challenging and interesting research topic than conventional optimization; the research outcome will substantially improve the real-time decision-making in general distributed systems using the developed resilient and intelligent RL algorithms.”
In addition to advancements in RL-based control of these heterogeneous multi-agent systems, the research activities from this award will also generate a positive educational impact on undergraduate and graduate students. The materials developed by this project will be integrated into courses on machine learning and optimization and will benefit interdisciplinary students majoring in electrical and computer engineering, statistics and computer science. The project will actively involve underrepresented students and integrate research with education for undergraduate and graduate students in STEM. It will also produce introductory materials for K-12 students to be used in engineering summer research programs.
“It is my great honor to receive this CAREER award, especially under the new John and Marcia Price College of Engineering,” says Zhou. “I am super excited about the potential research and educational impact that this long-term project will generate, particularly the development of a series of new undergrad/graduate ECE courses on machine learning, optimization and reinforcement learning. I hope that this will attract more electrical and computer engineering students to the exciting area of data science and machine learning and spark new research directions across different engineering disciplines.”
Learn more about Electrical and Computer Engineering faculty research topics and discover ways to get involved in undergraduate research.