Raginsky studies communication in complex systems for decision-making and control
Though he joined ECE as a faculty member in January 2012, ECE Assistant Professor Maxim Raginsky was no stranger to the University of Illinois. During 2004-2007 he was a Beckman fellow studying problems at the intersection of information theory, decision making, artificial intelligence, and machine learning.
Raginsky, a researcher in the Coordinated Science Lab, received his bachelor’s and master’s in electrical engineering from Northwestern University in 2000. He then received his PhD in electrical engineering from Northwestern in 2002. After completing his Beckman fellowship in 2007, Raginsky went to Duke University, where he was a research scientist and later a research assistant professor.
Raginsky’s research focuses primarily on communications and control. He is interested in theoretical and practical aspects of information processing and decision-making in complex systems. He uses tools from information theory, machine learning, game theory, optimal control, and signal processing. His doctoral work was in quantum information, which studies computing devices that are based on the principles of quantum mechanics and that can solve certain computational problems much more efficiently than what is currently known to be possible with ordinary computers.
While these quantum computers remain a tantalizing possibility, Raginsky wants to use the background he acquired to look at complex systems that are driven by today’s needs and that people are building and testing right now, such as the Internet and other large scale networks.
“You can build very reliable and complex systems from very simple components,” said Raginsky. “My experience with probability and physics in the context of quantum information science has helped me see how to take this idea and to venture out into other domains.”
He added, “We have to start paying attention to the purpose of communication, so that the users get only the data relevant to their functional demands. The way we need to encode and decode high-dimensional data required for monitoring and control of large-scale decentralized infrastructures, such as smart grids or smart transportation systems, will necessarily be very different from how we approach coding for more traditional data types. A synthesis of information theory with machine learning and control theory will give me the tools needed to tackle these challenges."
With a passion for physical sciences, Raginsky has long wanted to be an engineer. He wanted the chance to create things with the freedom to imagine systems with different capabilities and to understand what you need to do in order to develop them.
This fall Raginsky will teach ECE 534: Random Processes. He has always enjoyed teaching. While a graduate student working on quantum information science, he noticed that more and more students were entering that field. To help in their entry into the complex topic, he developed a graduate seminar during the last year of his PhD.
“Teaching is interesting because I like discussing ideas and seeing different points of view,” Raginsky said. “Without that flow of ideas and seeing new connections, you tend to develop a narrow perspective for your area of expertise. Sometimes you don’t see connections to other disciplines in your field.”
Raginsky enjoys helping students appreciate the big picture of electrical engineering. He likes projecting his excitement, while being on the receiving end of learning new things in his field.
“Mentoring both undergrads and graduates is an exciting thing,” he said. “It’s a two-way process. You get to learn a lot from your students as you work with them.”
Raginsky also enjoys the interaction with colleagues at Illinois. “I never really got to experience such a collaborative spirit that we have here,” he said. “There is a concentration of people here that share the same goal as you do.”