Anastasios Kyrillidis has been promoted to the Noah Harding associate professor of Computer Science with tenure. Kyrillidis leads Rice’s CS Optimization Lab (OptimaLab). OptimaLab specializes in optimizing very complicated calculations.
Kyrillidis is particularly passionate about optimization for machine learning, but he is also “deeply interested in convex and non-convex algorithms and their analysis. As I often say to my students, 'Any problem that includes a math-driven criterion and requires an efficient method for its solution, is right up my alley.’”
Computer Science Department Chair Chris Jermaine described Kyrillidis as "One of our star researchers in the general area of AI and machine learning. He's a gifted mathematician, but also deeply concerned about practicality. To him, math is a tool to help understand the practical limitations of AI. And Tasos is a wonderful, charismatic teacher."
“The collaborative spirit at Rice is one of the things I value most about working here,” said Kyrillidis. “I always say to newcomers, 'At Rice, your next big collaboration is just a conversation away.' It's this open, supportive environment that makes our work not just productive, but genuinely enjoyable."
Prior to joining Rice’s Department of Computer Science in 2018, he was a Goldstine Postdoctoral Fellow at the IBM T. J. Watson Research Center (NY) for a year, and a Simons Foundation Postdoc member at the University of Texas at Austin for three years. He completed his Ph.D. in Computer Science at EPFL in Switzerland.
In 2022, Kyrillidis was the recipient of a five-year, $650,000 National Science Foundation (NSF) CAREER Award for his research proposal “Algorithmic foundations for practical acceleration in computational sciences.” His goal is to devise algorithmic foundations and theory that will accelerate problem solving, which would include the design of fast algorithms as an active research area in machine learning, information processing, and optimization research.
CAREER Awards, among the most competitive given by the NSF, are annually awarded to 400 scientists and engineers in support of “early career faculty who have the potential to serve as academic role models in research and education and to lead advances in the mission of their department or organization.”
Kyrillidis and OptimaLab are involved in a number of other high-profile research projects as well.
In collaboration with other Rice Computer Science and Electrical and Computer Engineering researchers, Kyrillidis and his lab secured a competitive grant funded by the NSF and Intel to develop a new class of distributed neural network training algorithms, called Independent Subnetwork Training (IST).
IST continues under a one-year, unrestricted Amazon Research Awards (ARA), which Kyrillidis was awarded based on his proposed research, “Efficient and affordable transformers for distributed platforms.” This research builds on his previous breakthroughs in optimization for large-scale systems. He was one of only 79 award recipients who represented 54 universities in 14 countries.
He has also been the recipient of a Microsoft Research Award to explore complementary machine learning (ML) research axes: efficient large-scale training of neural networks and efficient adaptation of large neural network models to learn new tasks with the goal of continual learning. Additional research includes work on NonConvex Quantum Characterization, based on a collaboration with IBM.
Kyrillidis also leads two projects, supported by Rice’s Ken Kennedy Institute: the QuanTAS project, stating that through this quantum computing research, his lab is “making strides in understanding the limits of quantum theory, algorithms, and systems.” And, AI-OWLS project, a cross-department collaboration with his colleagues in Computer Science, Electrical and Computer Engineering, and Computational Applied Mathematics and Operations Research, “where we're advancing cutting-edge AI through optimization and distributed systems design.”
“Each of these projects,” Kyrillidis explains, “represents a step forward in our understanding and application of complex computational concepts.”
Future research that Kyrillidis is particularly excited about is “expanding our understanding of how neural networks learn features. This is a fundamental question that has implications across machine learning applications.” He is also keen to “further explore the idea of mixtures of experts or, agent-of-agents.” This could lead to breakthroughs in understanding multi-agent systems — how they interact, why they can learn in collaboration, and when and why they fail. Understanding failure modes is just as crucial as understanding success.”
Lastly, Kyrillidis wants to “continue our work on optimization methods, particularly in how they affect progress in areas like machine learning and quantum computing. The interplay between optimization techniques and these cutting-edge fields is fascinating and ripe for discovery.”
In all these areas, his guiding philosophy remains the same: “Push the boundaries but always with an eye on practical applications. I believe the most impactful research bridges the gap between theoretical advances and real-world problems."