Radu Balan

Radu Balan is a Professor of Applied Mathematics in the Department of Mathematics at the University of Maryland, College Park. He is affiliated with the Center for Scientific Computation and Mathematical Modeling (CSCAMM) and the Norbert Wiener Center for Harmonic Analysis and Applications. His research bridges harmonic analysis, signal processing, and machine learning, with a strong focus on phase retrieval, frame theory, and deep neural network analysis.

He earned his Ph.D. in Applied Mathematics from Princeton University in 1998, advised by Ingrid Daubechies, and holds B.Sc. degrees in Physics and B.E. in Electrical Engineering from Romanian institutions. His academic journey includes a postdoctoral stint at IMA and IBM, followed by research roles at Siemens before joining the University of Maryland faculty in 2006 as an Associate Professor, rising to full Professor in 2011.

His research interests include harmonic analysis, statistical signal processing, machine learning, phase retrieval, and mathematical modeling. He has made significant contributions to the theory of frames, phaseless reconstruction, and the stability of deep neural networks. His recent work explores permutation-invariant representations, Lipschitz bounds in CNNs, and graph deep learning.

His recent publications (2020–2024) reveal a strong trend toward theoretical machine learning and its mathematical foundations, particularly in the stability and invertibility of neural networks and representations. He frequently publishes in top-tier journals such as IEEE Transactions, SIAM Journal, and Foundations of Computational Mathematics, often on arXiv, indicating an active and current research program in mathematical signal processing and deep learning theory.

Radu Balan has authored or co-authored over 60 peer-reviewed papers, edited five volumes of Excursions in Harmonic Analysis, and holds 25 U.S. and international patents in signal processing and source separation. His work has been supported by the National Science Foundation (NSF), Army Research Office (ARO), Simons Foundation, Siemens, and Verisk.

He has advised numerous students and collaborators, with research projects spanning statistical signal processing, blind source separation, and deep learning. His lab and research group are associated with CSCAMM and the Norbert Wiener Center, fostering interdisciplinary collaboration in mathematical modeling and computation.