معرفی
Elad Romanov is a postdoctoral researcher in the Department of Statistics at Stanford University, hosted by David Donoho. He completed his PhD at the Hebrew University of Jerusalem's School of Computer Science and Engineering, advised by Or Ordentlich and Matan Gavish. His research focuses on information theory, signal processing, high-dimensional statistics, and mathematical foundations of data science. Key contributions include ScreeNOT—a method for optimal singular value thresholding in correlated noise—and work on Gaussian mixture models, modulo-reduced measurements, and matrix denoising.
Education: PhD in Computer Science (Hebrew University of Jerusalem, advised by Or Ordentlich and Matan Gavish), followed by postdoctoral research at Stanford under David Donoho. His work bridges statistical theory and practical algorithms, with applications in optimization, matrix recovery, and information-theoretic limits.
Research interests emphasize statistical learning theory, robust estimation techniques, and algorithmic methods for high-dimensional data. His publications span topics like phase transitions in PCA, optimal spectral shrinkage, and the interplay between channel capacity and learning Gaussian mixtures. The ScreeNOT algorithm, implemented in multiple programming languages, provides a mathematically grounded alternative to heuristic singular value thresholding.
Collaborative projects include work on multi-reference alignment and matrix denoising with partial noise statistics. His contributions to modulo-reduced signal processing demonstrate recovery guarantees under non-traditional measurement constraints.
