Mikhail Belkin

Mikhail Belkin is HDSI Endowed Chair Professor in AI at Halicioglu Data Science Institute and Computer Science and Engineering Department at UCSD. From 2023 to 2025 he was an Amazon Scholar. His research interests are in theory and applications of Artificial Intelligence and Machine Learning.

His well-known work includes widely used Laplacian Eigenmaps, Graph Regularization and Manifold Regularization algorithms, bringing ideas from classical differential geometry and spectral graph theory to data science. His more recent work has been concerned with understanding remarkable mathematical and statistical phenomena observed in deep learning. The empirical evidence necessitated revisiting some of the classical concepts in statistics and optimization, including the basic notion of over-fitting. One of his key findings has been the “double descent” risk curve that extends the textbook U-shaped bias-variance trade-off curve beyond the point of interpolation. His recent work focuses on understanding feature learning and over-parameterization in AI models.

Mikhail Belkin is an ACM Fellow and a recipient of a NSF Career Award and a number of other awards. He had served as editor-in-chief of SIAM Journal on Mathematics of Data Science (SIMODS).

AI2050 Project

Modern AI systems work remarkably well, yet even their creators cannot fully explain how they work or what is really happening inside them. Today we mostly judge these systems from the outside, by what they say, which can hide what is actually driving their behavior. Belkin’s project develops the science to look inside: to find the hidden patterns an AI uses to reason, to test what those patterns control, and to adjust them directly. The goal is AI we can understand, monitor, and trust, so that as these systems grow more powerful they remain safe and beneficial.