Aryan Mokhtari

University of California, Berkeley, The University of Texas at Austin

Papers

Total Citations

H-Index

About

Aryan Mokhtari is a leading researcher in the field of large-scale optimization, with a primary focus on non-convex stochastic optimization and safe machine learning. His work addresses critical challenges in training models over massive datasets where both the number of samples and parameter dimensions are high. Mokhtari’s major contributions include the development of doubly stochastic successive convex approximation methods, which efficiently handle non-convex objectives with sparsity-promoting penalties—a key advancement for problems like dictionary learning. His most cited paper (13 citations) introduces a parallel stochastic successive convex approximation framework for large-scale dictionary learning, while another influential work (11 citations) extends this to high-dimensional non-convex settings. Notably, Mokhtari has also pioneered work in safe learning under uncertain objectives and constraints, addressing real-world safety-critical applications in robotics and medical procedures where constraints are unknown. His research bridges theoretical guarantees with practical algorithmic efficiency, making him a significant figure in modern optimization theory and its applications to machine learning.