Project Grant 2240788

Northeastern University was awarded a $250,000 Project Grant from the National Science Foundation Division of Electrical, Communications and Cyber Systems to support research titled "COLLABORATIVE RESEARCH: CONSENSUS AND DISTRIBUTED OPTIMIZATION IN NON-CONVEX ENVIRONMENTS WITH APPLICATIONS TO NETWORKED MACHINE LEARNING." Under the grant, Northeastern University will conduct research and education activities to advance the fundamental understanding of distributed non-convex optimization and apply these techniques to machine learning problems. Specifically, the university will investigate distributed retraction-free manifold optimization methods and distributed non-smooth weakly-convex optimization algorithms. The goal is to establish convergence results for asynchronous subgradient descent methods. As part of the broader impacts, the university will develop educational components including seminars and an advanced course on state-of-the-art decentralized optimization techniques. The award falls under the National Science Foundation's Engineering program (CFDA 47.041), which supports innovation and excellence in engineering research and education.