Ph.D. Dissertation Defense in Computer Science titled “Optimization Methods and Software for Federated Learning” on May 8, 2025, at KAUST.

On May 8, 2025, Konstantin Burlachenko defended his Ph.D. in Computer Science at King Abdullah University of Science and Technology (KAUST).

The full text of the dissertation “Optimization Methods and Software for Federated Learning” has been archived in the KAUST Repository at http://hdl.handle.net/10754/704085.

The dissertation “Optimization Methods and Software for Federated Learning” identifies five key challenges in Federated Learning (FL), including data and device heterogeneity, communication issues, privacy concerns, and software implementations. More broadly, this work serves as a guide for researchers navigating the complexities of translating theoretical methods into efficient real-world implementations, while also offering insights into the reverse process of adapting practical implementation aspects back into theoretical algorithm design.

The dissertation, titled “Optimization Methods and Software for Federated Learning” was defended before a distinguished committee of global experts:

• Peter Richtarik - Professor at KAUST, Konstantin’s advisor, and one of the pioneers of Federated Learning. His contributions to randomized and distributed optimization methods are foundational and have shaped the future of the field.
• Stephen Boyd - Professor at Stanford University, a giant in convex optimization, control systems, and applied mathematics. His groundbreaking works continue to influence entire generations of engineers and applied mathematicians.
• Suhaib Fahmy - Professor at KAUST, a specialist in reconfigurable computing, systems architecture, and hardware-software co-design, helping to build the future of compute-efficient solutions at KAUST.
• Eric Feron - Professor at KAUST, an expert in control theory and autonomous aerospace systems, known for transforming our thoughts about intelligent systems and control theory.
• David Keyes - Professor at KAUST, Director of the Extreme Computing Research Center at KAUST, a pioneer in high-performance computing, parallel algorithms, and numerical analysis, pushing the limits of computational power.
• Nic Lane - Professor at the University of Cambridge, a leading voice in embedded machine learning and resource-efficient AI. His efforts to make Federated Learning practical and deployable pave the way for its real-world applications.