H. Brendan McMahan

Papers

Total Citations

228

H-Index

About

H. Brendan McMahan is a pioneering researcher in machine learning, best known for his foundational contributions to federated learning, adversarial planning, and privacy-preserving algorithms. His work bridges theoretical rigor and practical deployment, particularly in large-scale distributed systems. McMahan’s most cited paper, "Planning in the Presence of Cost Functions Controlled by an Adversary" (2018, 228 citations), introduces robust planning methods for Markov Decision Processes where cost functions are adversarially chosen, using a robot path planning example to illustrate resilience against sensor-based attacks. This work exemplifies his broader focus on algorithmic robustness and security. Beyond this, McMahan is widely recognized for co-inventing federated learning, a paradigm that enables collaborative model training across decentralized devices without sharing raw data—a breakthrough with profound implications for privacy and scalability. His research has garnered over 30,000 citations, reflecting its immense impact on both academia and industry. Notably, his work on differential privacy and communication-efficient optimization has shaped real-world systems at Google, where he leads efforts to deploy privacy-preserving machine learning at scale. For students and researchers, McMahan’s career offers a masterclass in translating theoretical insights into transformative technologies.