Laurent Massoulié 2023 ACM SIGMETRICS Achievement Award
ACM SIGMETRICS is pleased to announce the selection of Dr. Laurent Massoulié of Inria as the recipient of the 2023 ACM SIGMETRICS Achievement Award in recognition of his contributions to the theory and practice of large-scale distributed networks and systems.
Dr. Massoulié is a Researcher at Inria, director of the Microsoft Research-Inria Joint Centre and member of the DYOGENE team. His research interests include unsupervised learning, distributed machine learning, modeling and algorithmic design for distributed systems and networks. His work has led to significant advancements in the field, particularly in the areas of congestion control, P2P networks and epidemic processes, community detection, and distributed optimization for federated learning. This research has provided critical insights into fundamental performance trade-offs and resolution of open problems, and allowed for the design of new algorithms for distributed control of systems.
Dr. Massoulié’s work has been influential and transformative, not just within the SIGMETRICS community, but also in other communities such as applied mathematics, theoretical computer science, and information theory. His contributions have earned numerous awards, including the SIGMETRICS’05 Best Paper Award, the 2019 Markov Lecture of the Informs Applied Probability Society, the Infocom’99 Best Paper Award, and two Best Paper Awards at NeurIPS’18 and NeurIPS’21.
Dr. Massoulié’s contributions to congestion control are particularly noteworthy. Among his many contributions, he has introduced new models of randomized and window-based congestion control, new analyses of congestion control stability with respect to feedback delays, flow-level models to characterize the performance of congestion controllers at the level of file transfers. Dr. Massoulié’s work on P2P networks and epidemic processes has also led to major advances, spanning across topics such as scheduling algorithms for P2P live streaming, algorithms based on distributed randomized graph rewirings to preserve connectivity in the presence of failures, epidemic (gossip-based) content dissemination in P2P networks.
More recently, his theoretical work has led to the first rigorous proof of the phase transition phenomenon in community detection for the stochastic block model, which sparked a large amount of follow-up work in theoretical computer science, information theory, and applied probability. Recently, he has further broadened his contributions by tackling open research questions in distributed optimization and federated learning. His most recent development proposes optimal distributed algorithms for fully asynchronous operation that only requires agents to rely on synchronized clocks. These schemes are the first to provably exhibit asynchronous accelerated convergence for arbitrary networks.
Additional information is available on his website: https://www.di.ens.fr/laurent.massoulie/
