Laurent Cavalier

Laurent Cavalier, Professor of Mathematics at the Aix-Marseille Université, France, passed away on January 13, 2014, after being operated on for appendicitis in a hospital in Marseille, France. He was only 42 years old.

Laurent Cavalier was born on October 13, 1971. After studying pure mathematics at the University of Montpellier, he became interested in Statistics and moved to the Université Pierre et Marie Curie (Paris 6) where he was admitted to DEA of Statistics and then obtained his Master Degree in 1994. His PhD work at the Université Pierre et Marie Curie under my supervision was devoted to nonparametric estimation of level sets and statistical tomography. He defended his PhD thesis in 1998. The next year, he accepted a position of assistant professor (Maître de Conférences) at the Université Aix-Marseille 1 (later renamed as Aix-Marseille Université) where he worked until his death. He was promoted to the rank of associate professor and then to full professor.

Laurent Cavalier’s research was mainly devoted to statistical inverse problems. This new area has been rapidly developing over the last two decades and Laurent Cavalier was one of the leading researchers to whom this advance was due. He was working on many aspects of inverse problems. Among his most prominent contributions are the construction and refined study of adaptive methods (the risk hull method, and the block Stein methods), the derivation of exact asymptotics of the minimax risks in several statistical inverse problems, a pioneering work on the tomography problem with Poisson data, estimation in inverse problems with non-compact operators. His works on inverse problems with errors in the operator were also very influential.

Laurent Cavalier was a brilliant lecturer; he was invited to give advanced courses in several leading research institutions including Yale University, the University of Göttingen, Humboldt University and WIAS in Berlin, as well as at summer schools (Stats in the Château, 2009). He authored two surveys on inverse problems in statistics and was preparing a book on this subject but his premature death put a brutal end to all these projects.

Laurent Cavalier contributed a lot to the scientific life of the French statistical community. He was organizing, jointly with Oleg Lepski, the annual “Meeting on Mathematical Statistics” at CIRM, Luminy, France. This event was very stimulating for the development of statistical science in France and became a benchmark for several generations of young researchers. Laurent bore the main burden of local organizer of these meetings for many years. Later, with a group of other statisticians from France and Spain, he founded another important and now very popular event, Journées Statistiques du Sud, a two-day annual summer school featuring short courses on recent developments in mathematical and applied statistics. Laurent was also an active member of the scientific committee, Etats de la Recherche, of the French Mathematical Society (SMF).

Laurent was a “man of the south”: he enjoyed living in the south of France where he was born, near the sea and the mountains. He was very active in sports, such as hiking in the montains and especially football. He was an open and lovely person, and his untimely death profoundly saddened all those who knew him. Laurent is survived by his wife Elsa and his two sons, Gael and Romain.

Written by Alexandre Tsybakov, Professor at CREST-ENSAE