![Matías Salibián-Barrera Matías Salibián-Barrera](/sites/default/files/styles/award_winner/public/people/salibian-barrera-crm-ssc1.jpg?itok=MYyTNGV7)
The CRM-SSC Prize in statistics is awarded annually by the Centre de recherches mathématiques (CRM) and the Statistical Society of Canada (SSC). It is awarded in recognition of a statistical scientist's professional accomplishments in research during the first fifteen years after having received a doctorate. This year's winner is Matías Salibián-Barrera of the University of British Columbia (UBC).
Matías Salibián-Barrera is one of the brightest and most accomplished young statisticians in our country. He was born in Chile and grew up in Buenos Aires, Argentina. He obtained his Bachelor in Mathematics at the University of Buenos Aires, where he was introduced to Statistics, and in particular to Robustness, by Victor Yohai - himself a major force in this field.
Matías's doctoral dissertation was completed, in 2000, at UBC under the supervision of Ruben Zamar. The thesis, entitled Contributions to the Theory of Robust Inference, blends mathematical theory and computational procedures in a sophisticated manner that has continued throughout his career.
After graduation Matías was appointed Assistant Professor at Carleton University; after three years he returned in 2004 to UBC, where he is now Associate Professor. During his time at UBC he has also held Visiting Lectureships, designing and teaching short graduate level courses at Université libre de Bruxelles, Belgium, and at the University of Buenos Aires.
A remarkable feature of Matías's research is that his contributions are not only rigorously documented in good papers but also implemented in statistical freeware. He is well known and prized in the statistical community for his non-trivial implementation of 'state of the art' robust methods in R. His methodological contributions include the fast and robust bootstrap, uniform asymptotics for robust location and regression estimates, globally robust inference, robust smoothing, and robust functional data analysis. Complementing this, his computational work includes fast S- and fast tau-regression estimates, deep involvement with the construction of the S-plus 'robust' library and the R-package 'robustbase', linear clustering, and robust and sparse k-means.
The fast and robust bootstrap introduced in Matías's doctoral dissertation and subsequently developed in several joint papers with Stefan Van Aelst and Gert Willems represents a breakthrough in robust inference, by allowing the bootstrapping of robust methods. The straightforward application of the classical bootstrap to robust methods is not feasible because it does not yield robust inferences, and is much too slow. It has been adapted for numerous other scenarios, in particular for longitudinal studies and unbalanced clustering, by Alan Welsh (ANU) and collaborators.
Most proofs of asymptotic normality for robust procedures in the statistical literature use the unrealistic assumption of the validity of the central parametric model. This is unsatisfactory because robust methods are meant to be used with contaminated data. Matías's research deals with this problem and has produced very strong results on the uniform consistency and asymptotic normality in a neighbourhood of the central parametric model. Matías's introduction - jointly with Victor Yohai - of the fast regression S-estimator and the subsequent development of the fast tau-estimator are important breakthroughs for the efficient computation of these regression estimates. Similar ideas have also been used to compute multivariate location estimators.
More recently, Matías has turned his attention to functional principal component analysis. Dimension reduction associated either with variable selection in regression or the approximation of covariance matrices is an essential part of addressing the problems associated with high-dimensional data analysis. In a recent JASA paper Matías studies ways to find lower dimensional approximations which fit the functional data well and have minimum prediction error.
Matías's contributions to the profession go beyond his research. In service to the SSC he has served on the local organizing committee for the 2009 meeting in Vancouver and on the SSC Board. He has been a wonderful colleague in the Department of Statistics at UBC whose members, beyond pointing to his various research contributions, also emphasize his generous contributions to the department and the discipline. Matías is a valued Associate Editor for both The Canadian Journal of Statistics and Computational Statistics and Data Analysis.
Matías and his wife Veronica have been very busy raising three children. Matías enjoys hiking and learning photography. On a typical fall or winter evening you can find him at the soccer pitch, either coaching one of his sons or playing for one of his two teams. He enjoys a wide range of music styles, and will rarely miss a concert of his favourite Canadian band: Rush.
Matías Salibián-Barrera will present an overview of his work in a special session at the 43rd Annual Meeting of the Statistical Society of Canada to be held in Halifax, Nova Scotia, June 14 to 17, 2015.
"To Matías Salibián-Barrera for his fundamental contributions to the field of robust statistics, for the introduction of influential new methodology such as the fast and robust bootstrap and the fast S-estimator for robust regression, and for his breakthrough innovations in efficient computational algorithms for robust procedures."