In memory of Ruben H. Zamar (1949–2023)


Ruben H. Zamar, professor of statistics at the University of British Columbia, passed away on June 28, 2023. Ruben’s inquisitive mind combined with his innovative ideas expanded and improved the field of robust statistics making it more relevant.

Ruben was born in San Pedro, in the province of Jujuy, Argentina, on December 20, 1949. After graduating as an accountant from the Nacional University of Cordoba in 1973, he began teaching at the National University of Rio Cuarto, Cordoba, Argentina. He completed two master's degrees, one in statistics (Chile, 1977) and one in mathematics (Brazil, 1981), before moving with his family to Seattle to continue his studies at the University of Washington. In 1985, Ruben completed his PhD in statistics under the supervision of Professor Douglas Martin. In 1986, he joined the recently created Department of Statistics at the University of British Columbia, Vancouver, Canada, which became his academic home for the rest of his fruitful career.

Ruben’s career started at the peak time of development and growth in the then emerging field of robust statistics. Ruben first worked on the general theory of quantitative robustness. Despite many exciting foundational contributions in the field, Ruben was unsettled by the limited tools available to assess the degree of robustness of different estimators. During his PhD studies, Ruben was enthused by the concept of maximum asymptotic bias (maxbias), originally proposed by Peter Huber for the simple location model (1964) but not fully examined in other scenarios. In his doctoral thesis, Ruben derived and studied maxbias functions and minimax bias estimators for more complex models, including linear and orthogonal regression. Ruben’s pioneering work in minimax bias theory was established as a very important theoretical tool in quantitative robustness which enabled the development of globally robust inference tools.

Ruben’s research contributions are notable by both their breadth and their depth. He published more than 80 peer-reviewed articles, most of them with distinguished collaborators and graduate students. He maintained a prolific network of national and international collaborations that produced trailblazing contributions to the field. Noteworthy examples include the development of a robust bootstrapping method for linear regression estimators, robust estimators for measurement error models, and a robust model selection method for complex high-dimensional settings. Ruben’s research spanned over many other challenging problems such as robust clustering, image processing, bioinformatics, and covariance matrix estimation.

In 2009, Ruben and colleagues introduced what quickly became a new paradigm in robust statistics: the independent contamination model, which describes the independent propagation of outliers as observed in modern high-dimensional data. Their ground-breaking paper in the Annals of Statistics opened a new path for theoretical and computational contributions in the field. In Ruben’s own words, “outlier propagation is a serious statistical problem and our model may become an important tool in the context of the robust analysis of high dimensional datasets.” Ruben’s high research standards were also recognized with associate editorships in the Annals of Statistics (2004–2007), Test (1998–2001) and the Journal of the American Statistical Association (1991–1993). Among many important awards, in 2021 Ruben was named a fellow of the Institute of Mathematical Sciences for his fundamental contributions in robust statistics.

His relentless commitment to research went hand in hand with his full dedication to his graduate students and the statistics community. During one of our last conversations, Ruben said, “I always did what I thought was important.” He was a dedicated educator and an exemplary advisor and mentor who worked tirelessly to support and stimulate his students’ ideas. He supervised eight postdoctoral fellows, 10 PhDs, and 16 MScs and also maintained a tight academic bond with researchers in his native country, Argentina, which resulted in many collaborations and the co-supervision of graduate students. Ruben is remembered by his students and collaborators as a talented, humble, and joyful person to work with.

Ruben was a kind colleague, generous with his time and ideas, and also a great mentor and friend. He had an uncanny ability for identifying and fostering unique qualities in everyone, and inspired many of us to be the best version possible of ourselves. The poet Patrick Lane wrote in his book There is a Season: “Whether or not the cairn is gone, the stones remain like ghosts in my hands and that is enough.” Ruben’s countless lessons and discussions, his contagious excitement for research, and his friendly laughs will always remain in our hearts. Ruben is survived by his wife Magui, sons Ruben Jr. and David, and grandchild Sebastian. They have always been at Ruben’s side, supporting his endeavours. He will be greatly missed.

Gabriela Cohen Freue and Matias Salibián-Barrera

Wednesday, August 30, 2023

Liaison Newsletter: