Daniela Witten and Limin Peng
The Committee of Presidents of Statistical Societies selected Robert Tibshirani, professor of biomedical data science and professor of statistics at Stanford University, for the 2024 Distinguished Achievement Award and Lectureship. The award recognizes researchers who have made exceptional contributions to statistical methods with significant impact on scientific investigations. Tibshirani delivered the lecture at the 2024 Joint Statistical Meetings in Portland, Oregon. His lecture was titled “Pre-training and the Lasso.”
Tibshirani has played a key role in many of the most important statistical developments of the past 40 years, including generalized additive modeling, false discovery rate estimation, the lasso and related methods for high-dimensional modeling, and post-selection inference. His early research on generalized additive models dramatically extended the flexibility of traditional linear regression, and generalized additive models are now standard techniques for nonparametric multiple regression.
Citation: For fundamental contributions to statistics and machine learning that have deepened, broadened, and created a bridge between those fields; for bringing key statistical ideas in multiple testing and high-dimensional learning to the broader scientific community; for high-impact textbooks on generalized additive models, the bootstrap, high-dimensional statistics, and statistical learning that have come to define those fields; and for outstanding mentoring of PhD students and junior researchers.
Tibshirani’s work on the Least Absolute Shrinkage and Selection Operator (Lasso) was a breakthrough in statistical methodology and theory that has transformed the practice of feature selection and high-dimensional modeling in biomedicine and other scientific fields. The original paper on the Lasso has been cited over 55,000 times. Tibshirani’s foundational contributions to the field of machine learning have bridged the gap between the algorithmic-type thinking that is pervasive in the field, and “classical” statistical thinking. He has also had a substantial impact on the field of genomics, through his pioneering work tackling the statistical challenges associated with high-throughput gene expression data, and by popularizing ideas in multiple testing and false discovery rate estimation for a wide biological audience. More recently, Tibshirani and his collaborators played a key role in developing the area of post-selection inference for penalized regression models; this work has contributed to moving the field of statistical machine learning from its original focus on prediction to its more recent focus on uncertainty quantification.
Tibshirani’s work has received almost 500,000 citations, and he has an h-index of 181. In addition to hundreds of published articles, he has co-authored five best-selling textbooks on topics ranging from the bootstrap to generalized additive models to statistical machine learning. His talent of distilling a complicated idea into its most simple and accessible essence shines through in his textbooks. His co-authored textbook, The Elements of Statistical Learning (with T. Hastie and J. Friedman), is considered by many to be the “Bible of machine learning” and remains a key reference more than 20 years after its publication. As noted by his PhD advisor Brad Efron, “[Tibshirani] is arguably the most influential applied statistician working today.” In the words of his PhD student Larry Wasserman, “Few statisticians have made a single contribution that has had a lasting impact on either the field of statistics or on fields that use statistical methods … [he] has made a number of such contributions.” In addition to his statistical brilliance, Tibshirani is known for his infectious enthusiasm, and for the mentoring that he has provided for generations of trainees.
Tibshirani received a bachelor of science in mathematics and statistics from the University of Waterloo in 1979, a master of science in statistics from the University of Toronto in 1979, and a PhD in statistics from Stanford University in 1984. He then joined the University of Toronto faculty in 1985, where he stayed until moving to Stanford University in 1998.
Among his many awards, Tibshirani won the Guggenheim Foundation Fellowship (1994), the COPSS Presidents’ Award (1996), the Gold Medal from the Statistical Society of Canada (2012), and the International Statistical Institute’s Founders of Statistics Prize (2021). He is also a fellow of the Royal Society of Canada (2001), a fellow of the Royal Society of the UK (2019), and a member of the US National Academy of Sciences (2012). Tibshirani’s fundamental contributions to methods, theory, and applications of statistics and machine learning make him a highly deserving recipient of the COPSS Distinguished Achievement Award and Lectureship.
Leave a Reply