On behalf of WNAR and the WNAR Award Committee, we would like to congratulate Robert Tibshirani as the recipient of the 2024 WNAR Outstanding Impact Award and Lectureship. Congratulations, Dr. Tibshirani!
The WNAR of IBS Outstanding impact and Lectureship Award was established in 2021 to recognize an outstanding individual or team, regardless of race, gender, sexual orientation, nationality or citizenship, who has made a significant impact on our society through service and/or research in the development and application of statistical, mathematical, and data science theory and methods in the biomedical or environmental sciences. A significant impact can comprise either a single contribution of extraordinary merit or an outstanding aggregate of contributions that significantly impacts to biosciences and environmental sciences.
Dr. Robert J. Tibshirani is Professor of Biomedical Data Science and Statistics at Stanford University. Dr. Tibshirani’s nomination package clearly demonstrated his highly impactful research contributions to the fields of biometrics and science at large. The LASSO, his most renowned contribution, has ignited an entirely new realm of research and applications from genomics to finance, and is a cornerstone of the modern data science. WNAR is very proud of its outstanding members, represented by Dr. Tibshirani.
Congratulations as well to our other nominees, all of whom were outstanding and highly impressive in their contributions.
As a recipient of the award, Dr. Tibshirani will give a talk in the WNAR Outstanding Impact Award Lecture at the JSM on Monday, August 5, 2024, 2pm. More information about the lecture can be found below.
WNAR members, please plan to submit nomination materials for next year’s award in Fall 2024. We look forward to recognizing our outstanding members with this honor. More information about the award process and upcoming deadlines can be found on the WNAR award website.
WNAR Outstanding Impact Award Lecture at the JSM
Session 1403 Monday August 5, 2:00-3:50PM
Title: Cooperative learning and cooperative components analysis
Speaker: Dr. Robert Tibshirani (Stanford Biomedical Science and Statistics)
Abstract: We propose two methods --- one for supervised learning and the other for unsupervised learning-- both of which make use of an “agreement penalty”.
The first—“Cooperative learning”--- is designed for labelled data with multiple sets of features (“views”). The multiview problem is especially important in biology and medicine, where “-omics” data, such as genomics, proteomics, and radiomics, are measured on a common set of samples. Cooperative learning combines the usual squared-error loss of predictions with an “agreement” penalty to encourage the predictions from different data views to agree. By varying the weight of the agreement penalty, we get a continuum of solutions that include the well-known early and late fusion approaches.
Cooperative components analysis (“CoCA”) is a new method for unsupervised multi-view analysis. It identifies the component that simultaneously captures significant within-view variance and exhibits strong cross-view correlation. The challenge of integrating multi-view data is particularly important in biology and medicine, where various types of “-omic” data, ranging from genomics to proteomics, are collected from the same set of samples.
CoCA combines a reconstruction error loss to preserve information within data views and an “agreement penalty” to encourage alignment across views. By balancing the trade-off between these two key components in the objective, CoCA encompasses both principal component analysis and canonical correlation analysis as special cases.
This is joint work with DY Ding, S Li, B Narasimhan, Alden Green and Min Sun.
