Multiple canonical correlation analysis of tensor data
Prof. David Degras
(University of Massachusetts Boston)
Abstract:
Canonical correlation analysis (CCA) is a celebrated statistical technique for finding linear combinations of variables that are maximally correlated between two datasets. By reducing the dimension of data, CCA facilitates understanding relationships between groups of variables. It also provides canonical variables as potentially useful features in machine learning tasks, e.g., classification, clustering, and regression. Since the 1970s numerous extensions of CCA have been developed to handle multiple datasets (MCCA), high-dimensional data, nonlinearity patterns, and more. In recent years, the wide availability of multiple data sources has led to a regain of interest in multiblock analysis methods -including MCCA- for data integration and fusion. Related to this, new applications in biomedical research, computer vision, and remote sensing have prompted efforts to extend MCCA to tensor data, e.g., 2D/3D images and video sequences. In this talk I will present recent advances in MCCA and related techniques for tensor data, including some of my ongoing work on the subject. Focusing on computations, I will discuss challenges in initializing optimization algorithms, assessing the quality of solutions, determining higher-order canonical components, and processing large datasets. Time permitting, I will show numerical experiments and applications to the multimodal integration of brain imaging data.
Biographical Sketch:
Professor Degras is a French statistician who teaches and conducts research in the Mathematics Department.
Date and time:
Saturday, February 26, 2022
12:45 PM—2:30 PM
Zoom Meeting Link:
https://umassboston.zoom.us/j/92049921824