Tuesday, July 8

MS49
Tensor Computations For Large-Scale Data Applications

4:00 PM - 6:00 PM
Room: Royal Palm 2

The structure of large-scale scientific data is of an order higher than that of a matrix, especially when comparing and integrating different types of data from different studies. Unfolded into a matrix, much of the information in a data tensor might be lost. Tensor generalizations of matrix decompositions are not obvious, because such properties as orthogonality and diagonality are not guaranteed. In this minisymposium we will present recent studies of the mathematics and computations of several tensor decompositions, and illustrate their applications in such diverse fields as chemistry and genetics.

Organizer: Orly Alter
University of Texas at Austin
Sri Priya Ponnapalli
University of Texas at Austin

4:00-4:25 Large-Scale Tensor Computations and Applications to Data Mining

Tamara G. Kolda, Sandia National Laboratories

4:30-4:55 Multilinear Algebra Based Fitting of a Sum of Exponentials to Oversampled Data

Lieven De Lathauwer, Katholieke Universiteit Leuven, Belgium; Jean-Michel Papy, Flanders Mechatronics Technology Centre, Belgium; Sabine Van Huffel, Katholieke Universiteit Leuven, Belgium

5:00-5:25 Multilinear Algebra Computations in Quantum Chemistry

Charles Van Loan, Cornell University

5:30-5:55 A Novel Higher-Order Generalized Singular Value Decomposition for Comparative Analysis of DNA Microarray Data From Different Organisms

Sri Priya Ponnapalli, University of Texas at Austin; Michael A. Saunders and Gene H. Golub, Stanford University; Orly Alter, University of Texas at Austin