Mathematics of Information
As modern society creates, processes, and interprets data faster than ever before and then uses it for a variety of tasks, the need for rigorous understanding of the algorithms and mechanisms used to process this data is ever necessary. The mathematics of information group at UBC studies the theoretical underpinnings of modern data science with the aim of designing new algorithms, and understanding their mathematical mechanisms.
Areas of interest in the group include machine learning theory and applications, large-scale optimization, compressed sensing, deep learning theory, high-dimensional inference, matrix completion, ergodic theory, symbolic dynamics, causality, tensor methods, algebraic statistics, optimal transport theory and applications.
INFO FOR PROSPECTIVE STUDENTS
Students are encouraged to take topics courses offered by the core faculty in the group. Besides the Mathematics of Information seminar, a variety of other seminars (which can be found on the IAM events website) are relevant, for example, the IAM colloquium and the probability seminar.
|Michael Friedlander||optimization, variational and nonsmooth analysis, computational mathematics, applications in statistical learning and signal processing|
|Yaniv Plan||Compressed sensing, matrix completion, and generally low-dimensional structures in high-dimensional space; theory of deep learning and generally learning theory; probability in high dimensions, typically as motivated by the preceding areas.|
|Elina Robeva||Probabilistic graphical models, causality, algebraic statistics, tensor decompositions, super-resolution imaging, density estimation, applied algebraic geometry|
|Geoffrey Schiebinger||optimization, single cell analysis, applied probability|
|Ozgur Yilmaz||Mathematical problems related to analog-to-digital conversion, blind source separation, sparse approximations and compressed sensing, and applications in seismic signal processing.|