The focus of this workshop will be on recent developments in randomized linear algebra, with an emphasis on how algorithmic improvements from the theory of algorithms interact with statistical, optimization, inference, and related perspectives. One focus area of the workshop will be the broad use of sketching techniques developed in the data stream literature for solving optimization problems in linear and multi-linear algebra. The workshop will also consider the impact of theoretical developments in randomized linear algebra on (i) numerical analysis as a method for constructing preconditioners; (ii) applications as a principled feature selection method; and (iii) implementations as a way to avoid communication rather than computation. Another goal of this workshop is thus to bridge the theory-practice gap by trying to understand the needs of practitioners when working on real datasets.
Randomized Numerical Linear Algebra and Applications
Dates: September 24-27, 2018
Location: Calvin Hall, Simons Institute, UC Berkeley
This training workshop is part of the Simons Institute's Foundations of Data Science Program being offered during the Fall 2018 semester. BIDS Senior Fellow Michael Mahoney is co-organizing and presenting.
Speaker(s)
Michael Mahoney
Michael Mahoney works on algorithmic and statistical aspects of modern large-scale data analysis. Much of his recent research has focused on large-scale machine learning including randomized matrix algorithms and randomized numerical linear algebra; geometric network analysis tools for structure extraction in large informatics graphs; scalable implicit regularization methods; and applications in genetics, astronomy, medical imaging, social network analysis, and internet data analysis.