Alex Townsend
Associate Professor
Research Focus
Numerical analysis, scientific computing, and deep learning
I am interested in the study and development of numerical algorithms in applied mathematics. I mainly work in the following three areas: (1) Novel spectral methods for the solution of differential equations, (2) Lowrank techniques, and (3) Theoretical aspects of deep learning.
Publications

Dense networks that do not synchronize and sparse ones that do (with M. Stillman and S. H. Strogatz), Chaos, 30 (2020), 083142.

Fast algorithms using orthogonal polynomials (with S. Olver and R. M. Slevinsky), Acta Numerica, 29 (2020), pp. 573699.

Stable extrapolation of analytic functions (with L. Demanet), Foundations of Computational Mathematics, 19 (2019), pp. 297331.

Bounds on the singular values of matrices with displacement structure (with B. Beckermann), SIAM Review, 61 (2019), pp. 319344.

Why are big data matrices approximately low rank? (with M. Udell), SIAM Journal on Mathematics of Data Science, 1 (2019), pp. 144160.
In the news
 New center merges math, AI to push frontiers of science
 Arts and Sciences faculty featured on Academic Minute
 Weiss teaching award honors eight exceptional faculty
 Rational neural network advances machinehuman discovery
 Six A&S professors named 2022 Simons fellows
 Computing with rational functions
 Eleven assistant professors win NSF earlycareer awards
 30 Arts & Sciences faculty honored with endowed professorships
 Grants create engagement opportunities for students
 Grants create communityengaged opportunities for students
 Math professor mentors winner of science talent search