When Matrix Factorization meets Machine Learning:
An NlogN Parallel Fast Direct Solver for Kernel Matrices by Chenhan D. Yu, William B. March, George Biros
Kernel matrices appear in machine learning and non-parametric statistics. Given N points in d dimensions and a kernel function that requires O(d) work to evaluate, we present an O(dNlogN)-work algorithm for the approximate factorization of a regularized kernel matrix, a common computational bottleneck in the training phase of a learning task. With this factorization, solving a linear system with a kernel matrix can be done with O(NlogN) work. Our algorithm only requires kernel evaluations and does not require that the kernel matrix admits an efficient global low rank approximation. Instead our factorization only assumes low-rank properties for the off-diagonal blocks under an appropriate row and column ordering. We also present a hybrid method that, when the factorization is prohibitively expensive, combines a partial factorization with iterative methods. As a highlight, we are able to approximately factorize a dense 11M×11M kernel matrix in 2 minutes on 3,072 x86 "Haswell" cores and a 4.5M×4.5M matrix in 1 minute using 4,352 "Knights Landing" cores.
Liked this entry ? subscribe to Nuit Blanche's feed, there's more where that came from. You can also subscribe to Nuit Blanche by Email, explore the Big Picture in Compressive Sensing or the Matrix Factorization Jungle and join the conversations on compressive sensing, advanced matrix factorization and calibration issues on Linkedin.
No comments:
Post a Comment