TY  - CHAP
U1  - Konferenzveröffentlichung
A1  - Ram, Raju
A1  - Müller, Sabine
A1  - Pfreundt, Franz-Josef
A1  - Gauger, Nicolas R.
A1  - Keuper, Janis
T1  - Scalable Hyperparameter Optimization with Lazy Gaussian Processes
T2  - Proceedings of MLHPC 2019: 5th Workshop on Machine Learning in HPC Environments
N2  - Most machine learning methods require careful selection of hyper-parameters in order to train a high performing model with good generalization abilities. Hence, several automatic selection algorithms have been introduced to overcome tedious manual (try and error) tuning of these parameters. Due to its very high sample efficiency, Bayesian Optimization over a Gaussian Processes modeling of the parameter space has become the method of choice. Unfortunately, this approach suffers from a cubic compute complexity due to underlying Cholesky factorization, which makes it very hard to be scaled beyond a small number of sampling steps. In this paper, we present a novel, highly accurate approximation of the underlying Gaussian Process. Reducing its computational complexity from cubic to quadratic allows an efficient strong scaling of Bayesian Optimization while outperforming the previous approach regarding optimization accuracy. First experiments show speedups of a factor of 162 in single node and further speed up by a factor of 5 in a parallel environment.
Y1  - 2019
SN  - 978-1-7281-5985-0 (Online)
SB  - 978-1-7281-5985-0 (Online)
SN  - 978-1-7281-5986-7 (Print on Demand)
SB  - 978-1-7281-5986-7 (Print on Demand)
U6  - https://doi.org/10.1109/MLHPC49564.2019.00011
DO  - https://doi.org/10.1109/MLHPC49564.2019.00011
N1  - Conference held in conjunction with SC19: The International Conference for High Performance Computing, Networking, Storage and Analysis, Denver, Colorado, November 17-22, 2019
SP  - 56
EP  - 65
PB  - IEEE
ER  -