Optimization for Classical Machine Learning Problems on the GPU
Abstract
Constrained optimization problems arise frequently in classical machine learning. There exist frameworks addressing constrained optimization, for instance, CVXPY and GENO. However, in contrast to deep learning frameworks, GPU support is limited. Here, we extend the GENO framework to also solve constrained optimization problems on the GPU. The framework allows the user to specify constrained optimization problems in an easy-to-read modeling language. A solver is then automatically generated from this specification. When run on the GPU, the solver outperforms state-of-the-art approaches like CVXPY combined with a GPU-accelerated solver such as cuOSQP or SCS by a few orders of magnitude.
Cite
Text
Laue et al. "Optimization for Classical Machine Learning Problems on the GPU." AAAI Conference on Artificial Intelligence, 2022. doi:10.1609/AAAI.V36I7.20692Markdown
[Laue et al. "Optimization for Classical Machine Learning Problems on the GPU." AAAI Conference on Artificial Intelligence, 2022.](https://mlanthology.org/aaai/2022/laue2022aaai-optimization/) doi:10.1609/AAAI.V36I7.20692BibTeX
@inproceedings{laue2022aaai-optimization,
title = {{Optimization for Classical Machine Learning Problems on the GPU}},
author = {Laue, Sören and Blacher, Mark and Giesen, Joachim},
booktitle = {AAAI Conference on Artificial Intelligence},
year = {2022},
pages = {7300-7308},
doi = {10.1609/AAAI.V36I7.20692},
url = {https://mlanthology.org/aaai/2022/laue2022aaai-optimization/}
}