An ADMM Solver for the MKL-$L_{0/1}$-SVM
Machine Learning
2023-04-03 v2 Machine Learning
Abstract
We formulate the Multiple Kernel Learning (abbreviated as MKL) problem for the support vector machine with the infamous -loss function. Some first-order optimality conditions are given and then exploited to develop a fast ADMM solver for the nonconvex and nonsmooth optimization problem. A simple numerical experiment on synthetic planar data shows that our MKL--SVM framework could be promising.
Cite
@article{arxiv.2303.04445,
title = {An ADMM Solver for the MKL-$L_{0/1}$-SVM},
author = {Yijie Shi and Bin Zhu},
journal= {arXiv preprint arXiv:2303.04445},
year = {2023}
}
Comments
8 pages, 3 figures, 2 tables. Submitted to the 62nd IEEE Conference on Decision and Control as a Regular paper, with a shortened version (arXiv version 1) submitted to the 3rd Chinese Conference on Predictive Control and Intelligent Decision (CPCID) as an Extended Abstract