The development of Multiple Kernel Techniques has become of particular interest for machine learning researchers in Computer Vision topics like image processing, object classification, and object state recognition. Sparsity-inducing norms along with non-sparse formulations promote different degrees of sparsity at the kernel coefficient level, at the same time permitting non-sparse combination within each individual kernel. This makes MKL models very suitable for different problems, allowing adequate selection of the regularizer according to different norms and the nature of the problem. We formulate and discuss MKL regularizations and optimization approaches, as well asdemonstrate MKL effectiveness compared to the stateof-the-art SVM models using a Computer Vision Recognition problem.
Alioscha-Perez, M, Sahli, H, Gonzalez, I, Taboada-Crispi, A, Sidorov, G (ed.) & Cortes, U (ed.) 2012, 'Sparse and Non-sparse Multiple Kernel Learning for Recognition', Computacion y Sistemas - An International journal of computing science and applications, vol. 16, no. Machine Learning and Pattern Recognition, pp. 167-174. <http://www.cic.ipn.mx/sitioCIC/images/revista/vol16-2/art03.pdf>
Alioscha-Perez, M., Sahli, H., Gonzalez, I., Taboada-Crispi, A., Sidorov, G. (Ed.), & Cortes, U. (Ed.) (2012). Sparse and Non-sparse Multiple Kernel Learning for Recognition. Computacion y Sistemas - An International journal of computing science and applications, 16(Machine Learning and Pattern Recognition), 167-174. http://www.cic.ipn.mx/sitioCIC/images/revista/vol16-2/art03.pdf
@article{334a74b210f54e1da8d323329a36cea1,
title = "Sparse and Non-sparse Multiple Kernel Learning for Recognition",
abstract = "The development of Multiple Kernel Techniques has become of particular interest for machine learning researchers in Computer Vision topics like image processing, object classification, and object state recognition. Sparsity-inducing norms along with non-sparse formulations promote different degrees of sparsity at the kernel coefficient level, at the same time permitting non-sparse combination within each individual kernel. This makes MKL models very suitable for different problems, allowing adequate selection of the regularizer according to different norms and the nature of the problem. We formulate and discuss MKL regularizations and optimization approaches, as well asdemonstrate MKL effectiveness compared to the stateof-the-art SVM models using a Computer Vision Recognition problem.",
keywords = "Multiple Kernel Learning, object state recognition, norm regularizers, analytical updates, cutting plane method, Newton's method",
author = "Mitchel Alioscha-Perez and Hichem Sahli and Isabel Gonzalez and Alberto Taboada-Crispi and Grigori Sidorov and Ulises Cortes",
note = "Grigori Sidorov; Ulises Cortes",
year = "2012",
language = "English",
volume = "16",
pages = "167--174",
journal = "Computacion y Sistemas - An International journal of computing science and applications",
issn = "1405-5546",
publisher = "Centro de Investigacion en Computacion (CIC) del Instituto Politecnico Nacional (IPN)",
number = "Machine Learning and Pattern Recognition",
}