@InProceedings{Uricar-Franc-CVWW-2012, IS = { zkontrolovano 14 Jan 2013 }, UPDATE = { 2012-03-02 }, author = {U{\v{r}}i{\v{c}}{\'{a}}{\v{r}}, Michal and Franc, Vojt{\v{e}}ch }, language = {English}, title = {Efficient Algorithm for Regularized Risk Minimization}, c_title = {Efektivn{\' i} Algoritmus pro minimalizaci regularizovan{\' e}ho risku}, year = {2012}, pages = {57-64}, booktitle = {CVWW '12: Proceedings of the 17th Computer Vision Winter Workshop}, editor = {Kristan, Matej and Mandeljc, Rok and {\v C}echovin, Luka}, publisher = {Slovenian Pattern Recognition Society}, address = {Ljubljana, Slovenia}, isbn = {978-961-90901-6-9}, book_pages = {181}, month = {February}, day = {1-3}, venue = {Mala Nedelja, Slovenia}, annote = {Many machine learning algorithms lead to solving a convex regularized risk minimization problem. Despite its convexity the problem is often very demanding in practice due to a high number of variables or a complex objective function. The Bundle Method for Risk Minimization (BMRM) is a recently proposed method for minimizing a generic regularized risk. Unlike the approximative methods, the BMRM algorithm comes with convergence guarantees but it is often too slow in practice. We propose a modified variant of the BMRM algorithm which decomposes the objective function into several parts and approximates each part by a separate cutting plane model instead of a single cutting plane model used in the original BMRM. The finer approximation of the objective function can significantly decrease the number of iterations at the expense of higher memory requirements. A preliminary experimental comparison shows promising results.}, keywords = {Machine Learning, Regularized Risk Minimization, Cutting planes}, prestige = {international}, authorship = {50-50}, project = {FP7-ICT-247525 HUMAVIPS, PERG04-GA-2008-239455 SEMISOL}, psurl = {[UricarFranc-CVWW2012.pdf]}, www = {http://cvww2012.vicos.si/}, acceptance_ratio = {0.9}, }