1st Edition

Regularization, Optimization, Kernels, and Support Vector Machines

    525 Pages
    by Chapman & Hall

    525 Pages 93 B/W Illustrations
    by Chapman & Hall

    Regularization, Optimization, Kernels, and Support Vector Machines offers a snapshot of the current state of the art of large-scale machine learning, providing a single multidisciplinary source for the latest research and advances in regularization, sparsity, compressed sensing, convex and large-scale optimization, kernel methods, and support vector machines. Consisting of 21 chapters authored by leading researchers in machine learning, this comprehensive reference:





    • Covers the relationship between support vector machines (SVMs) and the Lasso


    • Discusses multi-layer SVMs


    • Explores nonparametric feature selection, basis pursuit methods, and robust compressive sensing


    • Describes graph-based regularization methods for single- and multi-task learning


    • Considers regularized methods for dictionary learning and portfolio selection


    • Addresses non-negative matrix factorization


    • Examines low-rank matrix and tensor-based models


    • Presents advanced kernel methods for batch and online machine learning, system identification, domain adaptation, and image processing


    • Tackles large-scale algorithms including conditional gradient methods, (non-convex) proximal techniques, and stochastic gradient descent


    Regularization, Optimization, Kernels, and Support Vector Machines is ideal for researchers in machine learning, pattern recognition, data mining, signal processing, statistical learning, and related areas.

    An Equivalence between the Lasso and Support Vector Machines. Regularized Dictionary Learning. Hybrid Conditional Gradient-Smoothing Algorithms with Applications to Sparse and Low Rank Regularization. Nonconvex Proximal Splitting with Computational Errors. Learning Constrained Task Similarities in Graph-Regularized Multi-Task Learning. The Graph-Guided Group Lasso for Genome-Wide Association Studies. On the Convergence Rate of Stochastic Gradient Descent for Strongly Convex Functions. Detecting Ineffective Features for Nonparametric Regression. Quadratic Basis Pursuit. Robust Compressive Sensing. Regularized Robust Portfolio Estimation. The Why and How of Nonnegative Matrix Factorization. Rank Constrained Optimization Problems in Computer Vision. Low-Rank Tensor Denoising and Recovery via Convex Optimization. Learning Sets and Subspaces. Output Kernel Learning Methods. Kernel Based Identification of Systems with Multiple Outputs Using Nuclear Norm Regularization. Kernel Methods for Image Denoising. Single-Source Domain Adaptation with Target and Conditional Shift. Multi-Layer Support Vector Machines. Online Regression with Kernels.

    Biography

    Johan A.K. Suykens is a professor at Katholieke Universiteit Leuven, Belgium, where he obtained a degree in electro-mechanical engineering and a Ph.D in applied sciences. He has been a visiting postdoctoral researcher at the University of California, Berkeley, USA, and a postdoctoral researcher with the Fonds Wetenschappelijk Onderzoek - Vlaanderen, Belgium. A senior IEEE member, he has co/authored and edited several books; received many prestigious awards; directed, co/organized, and co/chaired numerous international conferences; and served as associate editor for the IEEE Transactions on Circuits and Systems and the IEEE Transactions on Neural Networks.





    Marco Signoretto is currently a visiting lecturer at the Centre for Computational Statistics and Machine Learning (CSML), University College London, UK, in the framework of a postdoctoral fellowship with the Belgian Fund for Scientific Research (FWO). He holds a Ph.D in mathematical engineering from Katholieke Universiteit Leuven, Belgium; a degree in electronic engineering (Laurea Magistralis) from the University of Padova, Italy; and an M.Sc in methods for management of complex systems from the University of Pavia, Italy. His research interests include practical and theoretical aspects of mathematical modeling of structured data, with special focus on multivariate time-series, networks, and dynamical systems. His current work deals with methods based on (convex) optimization, structure-inducing penalties, and spectral regularization.





    Andreas Argyriou has received degrees in computer science from the Massachusetts Institute of Technology, Cambridge, USA, and a Ph.D in computer science from University College London (UCL), UK. The topic of his Ph.D work has been on machine learning methodologies integrating multiple tasks and data sources. He has held postdoctoral and research faculty positions at UCL; Toyota Technological Institute at Chicago, Illinois, USA; and Katholieke Universiteit Leuven, Belgium. He is currently serving an RBUCE-UP fellowship at École Centrale Paris, France. His current interests are in the areas of kernel methods, multitask learning, compressed sensing, and convex optimization methods.