ℓ_p-Norm Multiple Kernel One-Class Fisher Null-Space

2020/08/19 by Shervin Rahimzadeh Arashloo, Arashloo, Shervin Rahimzadeh
Computer Science · Engineering · #Anomaly Detection Techniques and Applications #FOS: Computer and information sciences #Machine Learning (cs.LG) #Machine Learning (stat.ML) #Machine Learning and ELM #Water Systems and Optimization

paper · pdf · doi:10.48550/arxiv.2008.08642

openalex publication_date 2020/08/19 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28

Abstract

The paper addresses the multiple kernel learning (MKL) problem for one-class classification (OCC). For this purpose, based on the Fisher null-space one-class classification principle, we present a multiple kernel learning algorithm where a general ℓ_p-norm constraint (p≥1) on kernel weights is considered. We cast the proposed one-class MKL task as a min-max saddle point Lagrangian optimisation problem and propose an efficient method to solve it. An extension of the proposed one-class MKL approach is also considered where several related one-class MKL tasks are learned jointly by constraining them to share common kernel weights. An extensive assessment of the proposed method on a range of data sets from different application domains confirms its merits against the baseline and several other algorithms.

Citations

• L2 Regularization for Learning Kernels
• The Pascal Visual Object Classes (VOC) Challenge
• Estimating the Support of a High-Dimensional Distribution
• ImageNet classification with deep convolutional neural networks
• Non-convex Optimization for Machine Learning
• Novelty Detection with GAN
• Anomaly Detection using One-Class Neural Networks

Related

• Multi-Task Kernel Null-Space for One-Class Classification
• A simple yet efficient algorithm for multiple kernel learning under elastic-net constraints
• Non-Sparse Regularization for Multiple Kernel Learning
• A Binary Classification Framework for Two-Stage Multiple Kernel Learning
• Multiple Locally Linear Kernel Machines