Twin Restricted Kernel Machines for Multiview Classification
A. Quadir, M. Sajid, Mushir Akhtar, M. Tanveer
TL;DR
This work addresses the computational and generalization challenges of kernel-based multiview classification by introducing TMvRKM, a twin restricted kernel machine that integrates visible and hidden variables in a kernelized, multiview framework. It combines early and late fusion through a coupled objective across views and learns two separating hyperplanes using a regularized least squares approach, avoiding expensive QPPs. Empirical results on 27 UCI/KEEL datasets and 27 AwA tasks show TMvRKM achieving superior generalization and robust performance, supported by Friedman and Nemenyi statistical analyses. The approach offers a scalable, RBM-inspired perspective on MVL with strong potential for extensions to clustering and dimensionality reduction.
Abstract
Multi-view learning (MVL) is an emerging field in machine learning that focuses on improving generalization performance by leveraging complementary information from multiple perspectives or views. Various multi-view support vector machine (MvSVM) approaches have been developed, demonstrating significant success. Moreover, these models face challenges in effectively capturing decision boundaries in high-dimensional spaces using the kernel trick. They are also prone to errors and struggle with view inconsistencies, which are common in multi-view datasets. In this work, we introduce the multiview twin restricted kernel machine (TMvRKM), a novel model that integrates the strengths of kernel machines with the multiview framework, addressing key computational and generalization challenges associated with traditional kernel-based approaches. Unlike traditional methods that rely on solving large quadratic programming problems (QPPs), the proposed TMvRKM efficiently determines an optimal separating hyperplane through a regularized least squares approach, enhancing both computational efficiency and classification performance. The primal objective of TMvRKM includes a coupling term designed to balance errors across multiple views effectively. By integrating early and late fusion strategies, TMvRKM leverages the collective information from all views during training while remaining flexible to variations specific to individual views. The proposed TMvRKM model is rigorously tested on UCI, KEEL, and AwA benchmark datasets. Both experimental results and statistical analyses consistently highlight its exceptional generalization performance, outperforming baseline models in every scenario.