Incorporate Hashing with Multi-view Learning

Abstract

Multi-view learning concentrates on multiple distinct feature sets mainly following either the consensus principle or the complementary principle. SVM-2K, as a typical SVM-based multi-view learning model, extends SVM for multi-view learning by combining Kernel Canonical Correlation Analysis (KCCA). However, SVM-2K cannot fully unleash the power of the complementary information among different feature views. Recently, a framework of Predictable Dual-View Hashing (PDH) has been proposed to process two-view data sharing the complementary information. Motivated by PDH, we propose a novel strategy: Incorporate Hashing with Multi-view Learning (IHMVL). The specific implementation of IHMVL is to incorporate the PDH algorithm with SVM-2K for two-view classification, which suffices to both the consensus and complementary principles. This is the first work that extends hashing to multi-view learning. Experimental results on the synthetic data sets confirm the effectiveness of the proposed method.

DOI: 10.1109/ICDMW.2016.0126

5 Figures and Tables

Cite this paper

@article{Tang2016IncorporateHW, title={Incorporate Hashing with Multi-view Learning}, author={Jingjing Tang and Dewei Li}, journal={2016 IEEE 16th International Conference on Data Mining Workshops (ICDMW)}, year={2016}, pages={853-859} }