Multi-View Oriented GPLVM: Expressiveness and Efficiency

2025-02-12

AI Summary

To address the dual bottlenecks of limited kernel expressiveness and inefficient variational inference in multi-view Gaussian process latent variable models (MV-GPLVM), this paper introduces a spectral-density-driven kernel construction paradigm. We establish an explicit duality between spectral densities and kernel functions, and propose a highly expressive Next-Gen Spectral Mixture (NGSM) kernel. By integrating bivariate Gaussian mixture spectral density modeling, random Fourier feature approximation, and reparameterization, we enable scalable variational inference. The proposed method is embedded within a variational autoencoder framework and achieves state-of-the-art performance on multiple multi-view benchmark datasets—outperforming existing MV-GPLVMs and deep multi-view models. It learns unified latent representations that are more robust, interpretable, and generalizable.

Technology Category

• Machine Learning: Multi-instance/Multi-view Learning
• Computer Vision: Multi-modal Vision
• Search and Optimization: Mixed Discrete/Continuous Search

Application Category

• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• User Modeling, Personalization and Recommendation: Explainable and interpretable methods for personalization
• Graph Algorithms and Modeling for the Web: Graph embeddings and representation learning for Web-related graphs

Abstract

The multi-view Gaussian process latent variable model (MV-GPLVM) aims to learn a unified representation from multi-view data but is hindered by challenges such as limited kernel expressiveness and low computational efficiency. To overcome these issues, we first introduce a new duality between the spectral density and the kernel function. By modeling the spectral density with a bivariate Gaussian mixture, we then derive a generic and expressive kernel termed Next-Gen Spectral Mixture (NG-SM) for MV-GPLVMs. To address the inherent computational inefficiency of the NG-SM kernel, we propose a random Fourier feature approximation. Combined with a tailored reparameterization trick, this approximation enables scalable variational inference for both the model and the unified latent representations. Numerical evaluations across a diverse range of multi-view datasets demonstrate that our proposed method consistently outperforms state-of-the-art models in learning meaningful latent representations.

Problem

• Enhances kernel expressiveness in MV-GPLVM
• Improves computational efficiency with approximation
• Outperforms state-of-the-art in latent representations

Innovation

• Introduces Next-Gen Spectral Mixture kernel
• Uses random Fourier feature approximation
• Implements scalable variational inference