Multiview learning with variational mixtures of Gaussian processes

作者：

Highlights：

• We present a framework of multiview learning for mixtures of Gaussian processes.

• MvMGPs maximize the posterior distribution of latent variables in each view.

• MvMGPs regularize the objective function to learn parameters of different views.

• The proposed model outperforms multiple baselines for classification tasks.

摘要

•We present a framework of multiview learning for mixtures of Gaussian processes.•MvMGPs maximize the posterior distribution of latent variables in each view.•MvMGPs regularize the objective function to learn parameters of different views.•The proposed model outperforms multiple baselines for classification tasks.

论文关键词：Mixtures of Gaussian processes,Co-regularization,Multiview learning,Variational inference,Supervised learning

论文评审过程：Received 15 November 2019, Revised 23 April 2020, Accepted 29 April 2020, Available online 5 May 2020, Version of Record 14 May 2020.

论文官网地址：https://doi.org/10.1016/j.knosys.2020.105990