A Quaternion-Valued Variational Autoencoder

Abstract

Deep probabilistic generative models have achieved incredible success in many fields of application. Among such models, variational autoencoders (VAEs) have proved their ability in modeling a generative process by learning a latent representation of the input. In this paper, we propose a novel VAE defined in the quaternion domain, which exploits the properties of quaternion algebra to improve performance while significantly reducing the number of parameters required by the network. The success of the proposed quaternion VAE with respect to traditional VAEs relies on the ability to leverage the internal relations between quaternion-valued input features and on the properties of second-order statistics which allow to define the latent variables in the augmented quaternion domain. In order to show the advantages due to such properties, we define a plain convolutional VAE in the quaternion domain and we evaluate its performance with respect to its real-valued counterpart on the CelebA face dataset.

Accepted for publication at the 2021 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)

References in corpus (10)

• Tutorial on Variational Autoencoders
• NVAE: A Deep Hierarchical Variational Autoencoder
• BIVA: A Very Deep Hierarchy of Latent Variables for Generative Modeling
• Variational Bayesian Inference with Stochastic Search
• Quaternion Convolutional Neural Networks for Heterogeneous Image Processing
• Quaternion Recurrent Neural Networks
• Quaternion Convolutional Neural Networks for Detection and Localization of 3D Sound Events
• Compressing deep quaternion neural networks with targeted regularization
• PixelVAE++: Improved PixelVAE with Discrete Prior
• Undirected Graphical Models as Approximate Posteriors