Adversarial Autoencoders with Constant-Curvature Latent Manifolds

Abstract

Constant-curvature Riemannian manifolds (CCMs) have been shown to be ideal embedding spaces in many application domains, as their non-Euclidean geometry can naturally account for some relevant properties of data, like hierarchy and circularity. In this work, we introduce the CCM adversarial autoencoder (CCM-AAE), a probabilistic generative model trained to represent a data distribution on a CCM. Our method works by matching the aggregated posterior of the CCM-AAE with a probability distribution defined on a CCM, so that the encoder implicitly learns to represent data on the CCM to fool the discriminator network. The geometric constraint is also explicitly imposed by jointly training the CCM-AAE to maximise the membership degree of the embeddings to the CCM. While a few works in recent literature make use of either hyperspherical or hyperbolic manifolds for different learning tasks, ours is the first unified framework to seamlessly deal with CCMs of different curvatures. We show the effectiveness of our model on three different datasets characterised by non-trivial geometry: semi-supervised classification on MNIST, link prediction on two popular citation datasets, and graph-based molecule generation using the QM9 chemical database. Results show that our method improves upon other autoencoders based on Euclidean and non-Euclidean geometries on all tasks taken into account.

arXiv commentSubmitted to Applied Soft Computing

References in corpus (10)

• Auto-Encoding Variational Bayes
• Geometric deep learning: going beyond Euclidean data
• Automatic chemical design using a data-driven continuous representation of molecules
• Semi-Supervised Learning with Deep Generative Models
• Variational Graph Auto-Encoders
• MolGAN: An implicit generative model for small molecular graphs
• Adversarial Autoencoders
• Hyperbolic Attention Networks
• Change Detection in Graph Streams by Learning Graph Embeddings on Constant-Curvature Manifolds
• Autoencoding topology

Cited by in corpus (8)

• Deep learning for molecular design - a review of the state of the art
• A Wrapped Normal Distribution on Hyperbolic Space for Gradient-Based Learning
• Hyperbolic Deep Neural Networks: A Survey
• Latent Variable Modelling with Hyperbolic Normalizing Flows
• Manifolds for Unsupervised Visual Anomaly Detection
• APo-VAE: Text Generation in Hyperbolic Space
• Randomized geometric tools for anomaly detection in stock markets
• Atlas Based Representation and Metric Learning on Manifolds