Hyperbolic Generative Adversarial Network
arXiv:2102.05567
Abstract
Recently, Hyperbolic Spaces in the context of Non-Euclidean Deep Learning have gained popularity because of their ability to represent hierarchical data. We propose that it is possible to take advantage of the hierarchical characteristic present in the images by using hyperbolic neural networks in a GAN architecture. In this study, different configurations using fully connected hyperbolic layers in the GAN, CGAN, and WGAN are tested, in what we call the HGAN, HCGAN, and HWGAN, respectively. The results are measured using the Inception Score (IS) and the Fréchet Inception Distance (FID) on the MNIST dataset. Depending on the configuration and space curvature, better results are achieved for each proposed hyperbolic versions than their euclidean counterpart.
References in corpus (9)
- Adam: A Method for Stochastic Optimization
- Conditional Generative Adversarial Nets
- GANs Trained by a Two Time-Scale Update Rule Converge to a Local Nash Equilibrium
- Geometric deep learning: going beyond Euclidean data
- Hyperbolic Geometry of Complex Networks
- Poincaré GloVe: Hyperbolic Word Embeddings
- A Wrapped Normal Distribution on Hyperbolic Space for Gradient-Based Learning
- Continuous Hierarchical Representations with Poincaré Variational Auto-Encoders
- APo-VAE: Text Generation in Hyperbolic Space