Corpus ID: 58028932

Hierarchical Representations with Poincaré Variational Auto-Encoders

@article{Mathieu2019HierarchicalRW,
  title={Hierarchical Representations with Poincar{\'e} Variational Auto-Encoders},
  author={Emile Mathieu and Charline Le Lan and Chris J. Maddison and Ryota Tomioka and Y. Teh},
  journal={ArXiv},
  year={2019},
  volume={abs/1901.06033}
}
  • Emile Mathieu, Charline Le Lan, +2 authors Y. Teh
  • Published 2019
  • Mathematics, Computer Science
  • ArXiv
  • The Variational Auto-Encoder (VAE) model is a popular method to learn at once a generative model and embeddings for data living in a high-dimensional space. [...] Key Method We therefore endow VAE with a Poincar\'e ball model of hyperbolic geometry and derive the necessary methods to work with two main Gaussian generalisations on that space. We empirically show better generalisation to unseen data than the Euclidean counterpart, and can qualitatively and quantitatively better recover hierarchical structures.Expand Abstract
    14 Citations

    Figures, Tables, and Topics from this paper.

    Mixed-curvature Variational Autoencoders
    • 9
    • PDF
    Diffusion Variational Autoencoders
    • 13
    • PDF
    Poincaré Wasserstein Autoencoder
    • 16
    • PDF
    Learning Flat Latent Manifolds with VAEs
    • 1
    • PDF
    Neural Manifold Ordinary Differential Equations
    • 2
    • Highly Influenced
    • PDF
    Constant Curvature Graph Convolutional Networks
    • 9
    • PDF

    References

    SHOWING 1-10 OF 55 REFERENCES
    Hyperspherical Variational Auto-Encoders
    • 135
    • PDF
    Explorations in Homeomorphic Variational Auto-Encoding
    • 37
    • PDF
    Poincaré Embeddings for Learning Hierarchical Representations
    • 400
    • PDF
    Poincaré Wasserstein Autoencoder
    • 16
    • PDF
    Adversarial Autoencoders with Constant-Curvature Latent Manifolds
    • 12
    • Highly Influential
    • PDF
    Wasserstein Auto-Encoders
    • 355
    • PDF
    Hamiltonian Variational Auto-Encoder
    • 22
    • PDF
    Learning Continuous Hierarchies in the Lorentz Model of Hyperbolic Geometry
    • 118
    • PDF
    Hyperbolic Neural Networks
    • 82
    • Highly Influential
    • PDF
    Poincaré GloVe: Hyperbolic Word Embeddings
    • 68
    • PDF