Normalizing Flows on Tori and Spheres

Abstract

Normalizing flows are a powerful tool for building expressive distributions in high dimensions. So far, most of the literature has concentrated on learning flows on Euclidean spaces. Some problems however, such as those involving angles, are defined on spaces with more complex geometries, such as tori or spheres. In this paper, we propose and compare expressive and numerically stable flows on such spaces. Our flows are built recursively on the dimension of the space, starting from flows on circles, closed intervals or spheres.

Accepted to the International Conference on Machine Learning (ICML) 2020

References in corpus (8)

• Variational Inference with Normalizing Flows
• Normalizing Flows for Probabilistic Modeling and Inference
• Generalized Euler Angle Paramterization for SU(N)
• Generalized Euler Angle Parameterization for U(N) with Applications to SU(N) Coset Volume Measures
• Explorations in Homeomorphic Variational Auto-Encoding
• Continuous Hierarchical Representations with Poincaré Variational Auto-Encoders
• Normalizing Flows on Riemannian Manifolds
• A Wasserstein Minimum Velocity Approach to Learning Unnormalized Models

Cited by in corpus (33)

• Normalizing Flows: An Introduction and Review of Current Methods
• Equivariant flow-based sampling for lattice gauge theory
• Efficient Modelling of Trivializing Maps for Lattice ${\varphi }^4$ Theory Using Normalizing Flows: A First Look at Scalability
• Quasi Anomalous Knowledge: Searching for new physics with embedded knowledge
• SoftFlow: Probabilistic Framework for Normalizing Flow on Manifolds
• Normalizing flows for atomic solids
• Via Machinae: Searching for Stellar Streams using Unsupervised Machine Learning
• normflows: A PyTorch Package for Normalizing Flows
• Riemannian Continuous Normalizing Flows
• Neural Manifold Ordinary Differential Equations
• Returning CP-Observables to The Frames They Belong
• The Power Spherical distribution
• Introduction to Normalizing Flows for Lattice Field Theory
• Neural Ordinary Differential Equations on Manifolds
• RG-Flow: A hierarchical and explainable flow model based on renormalization group and sparse prior
• Manifold GPLVMs for discovering non-Euclidean latent structure in neural data
• Deep Learning Hamiltonian Monte Carlo
• Efficient mapping of phase diagrams with conditional Boltzmann Generators
• Latent Variable Modelling with Hyperbolic Normalizing Flows
• ChartPointFlow for Topology-Aware 3D Point Cloud Generation
• Computing Absolute Free Energy with Deep Generative Models
• Adaptive Monte Carlo augmented with normalizing flows
• On Computational Poisson Geometry II: Numerical Methods
• Principled Interpolation in Normalizing Flows
• Tractable Density Estimation on Learned Manifolds with Conformal Embedding Flows
• NF-iSAM: Incremental Smoothing and Mapping via Normalizing Flows
• Ordering Dimensions with Nested Dropout Normalizing Flows
• Variational Determinant Estimation with Spherical Normalizing Flows
• Spherical Sliced-Wasserstein
• Manifold Density Estimation via Generalized Dequantization
• Continuous normalizing flows on manifolds
• Equivariant Manifold Flows
• Minimax Analysis for Inverse Risk in Nonparametric Planer Invertible Regression