Natural gradient via optimal transport
From MaRDI portal
Publication:1713655
DOI10.1007/s41884-018-0015-3zbMath1409.62022arXiv1803.07033OpenAlexW2963645788WikidataQ126033651 ScholiaQ126033651MaRDI QIDQ1713655
Publication date: 28 January 2019
Published in: Information Geometry (Search for Journal in Brave)
Full work available at URL: https://arxiv.org/abs/1803.07033
information geometrymachine learningoptimal transportdisplacement convexitymanifolds of probability distributionsWasserstein statistical manifold
Probability distributions: general theory (60E05) Learning and adaptive systems in artificial intelligence (68T05) Applications of global differential geometry to the sciences (53C80) Statistical aspects of information-theoretic topics (62B10)
Related Items (25)
Analysis of Asymptotic Escape of Strict Saddle Sets in Manifold Optimization ⋮ Sparse optimization on measures with over-parameterized gradient descent ⋮ When optimal transport meets information geometry ⋮ Pseudo-Riemannian geometry encodes information geometry in optimal transport ⋮ Transport information geometry: Riemannian calculus on probability simplex ⋮ Information geometry of physics-informed statistical manifolds and its use in data assimilation ⋮ Quantum statistical learning via quantum Wasserstein natural gradient ⋮ Neural Parametric Fokker--Planck Equation ⋮ High order spatial discretization for variational time implicit schemes: Wasserstein gradient flows and reaction-diffusion systems ⋮ Invariance properties of the natural gradient in overparametrised systems ⋮ Efficient Natural Gradient Descent Methods for Large-Scale PDE-Based Optimization Problems ⋮ Interacting Langevin Diffusions: Gradient Structure and Ensemble Kalman Sampler ⋮ Affine statistical bundle modeled on a Gaussian Orlicz-Sobolev space ⋮ Geometry and convergence of natural policy gradient methods ⋮ Wasserstein information matrix ⋮ Mirror descent algorithms for minimizing interacting free energy ⋮ Information Geometry of Smooth Densities on the Gaussian Space: Poincaré Inequalities ⋮ Optimal transport natural gradient for statistical manifolds with continuous sample space ⋮ Ricci curvature for parametric statistics via optimal transport ⋮ Natural gradient for combined loss using wavelets ⋮ Transport information Bregman divergences ⋮ Fisher information regularization schemes for Wasserstein gradient flows ⋮ Hessian transport gradient flows ⋮ Hessian metric via transport information geometry ⋮ Wasserstein proximal of GANs
Uses Software
Cites Work
- Unnamed Item
- Unnamed Item
- Unnamed Item
- Natural gradient flow in the mixture geometry of a discrete exponential family
- On the Fisher metric of conditional probability polytopes
- Wasserstein geometry of Gaussian measures
- Gradient flows of the entropy for finite Markov chains
- Information geometry and its applications
- Some geometric calculations on Wasserstein space
- Constrained steepest descent in the 2-Wasserstein metric
- Computations of optimal transport distance with Fisher information regularization
- Entropy dissipation of Fokker-Planck equations on graphs
- Information geometry of Wasserstein divergence
- Information geometry connecting Wasserstein distance and Kullback-Leibler divergence via the entropy-relaxed transportation problem
- A computational fluid mechanics solution to the Monge-Kantorovich mass transfer problem
- Ricci curvature for parametric statistics via optimal transport
- Fokker-Planck equations for a free energy functional or Markov process on a graph
- Geometry of matrix decompositions seen through optimal transport and information geometry
- THE GEOMETRY OF DISSIPATIVE EVOLUTION EQUATIONS: THE POROUS MEDIUM EQUATION
- A gradient structure for reaction–diffusion systems and for energy-drift-diffusion systems
- Axiomatic Geometry of Conditional Models
- An Extended Cencov Characterization of the Information Metric
- The Density Manifold and Configuration Space Quantization
- Geodesics of minimal length in the set of probability measures on graphs
- Towards the geometry of estimation of distribution algorithms based on the exponential family
- Information Geometry
- Optimal Transport
This page was built for publication: Natural gradient via optimal transport
