Quantum statistical learning via quantum Wasserstein natural gradient

DOI10.1007/S10955-020-02682-1zbMATH Open1460.81008arXiv2008.11135OpenAlexW3081367176WikidataQ125601851 ScholiaQ125601851MaRDI QIDQ2658080FDOQ2658080

Authors: Simon Becker, Wuchen Li 

Publication date: 18 March 2021

Published in: Journal of Statistical Physics (Search for Journal in Brave)

Abstract: In this article, we introduce a new approach towards the statistical learning problem $operatorname{argmin}_{ho(heta)inmathcal{P}_{heta}}{W}_Q^2(h{o}_{star},ho(heta))$ to approximate a target quantum state $h{o}_{star}$ by a set of parametrized quantum states $ho(heta)$ in a quantum $L^2$-Wasserstein metric. We solve this estimation problem by considering Wasserstein natural gradient flows for density operators on finite-dimensional $C^{\ast }$ algebras. For continuous parametric models of density operators, we pull back the quantum Wasserstein metric such that the parameter space becomes a Riemannian manifold with quantum Wasserstein information matrix. Using a quantum analogue of the Benamou-Brenier formula, we derive a natural gradient flow on the parameter space. We also discuss certain continuous-variable quantum states by studying the transport of the associated Wigner probability distributions.

Full work available at URL: https://arxiv.org/abs/2008.11135

Recommendations

• Optimal transport natural gradient for statistical manifolds with continuous sample space
• On quantum versions of the classical Wasserstein distance
• Natural gradient via optimal transport
• Quantum optimal transport with quantum channels
• Wasserstein geometry of quantum states and optimal transport of matrix-valued measures

zbMATH Keywords

quantum state estimationquantum transport information geometryquantum Wasserstein information matrixquantum Wasserstein natural gradientquantum Schrödinger bridge problem

Mathematics Subject Classification ID

Learning and adaptive systems in artificial intelligence (68T05) General theory of (C^*)-algebras (46L05) Quantum information, communication, networks (quantum-theoretic aspects) (81P45) Quantum state spaces, operational and probabilistic concepts (81P16) Quantum state estimation, approximate cloning (81P50) Differential geometric aspects of statistical manifolds and information geometry (53B12)

Cites Work

• Optimal Transport
• A computational fluid mechanics solution to the Monge-Kantorovich mass transfer problem
• THE GEOMETRY OF DISSIPATIVE EVOLUTION EQUATIONS: THE POROUS MEDIUM EQUATION
• On the mean field and classical limits of quantum mechanics
• On the relation between optimal transport and Schrödinger bridges: a stochastic control viewpoint
• On Matrix-Valued Monge–Kantorovich Optimal Mass Transport
• Differential equations methods for the Monge-Kantorovich mass transfer problem
• Optimal, reliable estimation of quantum states
• Quantum state estimation.
• Parametric Fokker-Planck equation
• The Density Manifold and Configuration Space Quantization
• An analog of the 2-Wasserstein metric in non-commutative probability under which the fermionic Fokker-Planck equation is gradient flow for the entropy
• On the detailed balance condition for non-Hamiltonian systems
• Concentration of quantum states from quantum functional and transportation cost inequalities
• Information geometry and its applications
• Non-commutative calculus, optimal transport and functional inequalities in dissipative quantum systems
• Relating relative entropy, optimal transport and Fisher information: a quantum HWI inequality
• Computations of optimal transport distance with Fisher information regularization
• Natural gradient via optimal transport
• Information geometry
• Title not available (Why is that?)
• An entropic gradient structure for Lindblad equations and couplings of quantum systems to macroscopic models
• Matrix Optimal Mass Transport: A Quantum Mechanical Approach
• Ricci curvature for parametric statistics via optimal transport
• Optimal transport natural gradient for statistical manifolds with continuous sample space
• Quantum Information Theory
• Wasserstein geometry of quantum states and optimal transport of matrix-valued measures

Cited In (2)

• Variational learning the SDC quantum protocol with gradient-based optimization
• A dual formula for the noncommutative transport distance

This page was built for publication: Quantum statistical learning via quantum Wasserstein natural gradient

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2658080)