Statistical inference of finite-rank tensors

DOI10.5802/AHL.146zbMATH Open1515.82083arXiv2104.05360OpenAlexW3155800910MaRDI QIDQ6159729FDOQ6159729

Authors: Hongbin Chen, J-C. Mourrat, Jiaming Xia 

Publication date: 20 June 2023

Published in: Annales Henri Lebesgue (Search for Journal in Brave)

Abstract: We consider a general statistical inference model of finite-rank tensor products. For any interaction structure and any order of tensor products, we identify the limit free energy of the model in terms of a variational formula. Our approach consists of showing first that the limit free energy must be the viscosity solution to a certain Hamilton-Jacobi equation.

Full work available at URL: https://arxiv.org/abs/2104.05360

Recommendations

• Hamilton-Jacobi equations for inference of matrix tensor products
• Hamilton-Jacobi equations for finite-rank matrix inference
• Mutual information for low-rank even-order symmetric tensor estimation
• Inference for low-rank tensors -- no need to debias
• Statistical thresholds for tensor PCA

zbMATH Keywords

tensorHamilton-Jacobi equationinference problem

Mathematics Subject Classification ID

Disordered systems (random Ising models, random Schrödinger operators, etc.) in equilibrium statistical mechanics (82B44) Statistical mechanics of random media, disordered materials (including liquid crystals and spin glasses) (82D30)

Cites Work

• Convex Analysis
• Matrix theory. Basic results and techniques
• A mechanical approach to mean field spin models
• Title not available (Why is that?)
• Fundamental limits of symmetric low-rank matrix estimation
• Nonconvex interactions in mean-field spin glasses
• Extending the Parisi formula along a Hamilton-Jacobi equation
• The solution of the deep Boltzmann machine on the Nishimori line
• Hamilton-Jacobi equations for finite-rank matrix inference
• Hamilton-Jacobi equations for nonsymmetric matrix inference
• Replica symmetry breaking in mean-field spin glasses through the Hamilton–Jacobi technique
• The adaptive interpolation method: a simple scheme to prove replica formulas in Bayesian inference
• Hamilton-Jacobi equations for mean-field disordered systems
• Mutual information for low-rank even-order symmetric tensor estimation
• The Parisi formula is a Hamilton–Jacobi equation in Wasserstein space
• Hamilton-Jacobi equations for inference of matrix tensor products
• The adaptive interpolation method for proving replica formulas. Applications to the Curie–Weiss and Wigner spike models

Cited In (8)

• Title not available (Why is that?)
• Infinite-dimensional Hamilton-Jacobi equations for statistical inference on sparse graphs
• Mutual information for the sparse stochastic block model
• Tensor Canonical Correlation Analysis With Convergence and Statistical Guarantees
• Tensor methods for full-information maximum likelihood estimation: Unconstrained estimation
• Free Energy in Spin Glass Models with Conventional Order
• On the free energy of vector spin glasses with nonconvex interactions
• \(L^{\infty}\) estimation of tensor truncations

This page was built for publication: Statistical inference of finite-rank tensors

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6159729)