Local convexity of the TAP free energy and AMP convergence for \(\mathbb{Z}_2\)-synchronization

DOI10.1214/23-aos2257arXiv2106.11428OpenAlexW4380490510MaRDI QIDQ6172186

Zhou Fan, Michael Celentano, Song Mei

Publication date: 19 July 2023

Published in: The Annals of Statistics (Search for Journal in Brave)

Full work available at URL: https://arxiv.org/abs/2106.11428

zbMATH Keywords

nonconvex optimizationvariational inferencelandscape analysisapproximate message passingTAP free energynatural gradient descent\(\mathbb{Z}_2\) synchronization

Mathematics Subject Classification ID

Bayesian problems; characterization of Bayes procedures (62C10)

Related Items

On the concavity of the TAP free energy in the SK model ⋮ The TAP free energy for high-dimensional linear regression

Cites Work

• Unnamed Item
• Unnamed Item
• Unnamed Item
• Unnamed Item
• Unnamed Item
• Asymptotic normality of maximum likelihood and its variational approximation for stochastic blockmodels
• An iterative construction of solutions of the TAP equations for the Sherrington-Kirkpatrick model
• Angular synchronization by eigenvectors and semidefinite programming
• Asymptotic normality and valid inference for Gaussian variational approximation
• Spin glasses and Stein's method
• The largest eigenvalue of small rank perturbations of Hermitian random matrices
• Une inegalité du type de Slepian et Gordon sur les processus gaussiens. (An inequality of Slepian and Gordon type for Gaussian processes)
• Some inequalities for Gaussian processes and applications
• Gaussian-width gradient complexity, reverse log-Sobolev inequalities and nonlinear large deviations
• On the TAP free energy in the mixed \(p\)-spin models
• The complexity of spherical \(p\)-spin models: a second moment approach
• Fundamental limits of symmetric low-rank matrix estimation
• A geometric analysis of phase retrieval
• Mirror descent and nonlinear projected subgradient methods for convex optimization.
• On the distribution of the largest eigenvalue in principal components analysis
• An introduction to variational methods for graphical models
• The landscape of empirical risk for nonconvex losses
• Universality of approximate message passing algorithms
• Accelerated Bregman proximal gradient methods for relatively smooth convex optimization
• The distribution of the Lasso: uniform control over sparse balls and adaptive parameter tuning
• A spin glass model for the loss surfaces of generative adversarial networks
• Nonlinear large deviations: beyond the hypercube
• \(\alpha\)-variational inference with statistical guarantees
• Concentration of tempered posteriors and of their variational approximations
• Nonlinear large deviation bounds with applications to Wigner matrices and sparse Erdős-Rényi graphs
• Convergence rates of variational posterior distributions
• Theoretical and computational guarantees of mean field variational inference for community detection
• TAP free energy, spin glasses and variational inference
• Thouless-Anderson-Palmer equations for generic \(p\)-spin glasses
• Universality in polytope phase transitions and message passing algorithms
• Universality of the mean-field for the Potts model
• The TAP-Plefka variational principle for the spherical SK model
• Complexity of random smooth functions on the high-dimensional sphere
• Phase transition of the largest eigenvalue for nonnull complex sample covariance matrices
• Nonlinear large deviations
• Scalable variational inference for Bayesian variable selection in regression, and its accuracy in genetic association studies
• Estimation of low-rank matrices via approximate message passing
• Perturbed Fenchel duality and first-order methods
• Complexity of Random Energy Landscapes, Glass Transition, and Absolute Value of the Spectral Determinant of Random Matrices
• Phase transitions in semidefinite relaxations
• Non-Negative Principal Component Analysis: Message Passing Algorithms and Sharp Asymptotics
• Inference for Generalized Linear Models via Alternating Directions and Bethe Free Energy Minimization
• A Morita Type Proof of the Replica-Symmetric Formula for SK
• The loss surfaces of neural networks with general activation functions
• Graphical Models, Exponential Families, and Variational Inference
• On the formal equivalence of the TAP and thermodynamic methods in the SK model
• Finite Sample Analysis of Approximate Message Passing Algorithms
• Relatively Smooth Convex Optimization by First-Order Methods, and Applications
• Asymptotic mutual information for the balanced binary stochastic block model
• 10.1162/jmlr.2003.3.4-5.993
• Random Matrices and Complexity of Spin Glasses
• Non-unique games over compact groups and orientation estimation in cryo-EM
• State evolution for approximate message passing with non-separable functions
• State evolution for general approximate message passing algorithms, with applications to spatial coupling
• The Landscape of the Spiked Tensor Model
• Vector Approximate Message Passing
• Capacity lower bound for the Ising perceptron
• Frequentist Consistency of Variational Bayes
• The Dynamics of Message Passing on Dense Graphs, with Applications to Compressed Sensing
• Semidefinite programs on sparse random graphs and their application to community detection
• A CDMA multiuser detection algorithm on the basis of belief propagation
• A Descent Lemma Beyond Lipschitz Gradient Continuity: First-Order Methods Revisited and Applications
• A Unifying Tutorial on Approximate Message Passing
• Variational Bayes for High-Dimensional Linear Regression With Sparse Priors
• Mean Field Models for Spin Glasses
• The free energy of spherical pure \(p\)-spin models: computation from the TAP approach