1-bit matrix completion: PAC-Bayesian analysis of a variational approximation

DOI10.1007/s10994-017-5667-zzbMath1461.15032arXiv1604.04191OpenAlexW2949864082MaRDI QIDQ1640399

Pierre Alquier, Vincent Cottet

Publication date: 14 June 2018

Published in: Machine Learning (Search for Journal in Brave)

Full work available at URL: https://arxiv.org/abs/1604.04191

zbMATH Keywords

matrix completionoracle inequalitiesvariational Bayessupervised classificationPAC-Bayesian boundsrisk convexification

Mathematics Subject Classification ID

Classification and discrimination; cluster analysis (statistical aspects) (62H30) Problem solving in the context of artificial intelligence (heuristics, search strategies, etc.) (68T20) Matrix completion problems (15A83) Numerical methods for low-rank matrix approximation; matrix compression (65F55)

Related Items

Tight risk bound for high dimensional time series completion ⋮ A reduced-rank approach to predicting multiple binary responses through machine learning ⋮ User-friendly Introduction to PAC-Bayes Bounds ⋮ Concentration of tempered posteriors and of their variational approximations ⋮ An efficient adaptive MCMC algorithm for pseudo-Bayesian quantum tomography ⋮ Pseudo-Bayesian quantum tomography with rank-adaptation ⋮ Consistency of variational Bayes inference for estimation and model selection in mixtures ⋮ Numerical comparisons between Bayesian and frequentist low-rank matrix completion: estimation accuracy and uncertainty quantification ⋮ Flexible low-rank statistical modeling with missing data and side information ⋮ On the exponentially weighted aggregate with the Laplace prior ⋮ Estimation bounds and sharp oracle inequalities of regularized procedures with Lipschitz loss functions

Uses Software

• softImpute
• Jellyfish
• PRMLT

Cites Work

• Unnamed Item
• Unnamed Item
• Unnamed Item
• Unnamed Item
• Unnamed Item
• Unnamed Item
• Distributed Optimization and Statistical Learning via the Alternating Direction Method of Multipliers
• Adaptive multinomial matrix completion
• Aggregation by exponential weighting, sharp PAC-Bayesian bounds and sparsity
• Smooth discrimination analysis
• Statistical learning theory and stochastic optimization. Ecole d'Eté de Probabilitiés de Saint-Flour XXXI -- 2001.
• Statistical behavior and consistency of classification methods based on convex risk minimization.
• Matrix estimation by universal singular value thresholding
• A Bayesian approach for noisy matrix completion: optimal rate under general sampling distribution
• Parallel stochastic gradient algorithms for large-scale matrix completion
• Penalized regression, standard errors, and Bayesian Lassos
• Exact matrix completion via convex optimization
• A Max-Norm Constrained Minimization Approach to 1-Bit Matrix Completion
• On the properties of variational approximations of Gibbs posteriors
• PAC-Bayesian Inequalities for Martingales
• The Bayesian Lasso
• A PAC-Bayesian margin bound for linear classifiers
• 1-Bit matrix completion
• The Power of Convex Relaxation: Near-Optimal Matrix Completion