Abstract:We study the problem of learning a partially observed matrix under the low rank assumption in the presence of fully observed side information that depends linearly on the true underlying matrix. This problem consists of an important generalization of the Matrix Completion problem, a central problem in Statistics, Operations Research and Machine Learning, that arises in applications such as recommendation systems, signal processing, system identification and image denoising. We formalize this problem as an optimization problem with an objective that balances the strength of the fit of the reconstruction to the observed entries with the ability of the reconstruction to be predictive of the side information. We derive a mixed-projection reformulation of the resulting optimization problem and present a strong semidefinite cone relaxation. We design an efficient, scalable alternating direction method of multipliers algorithm that produces high quality feasible solutions to the problem of interest. Our numerical results demonstrate that in the small rank regime ($k \leq 15$), our algorithm outputs solutions that achieve on average $79\%$ lower objective value and $90.1\%$ lower $\ell_2$ reconstruction error than the solutions returned by the best performing benchmark method on synthetic data. The runtime of our algorithm is competitive with and often superior to that of the benchmark methods. Our algorithm is able to solve problems with $n = 10000$ rows and $m = 10000$ columns in less than a minute. On large scale real world data, our algorithm produces solutions that achieve $67\%$ lower out of sample error than benchmark methods in $97\%$ less execution time.

Bayesian Learning for Low-Rank matrix reconstruction

Sparse Bayesian Learning Approach for Discrete Signal Reconstruction

Bayesian Low-rank Matrix Completion with Dual-graph Embedding: Prior Analysis and Tuning-free Inference

Bayesian Inference for Adaptive Low Rank and Sparse Matrix Estimation

Low-Rank Matrix Completion Based on Maximum Likelihood Estimation.

Bayesian Matrix Completion Via Adaptive Relaxed Spectral Regularization

Estimation of high-dimensional low-rank matrices

Low-Rank Tensor Completion: A Pseudo-Bayesian Learning Approach.

Learning-Based Low-Rank Approximations

Bayesian matrix completion: prior specification

Simultaneously Sparse and Low-Rank Matrix Reconstruction Via Nonconvex and Nonseparable Regularization

Bayesian sparse multiple regression for simultaneous rank reduction and variable selection

Fast Low-Rank Bayesian Matrix Completion with Hierarchical Gaussian Prior Models

Low rank matrix recovery using unconstrained smoothed lq minimization

Predictive Low Rank Matrix Learning under Partial Observations: Mixed-Projection ADMM

Bayesian Uncertainty Quantification for Low-Rank Matrix Completion

Fundamental limits of Non-Linear Low-Rank Matrix Estimation

Reconstruction of Structurally-Incomplete Matrices With Reweighted Low-Rank and Sparsity Priors.

Low-rank Bayesian matrix completion via geodesic Hamiltonian Monte Carlo on Stiefel manifolds

Can Learning Be Explained By Local Optimality In Low-rank Matrix Recovery?

Tuning Free Rank-Sparse Bayesian Matrix and Tensor Completion with Global-Local Priors