Sparse Bilinear Logistic Regression

Abstract

In this paper, we introduce the concept of sparse bilinear logistic regression for decision problems involving explanatory variables that are two-dimensional matrices. Such problems are common in computer vision, brain-computer interfaces, style/content factorization, and parallel factor analysis. The underlying optimization problem is bi-convex; we study its solution and develop an efficient algorithm based on block coordinate descent. We provide a theoretical guarantee for global convergence and estimate the asymptotical convergence rate using the Kurdyka-Łojasiewicz inequality. A range of experiments with simulated and real data demonstrate that sparse bilinear logistic regression outperforms current techniques in several important applications.

27 pages, 5 figures

References in corpus (10)

• Least Angle Regression
• Discussion of "Least angle regression" by Efron et al
• Discussion of "Least angle regression" by Efron et al
• Discussion of "Least angle regression" by Efron et al
• Discussion of "Least angle regression" by Efron et al
• Discussion of "Least angle regression" by Efron et al
• Discussion of "Least angle regression" by Efron et al
• Discussion of "Least angle regression" by Efron et al
• Discussion of "Least angle regression" by Efron et al
• Rejoinder to "Least angle regression" by Efron et al

Cited by in corpus (7)

• Bitcoin Volatility Forecasting with a Glimpse into Buy and Sell Orders
• A Primer on Coordinate Descent Algorithms
• Multidimensional Item Response Theory in the Style of Collaborative Filtering
• Momentum-based variance-reduced proximal stochastic gradient method for composite nonconvex stochastic optimization
• A Minimax Lower Bound for Low-Rank Matrix-Variate Logistic Regression
• Block stochastic gradient iteration for convex and nonconvex optimization
• Non-linear regression models for behavioral and neural data analysis