A Minimax Lower Bound for Low-Rank Matrix-Variate Logistic Regression

Abstract

This paper considers the problem of matrix-variate logistic regression. It derives the fundamental error threshold on estimating low-rank coefficient matrices in the logistic regression problem by obtaining a lower bound on the minimax risk. The bound depends explicitly on the dimension and distribution of the covariates, the rank and energy of the coefficient matrix, and the number of samples. The resulting bound is proportional to the intrinsic degrees of freedom in the problem, which suggests the sample complexity of the low-rank matrix logistic regression problem can be lower than that for vectorized logistic regression. The proof techniques utilized in this work also set the stage for development of minimax lower bounds for tensor-variate logistic regression problems.

arXiv comment8 pages; published in Proc. 55th Asilomar Conf. Signals, Systems, and Computers, Pacific Grove, CA, Oct. 31-Nov. 3, 2021

References in corpus (7)

• Restricted strong convexity and weighted matrix completion: Optimal bounds with noise
• Logistic Regression: The Importance of Being Improper
• Minimax Lower Bounds on Dictionary Learning for Tensor Data
• Sparse Bilinear Logistic Regression
• Minimax Lower Bounds for Kronecker-Structured Dictionary Learning
• Learning Product Graphs Underlying Smooth Graph Signals
• Minimax Bounds for Distributed Logistic Regression