LGAINAOCJun 4, 2023

Complexity of Block Coordinate Descent with Proximal Regularization and Applications to Wasserstein CP-dictionary Learning

arXiv:2306.02420v13 citationsh-index: 11
Originality Incremental advance
AI Analysis

This provides theoretical guarantees for a classical optimization method, with incremental improvements in complexity analysis and application to a specific domain of probability distribution approximation.

The paper establishes a worst-case complexity bound of O(1/epsilon) iterations for the block coordinate descent with proximal regularization algorithm to reach an epsilon-stationary point for general nonconvex smooth objectives, and applies it to develop a provable and efficient algorithm for Wasserstein CP-dictionary learning.

We consider the block coordinate descent methods of Gauss-Seidel type with proximal regularization (BCD-PR), which is a classical method of minimizing general nonconvex objectives under constraints that has a wide range of practical applications. We theoretically establish the worst-case complexity bound for this algorithm. Namely, we show that for general nonconvex smooth objectives with block-wise constraints, the classical BCD-PR algorithm converges to an epsilon-stationary point within O(1/epsilon) iterations. Under a mild condition, this result still holds even if the algorithm is executed inexactly in each step. As an application, we propose a provable and efficient algorithm for `Wasserstein CP-dictionary learning', which seeks a set of elementary probability distributions that can well-approximate a given set of d-dimensional joint probability distributions. Our algorithm is a version of BCD-PR that operates in the dual space, where the primal problem is regularized both entropically and proximally.

Foundations

The foundational work for this paper's niche, ranked by how specifically the neighbourhood builds on it — not by global fame.

Your Notes