<ul><li>This paper presents a subgradient-based algorithm for system identification with non-smooth objectives.</li><li>The algorithm is important for robust system identification in safety-critical applications.</li><li>The paper analyzes the subgradient method and establishes linear convergence results for both the best and Polyak step sizes.</li><li>The time complexity of the subgradient algorithm is compared with standard solvers, and experimental results support the findings.</li></ul>

Subgradient Method for System Identification with Non-Smooth Objectives

Discover more