2023/03/26 by Cameron J. Blocker, Blocker, Cameron J., Haroon Raja +5
Engineering · Mathematics · #Advanced Statistical Methods and Models #Computation (stat.CO) #FOS: Computer and information sciences #FOS: Electrical engineering #Signal Processing (eess.SP) #Sparse and Compressive Sensing Techniques #Statistical Methods and Inference #electronic engineering #information engineering
paper · pdf · doi:10.48550/arxiv.2303.14851
openalex publication_date 2023/03/26 · openalex created_date 2025/10/10 · openalex updated_date 2026/08/01
Dynamic subspace estimation, or subspace tracking, is a fundamental problem in statistical signal processing and machine learning. This paper considers a geodesic model for time-varying subspaces. The natural cost function for this model is non-convex. We propose a novel optimization algorithm for this cost that estimates the model parameters from data with Grassmannian constraints. We show that with this algorithm, the cost is monotonically non-increasing. We demonstrate the performance of this model and our algorithm on synthetic data, video data, and dynamic fMRI data.