Geometric Optimization Methods for Adaptive Filtering

Abstract

The techniques and analysis presented in this thesis provide new methods to solve optimization problems posed on Riemannian manifolds. These methods are applied to the subspace tracking problem found in adaptive signal processing and adaptive control. A new point of view is offered for the constrained optimization problem. Some classical optimization techniques on Euclidean space are generalized to Riemannian manifolds. Several algorithms are presented and their convergence properties are analyzed employing the Riemannian structure of the manifold. Specifically, two new algorithms, which can be thought of as Newton's method and the conjugate gradient method on Riemannian manifolds, are presented and shown to possess quadratic and superlinear convergence, respectively. These methods are applied to several eigenvalue and singular value problems, which are posed as constrained optimization problems. ...

PhD Dissertation, Harvard University, 1993. A thesis presented by Steven Thomas Smith to The Division of Applied Sciences in partial fulfillment for the degree of Doctor of Philosophy in the subject of Applied Mathematics

Cited by in corpus (18)

• Online Robust Subspace Tracking from Partial Information
• Cheap Orthogonal Constraints in Neural Networks: A Simple Parametrization of the Orthogonal and Unitary Group
• Unsupervised Learning Discriminative MIG Detectors in Nonhomogeneous Clutter
• LDA-MIG Detectors for Maritime Targets in Nonhomogeneous Sea Clutter
• Trivializations for Gradient-Based Optimization on Manifolds
• Manifold Optimization Over the Set of Doubly Stochastic Matrices: A Second-Order Geometry
• Double-bracket quantum algorithms for diagonalization
• The Comparison of Riemannian Geometric Matrix-CFAR Signal Detectors
• Towards a median signal detector through the total Bregman divergence and its robustness analysis
• Tensor completion using geodesics on Segre manifolds
• Stochastic Primal-Dual Method on Riemannian Manifolds with Bounded Sectional Curvature
• Double-bracket algorithm for quantum signal processing without post-selection
• Pactical Newton Methods for Electronic Structure Calculations
• Riemannian Stochastic Variance-Reduced Cubic Regularized Newton Method for Submanifold Optimization
• Convergence Analysis of Gradient Algorithms on Riemannian Manifolds Without Curvature Constraints and Application to Riemannian Mass
• Damped Newton's Method on Riemannian Manifolds
• Equilibrium problems on Riemannian manifolds with applications
• Convergence analysis of inexact descent algorithm for multiobjective optimizations on Riemannian manifolds without curvature constraints