Online learning of Riemannian hidden Markov models in homogeneous Hadamard spaces

TL;DR

This paper introduces an online algorithm for Riemannian hidden Markov models that significantly improves speed and accuracy over previous Euclidean-based methods, especially in signal and image processing applications.

Contribution

The paper presents a novel online learning algorithm for Riemannian HMMs that overcomes memory and speed limitations of prior Baum-Welch based approaches.

Findings

Enhanced speed and efficiency in Riemannian HMM learning

More accurate modeling of data on Riemannian manifolds

Reduced memory usage compared to previous methods

Abstract

Hidden Markov models with observations in a Euclidean space play an important role in signal and image processing. Previous work extending to models where observations lie in Riemannian manifolds based on the Baum-Welch algorithm suffered from high memory usage and slow speed. Here we present an algorithm that is online, more accurate, and offers dramatic improvements in speed and efficiency.