Hidden Markov Model Identifiability via Tensors

TL;DR

This paper establishes necessary and sufficient conditions for the identifiability of discrete-time finite alphabet hidden Markov models by linking the problem to tensor decomposition, resolving a long-standing open problem.

Contribution

It introduces a tensor-based approach to characterize HMM identifiability, providing a complete solution for single and multiple observer cases.

Findings

Derived necessary and sufficient conditions for HMM identifiability.

Connected HMM identifiability to tensor decomposition uniqueness.

Extended results to HMMs with multiple observers.

Abstract

The prevalence of hidden Markov models (HMMs) in various applications of statistical signal processing and communications is a testament to the power and flexibility of the model. In this paper, we link the identifiability problem with tensor decomposition, in particular, the Canonical Polyadic decomposition. Using recent results in deriving uniqueness conditions for tensor decomposition, we are able to provide a necessary and sufficient condition for the identification of the parameters of discrete time finite alphabet HMMs. This result resolves a long standing open problem regarding the derivation of a necessary and sufficient condition for uniquely identifying an HMM. We then further extend recent preliminary work on the identification of HMMs with multiple observers by deriving necessary and sufficient conditions for identifiability in this setting.