Hierarchical Semi-Markov Conditional Random Fields for Recursive Sequential Data

TL;DR

This paper introduces the hierarchical semi-Markov conditional random field (HSCRF), a discriminative model for complex hierarchical sequential data, with algorithms for learning and inference, demonstrated on activity recognition and noun-phrase chunking.

Contribution

The paper proposes HSCRF, a novel hierarchical semi-Markov model with polynomial algorithms, extending HHMMs to a discriminative framework with partially-supervised learning capabilities.

Findings

HSCRF effectively models hierarchical, nested sequential data.

HSCRF achieves reasonable accuracy in activity recognition and noun-phrase chunking.

Algorithms support both fully and partially observed data learning.

Abstract

Inspired by the hierarchical hidden Markov models (HHMM), we present the hierarchical semi-Markov conditional random field (HSCRF), a generalisation of embedded undirectedMarkov chains tomodel complex hierarchical, nestedMarkov processes. It is parameterised in a discriminative framework and has polynomial time algorithms for learning and inference. Importantly, we consider partiallysupervised learning and propose algorithms for generalised partially-supervised learning and constrained inference. We demonstrate the HSCRF in two applications: (i) recognising human activities of daily living (ADLs) from indoor surveillance cameras, and (ii) noun-phrase chunking. We show that the HSCRF is capable of learning rich hierarchical models with reasonable accuracy in both fully and partially observed data cases.