Monte Carlo Hidden Markov Models

Abstract

We present a learning algorithm for hidden Markov models with continuous state and observation spaces. All necessary probability density functions are approximated using samples, along with density trees generated from such samples. A Monte Carlo version of Baum-Welch (EM) is employed to learn models from data, just as in regular HMM learning. Regularization during learning is obtained using an exponential shrinking technique. The shrinkage factor, which determines the effective capacity of the learning algorithm, is annealed down over multiple iterations of Baum-Welch, and early stopping is applied to select the right model. We prove that under mild assumptions, Monte Carlo Hidden Markov Models converge to a local maximum in likelihood space, just like conventional HMMs. In addition, we provide empirical results obtained in a gesture recognition domain, which illustrate the appropriateness of the approach in practice.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Dec 01, 1998
Accession Number
ADA363714

Entities

People

  • John Langford
  • Sebastian Thrun

Organizations

  • Carnegie Mellon University

Tags

Communities of Interest

  • Autonomy
  • Energy and Power Technologies
  • Materials and Manufacturing Processes

DTIC Thesaurus Topics

  • Algorithms
  • Artificial Intelligence
  • Computational Complexity
  • Computational Science
  • Computer Science
  • Data Sets
  • Databases
  • Hidden Markov Models
  • Machine Learning
  • Markov Models
  • Monte Carlo Method
  • Probability
  • Probability Distributions
  • Random Variables
  • Recognition
  • Sampling
  • Signal Processing

Readers

  • Computational Modeling and Simulation
  • Mathematical Modeling and Probability Theory.
  • Speech Processing/Speech Recognition.

Technology Areas

  • Space