Automatic selection of Pareto-optimal topologies of hidden Markov models using multicriteria evolutionary algorithms

Authors:
Pawel Swietojanski;Robert Wielgat;Tomasz Zielinski
Affiliations:
Higher State Vocational School in Tarnow;Higher State Vocational School in Tarnow, Tarnow, Poland;AGH University of Science and Technology in Cracow, Department of Telecommunications, Cracow, Poland
Venue:
EvoApplications'11 Proceedings of the 2011 international conference on Applications of evolutionary computation - Volume Part I
Year:
2011

Citing 5
Cited 0

Evolving the Topology of Hidden Markov Models Using Evolutionary Algorithms

PPSN VII Proceedings of the 7th International Conference on Parallel Problem Solving from Nature
A Micro-Genetic Algorithm for Multiobjective Optimization

EMO '01 Proceedings of the First International Conference on Evolutionary Multi-Criterion Optimization
Genetic Algorithms as an Alternative Method of Parameter Estimation and Finding Most Likely Sequences of States of Hidden Markov Chains for HMMs and Hybrid HMM/ANN Models

Fundamenta Informaticae
Multiobjective Optimization: Interactive and Evolutionary Approaches

Multiobjective Optimization: Interactive and Evolutionary Approaches
Evolving the structure of hidden Markov models

IEEE Transactions on Evolutionary Computation

Quantified Score

Hi-index	0.00

Visualization

Abstract

In this paper a novel approach of automatic selection of Hidden Markov Models (HMM) structures under Pareto-optimality criteria is presented. Proof of concept is delivered in automatic speech recognition (ASR) discipline where two research scenarios including recognition of speech disorders as well as classification of bird species using their voice are performed. The conducted research unveiled that the Pareto Optimal Hidden Markov Models (POHMM) topologies outperformed both manual structures selection based on theoretical prejudices as well as the automatic approaches that used a single objective only.