Empirical Assessment of Two Strategies for Optimizing the Viterbi Algorithm

Authors:
Roberto Esposito;Daniele P. Radicioni
Affiliations:
Dipartimento di Informatica, Università di Torino,;Dipartimento di Informatica, Università di Torino,
Venue:
AI*IA '09: Proceedings of the XIth International Conference of the Italian Association for Artificial Intelligence Reggio Emilia on Emergent Perspectives in Artificial Intelligence
Year:
2009

Citing 7
Cited 0

The Harpy speech understanding system

Readings in speech recognition
Fast inference and learning in large-state-space HMMs

ICML '05 Proceedings of the 22nd international conference on Machine learning
Discriminative training methods for hidden Markov models: theory and experiments with perceptron algorithms

EMNLP '02 Proceedings of the ACL-02 conference on Empirical methods in natural language processing - Volume 10
Incremental parsing with the perceptron algorithm

ACL '04 Proceedings of the 42nd Annual Meeting on Association for Computational Linguistics
CarpeDiem: an algorithm for the fast evaluation of SSL classifiers

Proceedings of the 24th international conference on Machine learning
On learning linear ranking functions for beam search

Proceedings of the 24th international conference on Machine learning
Structured machine learning: the next ten years

Machine Learning

Quantified Score

Hi-index	0.01

Visualization

Abstract

The Viterbi algorithm is widely used to evaluate sequential classifiers. Unfortunately, depending on the number of labels involved, its time complexity can still be too high for practical purposes. In this paper, we empirically compare two approaches to the optimization of the Viterbi algorithm: Viterbi Beam Search and CarpeDiem . The algorithms are illustrated and tested on datasets representative of a wide range of experimental conditions. Results are reported and the conditions favourable to the characteristics of each approach are discussed.