Training Algorithms for Hidden Conditional Random Fields

Author

Mahajan, Milind ; Gunawardana, Asela ; Acero, Alex

Author_Institution

Microsoft Res., Redmond, WA

Volume

1

fYear

2006

fDate

14-19 May 2006

Abstract

We investigate algorithms for training hidden conditional random fields (HCRFs) - a class of direct models with hidden state sequences. We compare stochastic gradient ascent with the RProp algorithm, and investigate stochastic versions of RProp. We propose a new scheme for model flattening, and compare it to the state of the art. Finally we give experimental results on the TEMIT phone classification task showing how these training options interact, comparing HCRFs to HMMs trained using extended Baum-Welch as well as stochastic gradient methods

Keywords

gradient methods; hidden Markov models; speech recognition; HMM; extended Baum-Welch; hidden Markov model; hidden conditional random fields; hidden state sequences; model flattening; speech classification; speech recognition; stochastic gradient methods; training algorithms; Acoustics; Gradient methods; Hidden Markov models; Maximum likelihood estimation; Mutual information; Optimization methods; Speech recognition; Stochastic processes; Training data;

fLanguage

English

Publisher

ieee

Conference_Titel

Acoustics, Speech and Signal Processing, 2006. ICASSP 2006 Proceedings. 2006 IEEE International Conference on

Conference_Location

Toulouse

ISSN

1520-6149

Print_ISBN

1-4244-0469-X

Type

conf

DOI

10.1109/ICASSP.2006.1660010

Filename

1660010