A Hybrid MaxEnt/HMM Based ASR System

Yasser Hifny; Steve Renals; Neil D. Lawrence

A Hybrid MaxEnt/HMM Based ASR System

Yasser Hifny, Steve Renals, Neil D. Lawrence

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Abstract / Description of output

The aim of this work is to develop a practical framework, which extends the classical Hidden Markov Models (HMM) for continuous speech recognition based on the Maximum Entropy (MaxEnt) principle. The MaxEnt models can estimate the posterior probabilities directly as with Hybrid NN/HMM connectionist speech recognition systems. In particular, a new acoustic modelling based on discriminative MaxEnt models is formulated and is being developed to replace the generative Gaussian Mixture Models (GMM) commonly used to model acoustic variability. Initial experimental results using the TIMIT phone task are reported.

Original language	English
Title of host publication	Proceedings of the 9th European Conference on Speech Communication and Technology
Subtitle of host publication	Interspeech'2005 - Eurospeech
Publisher	ISCA
Pages	3017-3020
Publication status	Published - 2005
Event	9th European Conference on Speech Communication and Technology (Interspeech 2005 - Eurospeech) - Lisbon, Portugal Duration: 4 Sept 2005 → 8 Sept 2005

Conference

Conference	9th European Conference on Speech Communication and Technology (Interspeech 2005 - Eurospeech)
Country/Territory	Portugal
City	Lisbon
Period	4/09/05 → 8/09/05

Access to Document

http://www.isca-speech.org/archive/interspeech_2005/i05_3017.html

Cite this

@inproceedings{6e7451de2a604861aea2cd3132caaec4,

title = "A Hybrid MaxEnt/HMM Based ASR System",

abstract = "The aim of this work is to develop a practical framework, which extends the classical Hidden Markov Models (HMM) for continuous speech recognition based on the Maximum Entropy (MaxEnt) principle. The MaxEnt models can estimate the posterior probabilities directly as with Hybrid NN/HMM connectionist speech recognition systems. In particular, a new acoustic modelling based on discriminative MaxEnt models is formulated and is being developed to replace the generative Gaussian Mixture Models (GMM) commonly used to model acoustic variability. Initial experimental results using the TIMIT phone task are reported.",

author = "Yasser Hifny and Steve Renals and Lawrence, {Neil D.}",

year = "2005",

language = "English",

pages = "3017--3020",

booktitle = "Proceedings of the 9th European Conference on Speech Communication and Technology",

publisher = "ISCA",

note = "9th European Conference on Speech Communication and Technology (Interspeech 2005 - Eurospeech) ; Conference date: 04-09-2005 Through 08-09-2005",

}

Hifny, Y, Renals, S & Lawrence, ND 2005, A Hybrid MaxEnt/HMM Based ASR System. in Proceedings of the 9th European Conference on Speech Communication and Technology: Interspeech'2005 - Eurospeech. ISCA, pp. 3017-3020, 9th European Conference on Speech Communication and Technology (Interspeech 2005 - Eurospeech), Lisbon, Portugal, 4/09/05. <http://www.isca-speech.org/archive/interspeech_2005/i05_3017.html>

TY - GEN

T1 - A Hybrid MaxEnt/HMM Based ASR System

AU - Hifny, Yasser

AU - Renals, Steve

AU - Lawrence, Neil D.

PY - 2005

Y1 - 2005

N2 - The aim of this work is to develop a practical framework, which extends the classical Hidden Markov Models (HMM) for continuous speech recognition based on the Maximum Entropy (MaxEnt) principle. The MaxEnt models can estimate the posterior probabilities directly as with Hybrid NN/HMM connectionist speech recognition systems. In particular, a new acoustic modelling based on discriminative MaxEnt models is formulated and is being developed to replace the generative Gaussian Mixture Models (GMM) commonly used to model acoustic variability. Initial experimental results using the TIMIT phone task are reported.

AB - The aim of this work is to develop a practical framework, which extends the classical Hidden Markov Models (HMM) for continuous speech recognition based on the Maximum Entropy (MaxEnt) principle. The MaxEnt models can estimate the posterior probabilities directly as with Hybrid NN/HMM connectionist speech recognition systems. In particular, a new acoustic modelling based on discriminative MaxEnt models is formulated and is being developed to replace the generative Gaussian Mixture Models (GMM) commonly used to model acoustic variability. Initial experimental results using the TIMIT phone task are reported.

M3 - Conference contribution

SP - 3017

EP - 3020

BT - Proceedings of the 9th European Conference on Speech Communication and Technology

PB - ISCA

T2 - 9th European Conference on Speech Communication and Technology (Interspeech 2005 - Eurospeech)

Y2 - 4 September 2005 through 8 September 2005

ER -

A Hybrid MaxEnt/HMM Based ASR System

Abstract / Description of output

Conference

Access to Document

Fingerprint

Cite this