Language Model Adaptation

DeMori, R.; Federico, Marcello

doi:10.1007/978-3-642-60087-6_26

Cited by 21 publications

(8 citation statements)

References 12 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…computed by the j th language model. Given some target domain text, these weight can be set using an expectation-maximisation procedure [19]. When the target domain text is not available, some target domain audio can be recognised to yield hypothesised text [20].…”

Section: Language Model Interpolationmentioning

confidence: 99%

Automatic Speech Recognition System Development in the "Wild"

Ragni

Gales

2018

Interspeech 2018

View full text Add to dashboard Cite

The standard framework for developing an automatic speech recognition (ASR) system is to generate training and development data for building the system, and evaluation data for the final performance analysis. All the data is assumed to come from the domain of interest. Though this framework is matched to some tasks, it is more challenging for systems that are required to operate over broad domains, or where the ability to collect the required data is limited. This paper discusses ASR work performed under the IARPA MATERIAL program, which is aimed at cross-language information retrieval, and examines this challenging scenario. In terms of available data, only limited narrow-band conversational telephone speech data was provided. However, the system is required to operate over a range of domains, including broadcast data. As no data is available for the broadcast domain, this paper proposes an approach for system development based on scraping "related" data from the web, and using ASR system confidence scores as the primary metric for developing the acoustic and language model components. As an initial evaluation of the approach, the Swahili development language is used, with the final system performance assessed on the IARPA MATERIAL Analysis Pack 1 data.

show abstract

Section: Language Model Interpolationmentioning

confidence: 99%

Automatic Speech Recognition System Development in the "Wild"

Ragni

Gales

2018

Interspeech 2018

View full text Add to dashboard Cite

show abstract

“…It is for instance possible to apply a transformation to this matrix in order to accommodate for new LM data. Similar ideas have also been proposed for standard ¢ -gram LM (see for instance [8] for an review of LM adaptation techniques), but the discrete representation makes it mathematically less tractable.…”

Section: Discussion and Future Workmentioning

confidence: 99%

Connectionist language modeling for large vocabulary continuous speech recognition

Schwenk¹,

Gauvain²

2002

IEEE International Conference on Acoustics Speech and Signal Processing

View full text Add to dashboard Cite

This paper describes ongoing work on a new approach for language modeling for large vocabulary continuous speech recognition. Almost all state-of-the-art systems use statistical ¢-gram language models estimated on text corpora. One principle problem with such language models is the fact that many of the ¢-grams are never observed even in very large training corpora, and therefore it is common to back-off to a lower-order model. In this paper we propose to address this problem by carrying out the estimation task in a continuous space, enabling a smooth interpolation of the probabilities. A neural network is used to learn the projection of the words onto a continuous space and to estimate the ¢-gram probabilities. The connectionist language model is being evaluated on the DARPA HUB5 conversational telephone speech recognition task and preliminary results show consistent improvements in both perplexity and word error rate.

show abstract

“…It is for instance possible to apply a transformation to this matrix in order to accommodate for new LM data. Similar ideas have also been proposed for stan dard n-gram LM (see for instance [8] for an review of LM adap tation techniques), but the discrete representation makes it mathe matically less tractable.…”

Section: Discussion and Future Workmentioning

confidence: 99%

Connectionist language modeling for large vocabulary continuous speech recognition

Schwenk

Gauvain

2002

IEEE International Conference on Acoustics Speech and Signal Processing

View full text Add to dashboard Cite

This paper describes ongoing work on a new approach for lan guage modeling for large vocabulary continuous speech recogni tion. Almost all state-of-the-art systems use statistical n-gram lan guage models estimated on text corpora. One principle problem with such language models is the fact that many of the n-grams are never observed even in very large training corpora, and there fore it is common to back-off to a lower-order model. In this paper we propose to address this problem by carryi ng out the estima tion task in a continuous space, enabling a smooth interpolation of the probabilities. A neural network is used to learn the pro jection of the words onto a continuous space and to estimate the n-gram probabilities. The connectionist language model is being evaluated on the DARPA HUBS conversational telephone speech recognition task and preliminary results show consistent improve ments in both perplexity and word error rate.

show abstract

Language Model Adaptation

Cited by 21 publications

References 12 publications

Automatic Speech Recognition System Development in the "Wild"

Automatic Speech Recognition System Development in the "Wild"

Connectionist language modeling for large vocabulary continuous speech recognition

Connectionist language modeling for large vocabulary continuous speech recognition

Contact Info

Product

Resources

About