言語モデル作成装置、言語モデル作成方法、そのプログラムおよびその記録媒体 - Nippon Telegraph and Telephone Corporation

Title:

言語モデル作成装置、言語モデル作成方法、そのプログラムおよびその記録媒体

Document Type and Number:

Japanese Patent JP4537970

Kind Code:

B2

Abstract:

To generate a language model (adaptive language model) matching target use without requiring text data adapted to the target use.

Language models (cluster language models) corresponding to a plurality of text data clusters are generated from the respective text data clusters. Then a language model (a composite cluster language model) is composed of a plurality of combinations of respective cluster language models, and the respective cluster language models and/or respective composite cluster language models are evaluated by using data for evaluation and sound models for evaluation to output evaluation results of the respective language models. A language model giving a predetermined evaluation result among the respective evaluation results is selected.

More Like This:

JP4405542	Devices, methods and programs for clustering phoneme models
JPS6167899	VOICE RECOGNITION EQUIPMENT
JP2003295885	METHOD AND APPARATUS FOR CONSTRUCTING SPEECH RECOGNITION DICTIONARY, METHOD AND APPARATUS FOR SPEECH RECOGNITION, SPEECH RECOGNITION DICTIONARY CONSTRUCTING PROGRAM AND RECORDING MEDIUM WHERE THE PROGRAM IS RECORDED, AND SPEECH RECOGNIZING PROGRAM AND RECORDING MEDIUM WHERE THE PROGRAM IS RECORDED

Inventors:

Hirokazu Masataki
Satoshi Kobashigawa

Application Number:

JP2006075364A

Publication Date:

September 08, 2010

Filing Date:

March 17, 2006

Export Citation:

Click for automatic bibliography generation Help

Assignee:

Nippon Telegraph and Telephone Corporation

International Classes:

G10L15/06; G10L15/065; G10L15/187; G10L15/197

Domestic Patent References:

JP2004317845A
JP11259084A
JP2005284209A
JP2002091484A

Attorney, Agent or Firm:

Naoki Nakao
Taku Kusano
Yukio Nakamura

Previous Patent: 高周波パッケージモジュール

Next Patent: 乗用型田植機