Speech recognition with feedback from natural language processing for adaptation of acoustic model

• US 7,236,922 B2
• Filed: 03/08/2005
• Issued: 06/26/2007
• Est. Priority Date: 09/30/1999
• Status: Expired due to Fees

• Current Assignee: Sony Corporation (Sony Group Corp.)
• Original Assignee: Sony Corporation (Sony Group Corp.)
• Inventors: Honda, Hitoshi, Omote, Masanori, Ogawa, Hiroaki, Pao, Hongchang
• Primary Examiner(s): Lerner; Martin
• Application Number:
  US11/075,561
• Publication Number:
  US 20050149319A1
• Time in Patent Office: 840 Days
• Field of Search: 704/2, 704/9, 704/243, 704/244, 704/251, 704/257
• US Class Current: 704/2
• CPC Class Codes: G10L 15/063 Training
• CPC Class Codes: G10L 15/075 supervised, i.e. under machine guidance
• CPC Class Codes: G10L 2015/0638 Interactive procedures

Abstract

A speech processing system including a speech recognition unit to receive input speech, and a natural language processing unit. The speech recognition unit performs speech recognition on input speech using acoustic models to produce a speech recognition result. The natural-language processing unit performs natural language processing on the speech recognition result, and includes: a machine translation unit; a dialog management unit; and a feedback unit. The feedback unit configured to feed back information obtained as a result of the natural language processing performed in the machine translation unit and the dialog management unit. The speech recognition unit processes feedback information including adaptation zones to adapt the acoustic models so that it can produce the speech recognition result with higher precision than when adaptation zones are not used.

Claims

1. A speech processing system comprising:

a speech recognition unit to receive input speech, said speech recognition unit performing speech recognition on input speech using acoustic models to produce a speech recognition result; and

a natural-language processing unit to perform natural language processing on said speech recognition result, said natural-language processing unit including;

a machine translation unit configured to receive and translate said speech recognition result and a reply to said speech recognition result;

a dialog management unit configured to generate said reply to said speech recognition result; and

a feedback unit configured to feed back information obtained as a result of the natural language processing performed in said machine translation unit and said dialog management unit, said feedback information including adaptation zones,

wherein said speech recognition unit operates to process said feedback information including said adaptation zones to adapt the acoustic models so that said speech recognition unit produces said speech recognition result with higher precision than when said adaptation zones are not used.

2. A speech processing system according to claim 1,

wherein said feedback unit is configured to feed back at least one of said adaptation zones which are to be used for the adaptation of the acoustic models and at least one of said adaptation zones which are not to be used for the adaptation of the acoustic models.

3. A speech processing system according to claim 1,

wherein said feedback information includes reliability of said speech recognition result.