Method of adapting linguistic speech models
First Claim
Patent Images
1. A method of adapting linguistic speech models in automatic speech recognition systems, the method comprising the steps of:
- receiving speech recognition results during the operation of the systems; and
evaluating a list of best recognition result alternatives for a speech utterance to be recognized wherein a combination of a plurality of recognition result alternatives of the list is included in the adaptation, wherein in the combination of recognition result alternatives of the list, probability values assigned to these alternatives are weighted with a given numerical value, and an adaptation weight for a recognition result alternative used for the adaptation is formed in that the weighted probability value assigned to this recognition result alternative is related to the sum of the weighted probability values assigned to the other recognition result alternatives of the list.
2 Assignments
0 Petitions
Accused Products
Abstract
Method and apparatus are disclosed for adapting linguistic speech models in automatic speech recognition systems using speech recognition results obtained during operation of such systems, in which, during the adaptation, a list of N-best recognition result alternatives with N>1 for a speech utterance to be recognized is evaluated. To improve such an on-line adaptation of the linguistic speech models, a combination of a plurality of recognition result alternatives of the list is included in the adaptation.
-
Citations
3 Claims
-
1. A method of adapting linguistic speech models in automatic speech recognition systems, the method comprising the steps of:
-
receiving speech recognition results during the operation of the systems; and
evaluating a list of best recognition result alternatives for a speech utterance to be recognized wherein a combination of a plurality of recognition result alternatives of the list is included in the adaptation, wherein in the combination of recognition result alternatives of the list, probability values assigned to these alternatives are weighted with a given numerical value, and an adaptation weight for a recognition result alternative used for the adaptation is formed in that the weighted probability value assigned to this recognition result alternative is related to the sum of the weighted probability values assigned to the other recognition result alternatives of the list. - View Dependent Claims (2)
in which ω
i is the adaptation weight relating to the ith element of the list of N-best recognition result alternatives, andli is the probability value of the ith element of the list of N-best recognition result alternatives.
-
-
3. A speech recognition system wherein a linguistic speech model used for speech recognition is adapted, the system comprising:
- a processor configured to receive speech recognition results and to evaluate a list of best recognition result alternatives for a speech utterance to be recognized wherein a combination of a plurality of recognition result alternatives of the list is included in the adaptation, wherein in the combination of recognition result alternatives of the list, probability values assigned to these alternatives are weighted with a given numerical value, and an adaptation weight for a recognition result alternative used for the adaptation is formed in that the weighted probability value assigned to this recognition result alternative is related to the sum of the weighted probability values assigned to the other recognition
Specification