LEVERAGING INTERACTION CONTEXT TO IMPROVE RECOGNITION CONFIDENCE SCORES
2 Assignments
0 Petitions
Accused Products
Abstract
On a computing device a speech utterance is received from a user. The speech utterance is a section of a speech dialog that includes a plurality of speech utterances. One or more features from the speech utterance are identified. Each identified feature from the speech utterance is a specific characteristic of the speech utterance. One or more features from the speech dialog are identified. Each identified feature from the speech dialog is associated with one or more events in the speech dialog. The one or more events occur prior to the speech utterance. One or more identified features from the speech utterance and one or more identified features from the speech dialog are used to calculate a confidence score for the speech utterance.
11 Citations
30 Claims
-
1-10. -10. (canceled)
-
11. A method for improving speech recognition on a computing device, the method comprising:
-
on the computing device, obtaining a log file for a speech dialog, the speech dialog including one or more speech utterances; on the computing device, automatically extracting from the log file one or more dialog-level features from the speech dialog, each dialog-level feature being a specific characteristic of the speech dialog; on the computing device, automatically extracting from the log file a value associated with each of the one or more dialog-level features from the speech dialog. on the computing device, obtaining from the log file one or more features associated with the one or more speech utterances; on the computing device, obtaining from the log file a first confidence score associated with the one or more features associated with the one or more speech utterances; and on the computing device, using values associated with the one or more dialog-level features to adjust the first confidence score associated with one or more of the speech utterances. - View Dependent Claims (12, 15, 16, 17, 18)
-
-
19. A computer-readable storage medium comprising instructions that, when executed by a computing device, cause the computing device to:
-
receive a speech utterance from a user, the speech utterance being a section of a speech dialog, the speech dialog including a plurality of speech utterances; identify one or more features from the speech utterance, each identified feature from the speech utterance being a specific characteristic of the speech utterance; identify one or more features from the speech dialog, each identified feature from the speech dialog being associated with one or more events in the speech dialog, the one or more events occurring prior to the speech utterance; input the one or more features of the speech utterance to a first confidence classifier module; use the first confidence classifier module to calculate a first confidence score for the speech utterance; input the first confidence score for the speech utterance to a second confidence classifier module; input the one or more features from the speech dialog to the second confidence classifier module; and use the first confidence score and the one or more features from the speech dialog to calculate a second confidence score for the speech utterance. - View Dependent Claims (20)
-
-
21. A system for improving speech recognition on a computing device, the system comprising:
-
a confidence classifier module for; obtaining a log file for a speech dialog, the speech dialog including one or more speech utterances; automatically extracting from the log file one or more dialog-level features from the speech dialog, each dialog-level feature being a specific characteristic of the speech dialog, and wherein the one or more dialog-level features include at least a position of each utterance in the speech dialog; automatically extracting from the log file a value associated with each of the one or more dialog-level features from the speech dialog. obtaining from the log file one or more features associated with the one or more speech utterances; obtaining from the log file a first confidence score associated with the one or more features associated with the one or more speech utterances; and using values associated with the one or more dialog-level features to adjust the first confidence score associated with one or more of the speech utterances. - View Dependent Claims (22, 23, 24, 25, 26, 27, 28, 29, 30)
-
Specification