Google llc (20240257799). Training Speech Recognizers Based On Biased Transcriptions simplified abstract
Contents
Training Speech Recognizers Based On Biased Transcriptions
Organization Name
Inventor(s)
Dragan Zivkovic of Sunnyvale CA (US)
Training Speech Recognizers Based On Biased Transcriptions - A simplified explanation of the abstract
This abstract first appeared for US patent application 20240257799 titled 'Training Speech Recognizers Based On Biased Transcriptions
The method described in the abstract involves receiving a biased transcription of a voice command spoken by a user, instructing an application to perform an action based on the biased transcription, and generating a confidence score for the transcription.
- The method receives a biased transcription for a voice command spoken by a user.
- The biased transcription includes a biasing phrase specific to the user.
- An application on the user device is instructed to perform an action based on the biased transcription.
- User behavior signals are received in response to the application performing the action.
- A confidence score for the biased transcription is generated using a confidence model.
- The speech recognizer is trained based on the confidence score.
Potential Applications: - Personalized voice command recognition systems - User-specific voice assistant technology - Behavioral analysis in voice command interactions
Problems Solved: - Improving accuracy of voice command recognition - Enhancing user experience with voice-controlled applications - Tailoring voice recognition systems to individual users
Benefits: - Increased efficiency in executing voice commands - Enhanced user satisfaction with voice assistant technology - Customized user experience based on individual preferences
Commercial Applications: Title: Personalized Voice Command Recognition Technology This technology can be applied in various industries such as smart home devices, automotive systems, and customer service applications to provide personalized voice command recognition tailored to individual users.
Questions about Personalized Voice Command Recognition Technology: 1. How does the method ensure the biasing phrases are accurately incorporated into the transcription? 2. What are the potential privacy implications of using user-specific biasing phrases in voice command recognition systems?
Frequently Updated Research: Research on improving the accuracy and efficiency of voice command recognition systems through user-specific biasing phrases and behavioral analysis is ongoing. Stay updated on the latest advancements in this field to understand the evolving capabilities of personalized voice assistant technology.
Original Abstract Submitted
a method includes receiving a biased transcription for a voice command spoken by a user and captured by a user device, the biased transcription biased to include a biasing phrase from a set of biasing phrases specific to the user. the method also includes instructing an application executing on the user device to perform an action specified by the biased transcription for the voice command, and receiving one or more user behavior signals responsive to the application performing the action specified by the biased transcription. the method further includes generating, as output from a confidence model, a confidence score of the biased transcription based on the one or more user behavior signals input to the confidence model and, based on the confidence score output from the confidence model, training a speech recognizer on the biased transcription.
