17842166. CORRECTING SPEECH RECOGNITION ERRORS BY CONSIDERING PRIOR USER EDITS AND/OR ASSESSING FULFILLMENT DATA simplified abstract (Google LLC)
CORRECTING SPEECH RECOGNITION ERRORS BY CONSIDERING PRIOR USER EDITS AND/OR ASSESSING FULFILLMENT DATA
Organization Name
Inventor(s)
Miroslaw Michalski of Pfaffikon SZ (CH)
CORRECTING SPEECH RECOGNITION ERRORS BY CONSIDERING PRIOR USER EDITS AND/OR ASSESSING FULFILLMENT DATA - A simplified explanation of the abstract
This abstract first appeared for US patent application 17842166 titled 'CORRECTING SPEECH RECOGNITION ERRORS BY CONSIDERING PRIOR USER EDITS AND/OR ASSESSING FULFILLMENT DATA
Simplified Explanation
The patent application is about a system that corrects speech recognition errors based on user corrections and fulfillment data. Here is a simplified explanation of the abstract:
- The system generates a candidate speech recognition hypothesis when a user speaks to an application.
- If the confidence level of the candidate hypothesis is low, the system compares the terms in the hypothesis to user correction data.
- The system corrects the terms in the hypothesis based on the user's previous corrections.
- The system also considers fulfillment data associated with the candidate hypothesis to determine whether to use the candidate hypothesis or the corrected hypothesis in responding to the user.
Potential applications of this technology:
- Automated assistants: This system can be used in automated assistants like voice-activated virtual assistants to improve speech recognition accuracy and provide more accurate responses to user queries.
- Transcription services: The system can be used in transcription services to correct any speech recognition errors and provide accurate transcriptions of spoken content.
Problems solved by this technology:
- Speech recognition errors: The system addresses the problem of incorrect speech recognition by using user corrections and fulfillment data to improve the accuracy of the recognized speech.
- Misinterpretation of user queries: By correcting speech recognition errors, the system can better understand user queries and provide more relevant and accurate responses.
Benefits of this technology:
- Improved user experience: By correcting speech recognition errors, the system can provide more accurate and relevant responses, leading to a better user experience.
- Time-saving: The system reduces the need for manual correction of speech recognition errors, saving time for both users and service providers.
- Enhanced accuracy: By utilizing user corrections and fulfillment data, the system improves the accuracy of speech recognition, leading to more precise and reliable results.
Original Abstract Submitted
Implementations relate to correcting a speech recognition hypothesis based on prior correction(s) made by a user and/or fulfillment data associated with fulfilling a request embodied in the speech recognition hypothesis. A candidate speech recognition hypothesis can be generated in response to the user providing a spoken utterance to an application, such as an automated assistant. When a confidence metric for the candidate speech recognition hypothesis does not satisfy a threshold, one or more terms of the candidate speech recognition hypothesis can be compared to correcting data. The correcting data can indicate whether the user previously corrected any term(s) present in the candidate speech recognition hypothesis and, if so, correct the term(s) accordingly. Fulfillment data generated for the candidate hypothesis and/or for the corrected hypothesis can also be processed to determine whether to utilize the candidate hypothesis or the corrected hypothesis in responding to the user.
