18152749. Speech Recognition Biasing simplified abstract (GOOGLE LLC)
Contents
Speech Recognition Biasing
Organization Name
Inventor(s)
Olawale Abiri of Westfield NJ (US)
Dharmeshkumar Jayantilal Mokani of Milpitas CA (US)
Speech Recognition Biasing - A simplified explanation of the abstract
This abstract first appeared for US patent application 18152749 titled 'Speech Recognition Biasing
The abstract describes a method for biasing speech recognition based on context data, where a speech recognition request includes audio data and configuration parameters for biasing a model.
- Receiving speech recognition requests from an application on a user device
- Processing audio data using a speech recognition model to generate scores for speech elements
- Determining context scores for speech elements based on configuration parameters and context data
- Biasing speech recognition scores using context scores
- Determining the transcription of the utterance based on the biased speech recognition scores
Potential Applications: - Improving accuracy of speech recognition in various applications such as virtual assistants, dictation software, and customer service chatbots
Problems Solved: - Addressing bias in speech recognition models by incorporating context data to improve transcription accuracy
Benefits: - Enhanced user experience with more accurate speech recognition results - Increased efficiency in voice-controlled applications
Commercial Applications: - This technology can be utilized in industries such as healthcare, telecommunications, and automotive for voice-controlled systems and applications.
Questions about Speech Recognition Biasing: 1. How does biasing speech recognition scores using context data improve transcription accuracy? 2. What are the potential limitations or challenges of implementing this method in real-world applications?
Frequently Updated Research: - Stay updated on advancements in speech recognition technology and the integration of context data for biasing models.
Original Abstract Submitted
A method for speech recognition biasing includes receiving, from an application executing on a user device, at a speech service interface, a speech recognition request requesting a transcription of an utterance. The speech recognition request includes audio data encoding the utterance and configuration parameters for biasing a speech recognition model based on context data. The method includes processing, using the speech recognition model, the audio data to generate speech recognition scores for speech elements and determining context scores for the speech elements based on the configuration parameters and the context data. The method includes biasing the speech recognition scores using the context scores. The method also includes determining the transcription for the utterance based on the biased speech recognition scores.
