18063118. Voice-history Based Speech Biasing simplified abstract (Google LLC)
Contents
- 1 Voice-history Based Speech Biasing
- 1.1 Organization Name
- 1.2 Inventor(s)
- 1.3 Voice-history Based Speech Biasing - A simplified explanation of the abstract
- 1.4 Simplified Explanation
- 1.5 Potential Applications
- 1.6 Problems Solved
- 1.7 Benefits
- 1.8 Commercial Applications
- 1.9 Prior Art
- 1.10 Frequently Updated Research
- 1.11 Questions about Voice Query History
- 1.12 Original Abstract Submitted
Voice-history Based Speech Biasing
Organization Name
Inventor(s)
Mikhail Dektiarev of Zurich (CH)
Voice-history Based Speech Biasing - A simplified explanation of the abstract
This abstract first appeared for US patent application 18063118 titled 'Voice-history Based Speech Biasing
Simplified Explanation
The patent application describes a method of using voice query history to enhance speech recognition. By analyzing previous queries spoken by a user, the system can improve the accuracy of transcribing current queries.
- Receiving audio data of the current query spoken by the user.
- Generating a lattice of candidate hypotheses based on the audio data.
- Extracting n-grams from transcriptions of previous queries to create voice query history data.
- Using a biasing context model to generate a biasing context vector indicating the likelihood of n-grams appearing in the current query.
- Augmenting the lattice of candidate hypotheses with the biasing context vector.
- Determining a transcription for the current query based on the augmented lattice of candidate hypotheses.
Potential Applications
This technology can be applied in various fields such as virtual assistants, customer service call centers, and voice-controlled devices to enhance speech recognition accuracy and user experience.
Problems Solved
- Improves speech recognition accuracy by leveraging voice query history. - Reduces errors in transcribing user queries. - Enhances user interaction with voice-activated systems.
Benefits
- Increased accuracy in transcribing user queries. - Enhanced user experience with voice-controlled devices. - Improved efficiency in processing voice commands.
Commercial Applications
Title: Enhanced Speech Recognition Technology for Virtual Assistants and Voice-Controlled Devices This technology can be utilized in virtual assistants like Siri, Alexa, and Google Assistant, as well as in customer service call centers to improve speech recognition accuracy and streamline interactions with users.
Prior Art
Further research can be conducted in the field of natural language processing, machine learning, and speech recognition technologies to explore similar methods of utilizing voice query history for improving speech recognition systems.
Frequently Updated Research
Stay updated on advancements in natural language processing, machine learning algorithms, and voice recognition technologies to enhance the performance of speech recognition systems.
Questions about Voice Query History
How does voice query history data improve speech recognition accuracy?
Voice query history data provides insights into the user's speech patterns and commonly used phrases, allowing the system to make more accurate predictions about the current query.
What are the potential privacy concerns associated with storing voice query history data?
Storing voice query history data raises privacy concerns regarding the security and confidentiality of user information. Implementing robust data protection measures is essential to address these concerns.
Original Abstract Submitted
A method of using voice query history to improve speech recognition includes receiving audio data corresponding to a current query spoken by a user and processing the audio data to generate a lattice of candidate hypotheses. The method also includes obtaining voice query history data associated with the user that includes n-grams extracted from transcriptions of previous queries spoken by the user, and generating, using a biasing context model configured to receive the voice query history data, a biasing context vector. The biasing context vector indicates a likelihood that each n-gram from the n-grams extracted from the transcriptions of the previous queries spoken by the user will appear in the current query. The method also includes augmenting the lattice of candidate hypotheses based on the biasing context vector and determining a transcription for the current query based on the augmented lattice of candidate hypotheses.
