Google LLC (20240290333). PROVIDING PRE-COMPUTED HOTWORD MODELS simplified abstract
Contents
PROVIDING PRE-COMPUTED HOTWORD MODELS
Organization Name
Inventor(s)
Matthew Sharifi of Kilchberg (CH)
PROVIDING PRE-COMPUTED HOTWORD MODELS - A simplified explanation of the abstract
This abstract first appeared for US patent application 20240290333 titled 'PROVIDING PRE-COMPUTED HOTWORD MODELS
The patent application describes methods, systems, and apparatus for obtaining audio data for multiple users speaking specific words or sub-words, training pre-computed hotword models based on the audio data, receiving candidate hotwords, identifying corresponding pre-computed hotword models, and providing these models to a computing device.
- Audio data is collected for multiple users speaking specific words or sub-words.
- Pre-computed hotword models are trained based on the collected audio data.
- Candidate hotwords are received from a computing device.
- Corresponding pre-computed hotword models are identified for the candidate hotwords.
- The identified pre-computed hotword models are provided to the computing device.
Potential Applications: - Speech recognition systems - Voice-controlled devices - Language learning tools
Problems Solved: - Improving accuracy in recognizing specific words or sub-words - Enhancing user experience with voice-activated technology
Benefits: - Increased efficiency in speech recognition - Enhanced user interaction with devices - Personalized voice recognition capabilities
Commercial Applications: Title: Advanced Speech Recognition Technology for Voice-Controlled Devices This technology can be utilized in smart speakers, virtual assistants, language learning apps, and customer service chatbots to enhance user experience and improve accuracy in speech recognition.
Questions about the technology: 1. How does this technology improve the accuracy of speech recognition systems? 2. What are the potential limitations of using pre-computed hotword models in voice-controlled devices?
Original Abstract Submitted
methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for obtaining, for each of multiple words or sub-words, audio data corresponding to multiple users speaking the word or sub-word; training, for each of the multiple words or sub-words, a pre-computed hotword model for the word or sub-word based on the audio data for the word or sub-word; receiving a candidate hotword from a computing device; identifying one or more pre-computed hotword models that correspond to the candidate hotword; and providing the identified, pre-computed hotword models to the computing device.
