18610233. KEY PHRASE SPOTTING simplified abstract (Google LLC)

KEY PHRASE SPOTTING

Organization Name

Google LLC

Inventor(s)

Wei Li of Mountain View CA (US)

Rohit Prakash Prabhavalkar of Santa Clara CA (US)

Kanury Kanishka Rao of Santa Clara CA (US)

Yanzhang He of Mountain View CA (US)

Ian C. Mcgraw of Menlo Park CA (US)

Anton Bakhtin of New York NY (US)

KEY PHRASE SPOTTING - A simplified explanation of the abstract

This abstract first appeared for US patent application 18610233 titled 'KEY PHRASE SPOTTING

The patent application describes methods, systems, and apparatus for detecting utterances of a key phrase in an audio signal using an attention mechanism and neural network layers.

• Receiving an audio signal encoding one or more utterances.
• Generating an attention output using an attention mechanism based on encodings from neural network layers.
• Outputting whether the audio signal likely encodes the key phrase.
• Providing the output indicating the likelihood of the key phrase in the audio signal.

Potential Applications: - Speech recognition technology - Voice-activated devices - Security systems for keyword detection

Problems Solved: - Efficiently detecting specific key phrases in audio signals - Enhancing accuracy in speech recognition systems

Benefits: - Improved accuracy in detecting key phrases - Enhanced performance in voice-activated systems - Increased security in keyword detection applications

Commercial Applications: - Integration into smart home devices - Implementation in security systems for keyword detection - Utilization in speech-to-text transcription services

Questions about Key Phrase Detection: 1. How does the attention mechanism improve the accuracy of key phrase detection in audio signals? 2. What are the potential limitations of using neural network layers in this technology?

Frequently Updated Research: - Stay updated on advancements in neural network technology for key phrase detection in audio signals.

Original Abstract Submitted

Methods, systems, and apparatus, including computer programs encoded on computer storage media, for detecting utterances of a key phrase in an audio signal. One of the methods includes receiving, by a key phrase spotting system, an audio signal encoding one or more utterances; while continuing to receive the audio signal, generating, by the key phrase spotting system, an attention output using an attention mechanism that is configured to compute the attention output based on a series of encodings generated by an encoder comprising one or more neural network layers; generating, by the key phrase spotting system and using attention output, output that indicates whether the audio signal likely encodes the key phrase; and providing, by the key phrase spotting system, the output that indicates whether the audio signal likely encodes the key phrase.