US Patent Application 17736959. EFFICIENT EMBEDDING FOR ACOUSTIC MODELS simplified abstract

From WikiPatents
Jump to navigation Jump to search

EFFICIENT EMBEDDING FOR ACOUSTIC MODELS

Organization Name

Apple Inc.


Inventor(s)

Daniel C. Klingler of San Jose CA (US)

Carlos M. Avendano of Campbell CA (US)

Jonathan Huang of Pleasanton CA (US)

Miquel Espi Marques of Cupertino CA (US)

EFFICIENT EMBEDDING FOR ACOUSTIC MODELS - A simplified explanation of the abstract

This abstract first appeared for US patent application 17736959 titled 'EFFICIENT EMBEDDING FOR ACOUSTIC MODELS

Simplified Explanation

This patent application describes a system and method for generating and storing learned embeddings of audio inputs on an electronic device.

  • The electronic device can encode and store audio inputs as well as the learned embeddings of these inputs.
  • When a new audio input is received, the device can compare its encoded version to the stored encoded versions of prior inputs.
  • If a match is found, the device can provide the corresponding learned embedding to a detection model on the device.
  • The stored embeddings can be used by locally trained models to detect individual sounds using electronic devices.


Original Abstract Submitted

The subject disclosure provides systems and methods for generating and storing learned embeddings of audio inputs to an electronic device. The electronic device may generate and store encoded versions of audio inputs and learned embeddings of the audio inputs. When a new audio input is obtained, the electronic device can generate an encoded version of the new audio input, compare the encoded version of the new audio input to the stored encoded versions of prior audio inputs, and if the encoded version of the new audio input matches one of the stored encoded versions of the prior audio inputs, the electronic device can provide a stored learned embedding that corresponds to the one of the stored encoded versions of the prior audio inputs to a detection model at the electronic device. The cached embeddings can be provided to locally trained models for detecting individual sounds using electronic devices.

Retrieved from "http://wikipatents.org/index.php?title=US_Patent_Application_17736959._EFFICIENT_EMBEDDING_FOR_ACOUSTIC_MODELS_simplified_abstract&oldid=6649"