US 11,727,919 B2
Memory allocation for keyword spotting engines
Hubert de Taffanel de La Jonquière, Paris (FR)
Assigned to Sonos, Inc., Santa Barbara, CA (US)
Filed by Sonos, Inc., Santa Barbara, CA (US)
Filed on May 19, 2021, as Appl. No. 17/303,066.
Claims priority of provisional application 63/027,500, filed on May 20, 2020.
Prior Publication US 2021/0366466 A1, Nov. 25, 2021
Int. Cl. G10L 15/00 (2013.01); G10L 15/16 (2006.01); G10L 15/22 (2006.01); G06N 3/04 (2023.01); G06N 3/08 (2023.01); G10L 15/02 (2006.01); G10L 15/08 (2006.01)
CPC G10L 15/16 (2013.01) [G06N 3/04 (2013.01); G06N 3/08 (2013.01); G10L 15/02 (2013.01); G10L 15/22 (2013.01); G10L 2015/088 (2013.01)] 20 Claims
OG exemplary drawing
 
1. A device configured to detect keywords in sound samples, the device comprising:
one or more microphones;
one or more processors;
a dynamic random-access memory block; and
a computer-readable memory storing instructions that, when executed by the one or more processors, cause the device to perform operations comprising:
capturing a plurality of sound samples via the one or more microphones;
extracting features from the plurality of sound samples, the extraction comprising:
storing the sound samples in a first portion of the memory block;
performing first computations based on spectral coefficients of the sound samples using a second portion of the memory block; and
storing results of the first computations as extracted features in a third portion of the memory block;
designating as temporary memory the second and third portions of the memory block; and
processing the extracted features utilizing a neural network, the processing comprising:
storing the extracted features in a fourth portion of the memory block;
performing second computations on the extracted features using the temporary memory, the second computations comprising computing at least one layer of the neural network; and
storing an output of the neural network as a classification in the temporary memory.