US 12,080,276 B2
Adapting automated speech recognition parameters based on hotword properties
Matthew Sharifi, Kilchberg (CH); and Aleksandar Kracun, New York, NY (US)
Assigned to Google LLC, Mountain View, CA (US)
Filed by Google LLC, Mountain View, CA (US)
Filed on Mar. 22, 2023, as Appl. No. 18/188,238.
Application 18/188,238 is a continuation of application No. 17/120,033, filed on Dec. 11, 2020, granted, now 11,620,990.
Prior Publication US 2023/0223014 A1, Jul. 13, 2023
This patent is subject to a terminal disclaimer.
Int. Cl. G10L 15/06 (2013.01); G10L 15/16 (2006.01); G10L 15/22 (2006.01); G10L 15/28 (2013.01); G10L 25/90 (2013.01); G10L 15/08 (2006.01); G10L 25/78 (2013.01)
CPC G10L 15/06 (2013.01) [G10L 15/16 (2013.01); G10L 15/22 (2013.01); G10L 15/28 (2013.01); G10L 25/90 (2013.01); G10L 2015/088 (2013.01); G10L 2025/783 (2013.01)] 20 Claims
OG exemplary drawing
 
1. A computer-implemented method executed on data processing hardware that causes the data processing hardware to perform operations comprising:
receiving a first acoustic segment characterizing a hotword detected by a hotword detector in streaming audio captured by a user device;
without performing speech recognition processing on the streaming audio, extracting one or more hotword attributes from the first acoustic segment;
adjusting, based on the one or more hotword attributes extracted from the first acoustic segment, one or more speech recognition parameters of an automated speech recognition (ASR) model; and
after adjusting the speech recognition parameters of the ASR model, processing, using the ASR model, a second acoustic segment to generate a speech recognition result, the second acoustic segment characterizing a spoken query/command that follows the first acoustic segment in the streaming audio captured by the user device.