US 12,230,251 B2
Language model biasing modulation
Pedro J. Moreno Mengibar, Jersey City, NJ (US); and Petar Aleksic, Jersey City, NJ (US)
Assigned to Google LLC, Mountain View, CA (US)
Filed by Google LLC, Mountain View, CA (US)
Filed on Dec. 12, 2022, as Appl. No. 18/064,917.
Application 18/064,917 is a continuation of application No. 16/896,779, filed on Jun. 9, 2020, granted, now 11,532,299.
Application 16/896,779 is a continuation of application No. 16/381,167, filed on Apr. 11, 2019, granted, now 10,714,075, issued on Jul. 14, 2020.
Application 16/381,167 is a continuation of application No. 15/874,075, filed on Jan. 18, 2018, granted, now 10,297,248, issued on May 21, 2019.
Application 15/874,075 is a continuation of application No. 15/263,714, filed on Sep. 13, 2016, granted, now 9,886,946, issued on Feb. 6, 2018.
Application 15/263,714 is a continuation of application No. 14/673,731, filed on Mar. 30, 2015, granted, now 9,460,713, issued on Oct. 4, 2016.
Prior Publication US 2023/0109903 A1, Apr. 13, 2023
Int. Cl. G10L 15/00 (2013.01); G10L 15/07 (2013.01); G10L 15/183 (2013.01); G10L 15/197 (2013.01); G10L 15/24 (2013.01)
CPC G10L 15/07 (2013.01) [G10L 15/183 (2013.01); G10L 15/197 (2013.01); G10L 15/24 (2013.01)] 16 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 audio data corresponding to a voice query spoken by a user;
processing the audio data to determine a particular context associated with a type of words used in the voice query, the particular context associated with biasing an automated speech recognition (ASR) module;
biasing, based on the type of words used in the voice query, the ASR module to increase a likelihood of recognizing biasing terms from a set of biasing terms in the audio data;
receiving additional audio data corresponding to an utterance spoken by the user;
determining, that, when the additional audio data was received, the particular context associated with biasing the ASR module is no longer applicable; and
based on determining that the particular context associated with biasing the ASR module is no longer applicable, returning the biased ASR module to a baseline state.