US 12,334,071 B2
Decaying automated speech recognition processing results
Matthew Sharifi, Kilchberg (CH); and Victor Carbune, Zürich (CH)
Assigned to Google LLC, Mountain View, CA (US)
Filed by Google LLC, Mountain View, CA (US)
Filed on Apr. 26, 2023, as Appl. No. 18/307,736.
Application 18/307,736 is a continuation of application No. 17/111,467, filed on Dec. 3, 2020, granted, now 11,676,594.
Prior Publication US 2024/0096320 A1, Mar. 21, 2024
This patent is subject to a terminal disclaimer.
Int. Cl. G10L 15/22 (2006.01); G10L 25/78 (2013.01)
CPC G10L 15/22 (2013.01) [G10L 25/78 (2013.01)] 20 Claims
OG exemplary drawing
 
1. A computer-implemented method when executed by data processing hardware causes the data processing hardware to perform operations comprising:
receiving audio data corresponding to an utterance and captured by an open microphone;
generating, by processing the audio data using a first level of automated speech recognition (ASR) processing, a first-pass speech recognition result for the audio data;
determining a confidence level for the first-pass speech recognition result;
based on the confidence level for the first-pass speech recognition result, generating, by processing the audio data using a second level of ASR processing, a second-pass speech recognition result for the audio data, the second level of ASR processing greater than the first level of ASR processing, the second level of ASR processing decaying over time while processing the audio data using the second level of ASR processing; and
when the second level of ASR processing is equal to zero, instructing the microphone to close.