US 11,990,135 B2
Methods and apparatus for hybrid speech recognition processing
Daniel Willett, Walluf (DE); Joel Pinto, Aachen (DE); and William F. Ganong, III, Brookline, MA (US)
Assigned to Microsoft Technology Licensing, LLC, Redmond, WA (US)
Filed by Microsoft Technology Licensing, LLC, Redmond, WA (US)
Filed on Feb. 9, 2021, as Appl. No. 17/171,513.
Application 17/171,513 is a continuation of application No. 15/403,762, filed on Jan. 11, 2017, granted, now 10,971,157.
Prior Publication US 2021/0166699 A1, Jun. 3, 2021
This patent is subject to a terminal disclaimer.
Int. Cl. G10L 15/30 (2013.01); G10L 15/00 (2013.01); G10L 15/18 (2013.01)
CPC G10L 15/30 (2013.01) [G10L 15/005 (2013.01); G10L 15/1815 (2013.01)] 20 Claims
OG exemplary drawing
 
8. A method comprising:
receiving, by a processor in a server, text and a confidence score associated with the text from a mobile electronic device remotely located from the processor, the text comprising a result of an embedded speech recognizer of the mobile electronic device performing speech recognition on audio data associated with the text;
evaluating, by the processor, the text received from the mobile electronic device;
determining, by the processor, that the confidence score associated with the text does not exceed a threshold, wherein the confidence score indicates a confidence in the accuracy of the speech recognition performed by the embedded speech recognizer of the mobile electronic device; and
requesting, from the mobile electronic device, the audio data associated with the text in response to determining that the confidence score does not exceed the threshold.