US 11,782,496 B2
Smart context subsampling on-device system
Alexander Varshavsky, Summit, NJ (US); Bhaskar Mehta, Menlo Park, CA (US); Brian Coopersmith, Millburn, NJ (US); and Giovanni Botta, Brooklyn, NY (US)
Assigned to Google LLC, Mountain View, CA (US)
Filed by Google LLC, Mountain View, CA (US)
Filed on Oct. 27, 2022, as Appl. No. 17/975,042.
Application 17/975,042 is a continuation of application No. 16/496,225, granted, now 11,507,172, previously published as PCT/US2017/068923, filed on Dec. 29, 2017.
Prior Publication US 2023/0050146 A1, Feb. 16, 2023
This patent is subject to a terminal disclaimer.
Int. Cl. G06F 1/3234 (2019.01); G06N 20/00 (2019.01); G06F 1/3209 (2019.01); H04W 8/10 (2009.01); H04W 60/04 (2009.01)
CPC G06F 1/3234 (2013.01) [G06F 1/3209 (2013.01); G06N 20/00 (2019.01); H04W 8/10 (2013.01); H04W 60/04 (2013.01); H04M 2242/30 (2013.01)] 20 Claims
OG exemplary drawing
 
1. A method for reporting by a mobile computing device, the method comprising:
detecting, by the mobile computing device, an activity associated with the mobile computing device;
determining whether the detected activity corresponds to one or more trigger conditions in a first policy for reporting information, wherein the first policy is based on training data against a cost function associated with one or more activities; and
reporting information when the detected activity corresponds to the one or more trigger conditions, wherein the information is used for updating the first policy by updating the cost function.