US 11,734,977 B2
Machine learning-based platform for user identification
Lucas Allen Budman, Denver, CO (US); Amitabh Agrawal, Superior, CO (US); Andrew Weber Spott, Boulder, CO (US); and Michael Ross Graf, Lakewood, CO (US)
Assigned to TruU, Inc., Palo Alto, CA (US)
Filed by TruU, Inc., Palo Alto, CA (US)
Filed on Oct. 6, 2022, as Appl. No. 17/961,517.
Application 17/961,517 is a continuation of application No. 17/351,108, filed on Jun. 17, 2021, granted, now 11,514,739.
Application 17/351,108 is a continuation of application No. 16/897,174, filed on Jun. 9, 2020, granted, now 11,069,165, issued on Jul. 20, 2021.
Application 16/897,174 is a continuation of application No. 16/550,112, filed on Aug. 23, 2019, granted, now 10,713,874, issued on Jul. 14, 2020.
Claims priority of provisional application 62/722,701, filed on Aug. 24, 2018.
Prior Publication US 2023/0033188 A1, Feb. 2, 2023
This patent is subject to a terminal disclaimer.
Int. Cl. H04N 5/44 (2011.01); G07C 9/22 (2020.01); G06N 20/00 (2019.01); G07C 9/28 (2020.01); H04L 9/40 (2022.01); H04W 12/64 (2021.01); H04W 12/63 (2021.01)
CPC G07C 9/22 (2020.01) [G06N 20/00 (2019.01); G07C 9/28 (2020.01); H04L 63/0272 (2013.01); H04W 12/63 (2021.01); H04W 12/64 (2021.01)] 20 Claims
OG exemplary drawing
 
1. A non-transitory computer-readable medium comprising stored computer-readable instructions that, when executed by one or more processors, cause the one or more processors to:
identify a target user requesting access to an operational context assigned an operational security threshold at a particular time, wherein the target user performs an action to request access;
access a machine-learned model trained based on training data representing the action of performed by the target user;
input data representing the action to the machine-learned model to generate an identity confidence value describing a probability that the action was performed by the target user;
determine a decayed identity confidence value by applying a set of decay parameters to the identity confidence value, wherein the set of decay parameters are determined based on the operational context and the decayed identity confidence value represents a confidence in an identity of the target user at the particular time; and
grant the target user access to the operational context responsive to determining the decayed identity confidence value to be greater than the operational security threshold.