US 12,149,492 B2
Engaging unknowns in response to interactions with knowns
Sudeshna Banerjee, Waxhaw, NC (US); and Paul Gerard Mistor, Winston-Salem, NC (US)
Assigned to TRUIST BANK, Charlotte, NC (US)
Filed by Truist Bank, Charlotte, NC (US)
Filed on Apr. 29, 2022, as Appl. No. 17/661,413.
Prior Publication US 2023/0353524 A1, Nov. 2, 2023
Int. Cl. H04L 51/214 (2022.01); G06Q 30/0204 (2023.01); G06Q 30/0251 (2023.01); H04L 51/226 (2022.01)
CPC H04L 51/214 (2022.05) [G06Q 30/0204 (2013.01); G06Q 30/0254 (2013.01); H04L 51/226 (2022.05)] 20 Claims
OG exemplary drawing
 
1. A system for improving distributed network data flow efficiency by using a machine learning model to engage a reduced set of unknown objects, the system comprising:
at least one processor;
a communication interface communicatively coupled to the at least one processor; and
a memory device storing executable code that, when executed, causes the at least one processor to:
receive interaction data corresponding to interactions between known objects and unknown objects and comprising interactors and interaction descriptions;
develop an interaction map from the interaction data, wherein the interaction map comprises unknown objects, known objects, and connections between unknown objects and known objects;
develop a profile for each unknown object, wherein the profile comprises interactions and relationships with known objects;
train, via machine learning and using a set of training data, a machine learning model configured to identify a reduced set of unknown objects most likely to reciprocate, the training including:
iteratively predicting which of the unknown objects, based on the profile for each unknown object, which of the unknown objects are most likely to reciprocate, the predicting being based on at least one output category;
testing and comparing the unknown objects predicted during each iteration against a target variable; and
indicating, via a feedback loop, for each iteration whether modifications to weights assigned to certain profile data for each of the unknown objects are necessary to improve predictability of the target variable:
deploy the trained machine learning model to generate a reduced set of unknown objects most likely to reciprocate, and based thereon determining a reduced set of unknown objects most likely to reciprocate and, pull properties of each unknown object in the reduced set from a database storing the profiles of each unknown object; and
trigger a communication, based on the profiles of each unknown object, to each unknown object of the reduced set of unknown objects.