	US 11,681,916 B2
	Complex system for knowledge layout facilitated analytics-based action selection
Michael Thomas Giba, San Francisco, CA (US); Teresa Sheausan Tung, Tustin, CA (US); and Colin Anil Puri, Walnut Creek, CA (US)
Assigned to ACCENTURE GLOBAL SOLUTIONS LIMITED, Dublin (IE)
Filed by Accenture Global Solutions Limited, Dublin (IE)
Filed on Jul. 24, 2020, as Appl. No. 16/938,170.
Claims priority of provisional application 62/877,921, filed on Jul. 24, 2019.
Prior Publication US 2021/0027158 A1, Jan. 28, 2021
Int. Cl. G06N 3/08 (2023.01); G06N 3/04 (2023.01)

CPC G06N 3/08 (2013.01) [G06N 3/04 (2013.01)]

13 Claims

1. A method including:

at an input layer of an event resolution knowledge stack receiving previous challenge data associated with a previous challenge event:

determining a used analytics model type from among multiple available analytics model types for selection of an action for the previous challenge event;

determining a previous selected action for the previous challenge event, the previous selected action selected using a used analytics model type;

receiving resolution data responsive to execution of the previous selected action;

at knowledge mapping layer of the event resolution knowledge stack:

onboarding the previous challenge event to a knowledge layout by generating nodes on the knowledge layout, the knowledge layout comprising a collection of interconnected nodes representing mapping of resources, the nodes including:

a previous challenge event node to represent the previous challenge event;

a used analytics model type node to represent to the used analytics model type;

a previous selected action node to represent previous selected action; and

a resolution data node to represent the resolution data, wherein the nodes comprise pointers for data storage for information associated with each of the nodes;

building event and analytics models in parity by:

interconnecting the previous selected action node with the previous challenge event node on the knowledge layout;

interconnecting the used analytics model type node with the previous challenge event node; and

interconnecting the resolution data node with the previous challenge event node;

at the input layer, receiving active event data associated with an active challenge event;

at an analytics model layer of the event resolution knowledge stack:

traversing the knowledge layout to select an analytics model type by:

based on the active event data and the previous challenge data, determining an analytic correspondence between the active challenge event and the previous challenge event;

based on the analytic correspondence determining to implement the used analytics model type for selection of an action for the active challenge event; and

at a next action layer of the event resolution knowledge stack:

traversing the knowledge layout to implement a selected action by:

applying the resolution data to an analytics model of the used analytics model type to determine whether to implement the previous selected action for the active challenge event; and

when the previous selected action is determined to be implemented, controlling, with a processing resource, execution of a robotic task defined within the previous selected action, the robotic task representing automated command configured to implement the selected action;

generating a processing resource node representing the processing resource and adding the processing resource node to the knowledge layout.