| CPC G06N 5/02 (2013.01) [G06F 18/214 (2023.01); G06F 18/22 (2023.01); G06F 18/25 (2023.01); G06F 18/29 (2023.01); G06N 20/00 (2019.01); G06F 18/295 (2023.01); G06N 5/022 (2013.01)] | 18 Claims |
|
1. A computer-based method for aligning knowledge graphs, the method comprising:
assigning, by an electronic controller, a subgraph type to each node of a plurality of nodes in a first knowledge graph and a second knowledge graph, wherein the first knowledge graph and the second knowledge graph each include a plurality of nodes each coupled to one or more other nodes by a plurality of edges, wherein the assigned subgraph type to a node is defined by at least one selected from a group consisting of a set of labels of one or more edges extending from the node and a set of labels of other nodes to which the node is coupled;
performing, by the electronic controller, an initial matching to the first knowledge graph and the second knowledge graph to identify a plurality of candidate node-node pairs, wherein each candidate node-node pair of the plurality of candidate node-node pairs includes a node from the first knowledge graph and a node from the second knowledge graph;
identifying a first candidate node-node pair as a first valid node-node mapping;
identifying a second candidate node-node pair as a second valid node-node mapping based at least in part on a determination that a subgraph type combination of the second candidate node-node pair matches a subgraph type combination of the first candidate node-node pair, wherein the subgraph type combination of the first candidate node-node pair includes the subgraph type assigned to the node from the first knowledge graph and the subgraph type assigned to the node from the second knowledge graph in the candidate node-node pair;
connecting the first knowledge graph and the second knowledge graph by adding a new subgraph type node associated with the subgraph type and adding a plurality of new edges coupling the new subgraph type node to a first node assigned the subgraph type in the first knowledge graph and to a second node assigned the subgraph type in the second knowledge graph;
receiving a natural language inquiry via a user input;
generating a natural language output based on the natural language input and the new subgraph type node; and,
outputting the natural language output via a display screen or a speaker.
|