Back to results
US 11,636,123 B2Granted
Density-based computation for information discovery in knowledge graphs
AssigneeAccenture Global Solutions Limited
InventorsFreddy Lecue, Chahrazed Bouhini, Jeremiah Hayes, Mykhaylo Zayats, Nicholas McCarthy, Qurrat Ul Ain
Life of this patent
Estimating the term…
Abstract
Knowledge graph systems are disclosed for enhancing a knowledge graph by generating a new node. The knowledge graph system converts a knowledge graph into an embedding space, and selects a region of interest from within the embedding space. The knowledge graph system further identifies, from the region of interest, one or more gap regions, and calculates a center for each gap region. A node is generated for each gap region, and the information represented by the node is added to the original knowledge graph to generate an updated knowledge graph.
Claims18
A system comprising:
- a hardware memory for storing instructions;
- and
- a hardware processor in communication with the hardware memory and configured to execute the instructions to:
- receive an original knowledge graph including a set of structured data;
- convert the set of structured data of the original knowledge graph to nodes in an embedding space;
- determine a region of interest within the embedding space representing a region of a common concept in the embedding space;
- determine a gap region within the region of interest, the gap region lacking corresponding structured data in the original knowledge graph;
- determine a location of a center of the gap region within the region of interest in the embedding space based on locations of nodes corresponding to structured data in the original knowledge graph within the region of interest but outside of the gap region in the embedding space;
- create a center node at the location of the center of the gap region in the embedding space, the center node incorporating ingredients found from surrounding existing nodes in the embedding space and the original knowledge graph;
- create a new structured data corresponding to the center node, the new structured data including the ingredients found from the original knowledge graph corresponding to the surrounding existing nodes of the center node in the embedding space;
- and
- reconstruct relationships associated with the center node in the embedding space by finding links of the center node in the embedding space with existing nodes in the embedding space and updating the original knowledge graph by adding the new structured data generated corresponding to the center node and links in the embedding space.
9 more elements
A method comprising:
- receiving, by a knowledge graph reception circuitry, an original knowledge graph including a set of structured data;
- converting, by a knowledge graph embedding circuitry, the set of structured data of the original knowledge graph to nodes in an embedding space;
- determining, by a region identification circuitry, a region of interest within the embedding space representing a region of a common concept in the embedding space;
- determining, by the region identification circuitry, a gap region within the region of interest, the gap region lacking corresponding structured data in the original knowledge graph;
- determining, by a computation circuitry, a location of a center of the gap region within the region of interest in the embedding space based on locations of nodes corresponding to structured data in the original knowledge graph within the region of interest but outside of the gap region in the embedding space;
- creating a center node at the location of the center of the gap region in the embedding space, the center node incorporating ingredients found from surrounding existing nodes in the embedding space and the original knowledge graph;
- creating a new structured data corresponding to the center node, the new structured data including the ingredients found from the original knowledge graph corresponding to the surrounding existing nodes of the center node in the embedding space;
- discovering relationships between the center node and other nodes in the embedding space;
- and
- enhancing, by a reconstruction circuitry, the original knowledge graph into an updated knowledge graph by adding the new structured data corresponding to the center node and the relationships.
7 more elements
A product comprising:
- a machine-readable medium, other than a transitory signal;
- and
- instructions stored on the machine-readable medium, the instructions configured to, when executed, cause processing circuitryto:
- receive an original knowledge graph including a set of structured data;
- convert the set of structured data of the original knowledge graph to nodes in an embedding space;
- determine a region of interest within the embedding space representing a region of a common concept in the embedding space;
- determine a gap region within the region of interest, the gap region lacking corresponding structured data in the original knowledge graph;
- determine a location of a center of the gap region within the region of interest in the embedding space based on locations of nodes corresponding to structured data in the original knowledge graph within the region of interest but outside of the gap region in the embedding space;
- create a center node at the location of the center of the gap region in the embedding space, the center node incorporating ingredients found from surrounding existing nodes in the embedding space and the original knowledge graph;
- create a new structured data corresponding to the center node, the new structured data including the ingredients found from the original knowledge graph corresponding to the surrounding existing nodes of the center node in the embedding space;
- discover relationships between the center node and other nodes in the embedding space;
- and
- reconstruct the center node by finding links with existing nodes from the original knowledge graph and updating the original knowledge graph by adding the new structured data corresponding to the center node and links in the embedding space.
10 more elements
Dependencies are inferred from the claim text.