System for logical rule induction of knowledge graph of engineering system
Systems and methods for logical rule formula induction of knowledge graphs for engineering system design include receiving a plurality of knowledge graphs for an engineering system. For the broken knowledge graph, the aggregated bundling search is limited by the edge of the node-of-interest connecti...
Saved in:
Main Authors | , , , |
---|---|
Format | Patent |
Language | Chinese English |
Published |
14.07.2023
|
Subjects | |
Online Access | Get full text |
Cover
Loading…
Summary: | Systems and methods for logical rule formula induction of knowledge graphs for engineering system design include receiving a plurality of knowledge graphs for an engineering system. For the broken knowledge graph, the aggregated bundling search is limited by the edge of the node-of-interest connection, and candidate formulas representing the respective edge found by the bundling search engine are generated, each formula being constrained by the requirements of at least two arguments of the defined formula chain length. A formula evaluation determines whether each candidate formula is valid. And selecting the formula with the highest ranking from the candidate formulas according to the defined standard. For well-connected graphs, a graph neural network is trained to predict a first class of query graphs and a second class of interference graphs. The anti-fact resolver engine solves the minimum number of edits of the query graph towards the interference graph, and converts the predicted query graph first class |
---|---|
Bibliography: | Application Number: CN202080106713 |