View source for 20240028868. SYSTEM FOR LOGIC RULE INDUCTION ON KNOWLEDGE GRAPHS OF ENGINEERING SYSTEMS simplified abstract (Siemens Aktiengesellschaft)

=SYSTEM FOR LOGIC RULE INDUCTION ON KNOWLEDGE GRAPHS OF ENGINEERING SYSTEMS=


==Organization Name==

[[:Category:Siemens Aktiengesellschaft|Siemens Aktiengesellschaft]]


[[Category:Siemens Aktiengesellschaft]]

==Inventor(s)==

[[:Category:Arun Ramamurthy of Plainsboro NJ (US)|Arun Ramamurthy of Plainsboro NJ (US)]][[Category:Arun Ramamurthy of Plainsboro NJ (US)]]

[[:Category:Yuyu Zhang of Atlanta GA (US)|Yuyu Zhang of Atlanta GA (US)]][[Category:Yuyu Zhang of Atlanta GA (US)]]

[[:Category:Sanjeev Srivastava of Chantilly VA (US)|Sanjeev Srivastava of Chantilly VA (US)]][[Category:Sanjeev Srivastava of Chantilly VA (US)]]

[[:Category:Livio Dalloro of Plainsboro NJ (US)|Livio Dalloro of Plainsboro NJ (US)]][[Category:Livio Dalloro of Plainsboro NJ (US)]]

==SYSTEM FOR LOGIC RULE INDUCTION ON KNOWLEDGE GRAPHS OF ENGINEERING SYSTEMS - A simplified explanation of the abstract==


This abstract first appeared for US patent application 20240028868 titled 'SYSTEM FOR LOGIC RULE INDUCTION ON KNOWLEDGE GRAPHS OF ENGINEERING SYSTEMS

==Simplified Explanation==

The patent application describes a system and method for inducing logic rule formulas on knowledge graphs for engineering system designs. Here is a simplified explanation of the abstract:

* The system receives multiple knowledge graphs for an engineering system.
* For disconnected knowledge graphs, an agglomerative beam search is used to find edges connected to a specific node of interest.
* Candidate formulas are generated to represent each found edge, with each formula having at least two arguments for a defined formula chain length.
* Formula evaluation determines the validity of each candidate formula.
* The top-ranked formulas are selected based on defined criteria.
* For well-connected graphs, a graph neural network is trained to predict the first class for a query graph and the second class for a distractor graph.
* A counterfactual solver engine solves for the minimum number of edits needed to transform the predicted first class of the query graph to the predicted second class of the distractor graph.

Potential applications of this technology:

* Engineering system design: The system can be used to generate logic rule formulas for designing engineering systems based on knowledge graphs.
* Knowledge graph analysis: The system can analyze knowledge graphs to identify relevant edges and generate formulas representing the relationships between nodes.

Problems solved by this technology:

* Complex system design: The system simplifies the process of designing complex engineering systems by automatically generating logic rule formulas.
* Knowledge graph analysis: The system helps in analyzing and understanding the relationships between nodes in knowledge graphs.

Benefits of this technology:

* Efficiency: The system automates the process of generating logic rule formulas, saving time and effort in engineering system design.
* Accuracy: By using a combination of beam search and formula evaluation, the system ensures that only valid formulas are selected.
* Insightful analysis: The system provides insights into the relationships within knowledge graphs, aiding in decision-making and problem-solving.


==Original Abstract Submitted==

system and method for logic rule formula induction on knowledge graphs for engineering system designs include receiving plurality of knowledge graphs for an engineering system. for a disconnected knowledge graph, agglomerative beam search is constrained to edges connected from node of interest, and candidate formulas are generated representing a respective edge found by the beam search engine, each formula constrained by a requirement of at least two arguments for defined formula chain length. formula evaluation establishes whether each candidate formula is valid. top ranked formulas are selected from the candidate formulas according to defined criteria. for well-connected graphs, a graph neural network is trained to predict first class for a query graph and second class for distractor graph. counterfactual solver engine solves for minimum number of edits to query graph toward distractor graph which transforms predicted first class of the query graph to predicted second class.



[[Category:G06N3/042]]