US 12,079,716 B2
Method and system for optimizing results of a function in a knowledge graph using neural networks
Pavlo Malynin, Menlo Park, CA (US); Gregory Kenneth Coulombe, Sherwood Park (CA); Sricharan Kallur Palli Kumar, Mountain View, CA (US); Cynthia Joann Osmon, Sunnyvale, CA (US); and Roger C. Meike, Redwood City, CA (US)
Assigned to INTUIT INC., Mountain View, CA (US)
Filed by INTUIT INC., Mountain View, CA (US)
Filed on Feb. 28, 2020, as Appl. No. 16/805,242.
Prior Publication US 2021/0271965 A1, Sep. 2, 2021
Int. Cl. G06F 40/30 (2020.01); G06N 3/042 (2023.01); G06N 3/047 (2023.01); G06N 3/08 (2023.01)
CPC G06N 3/08 (2013.01) [G06N 3/042 (2023.01); G06N 3/047 (2023.01)] 20 Claims
OG exemplary drawing
 
1. A method for optimizing results generated by functions executed using a rule-based knowledge graph, comprising:
generating a neural network based on a knowledge graph defining a function, wherein:
the knowledge graph includes a first plurality of nodes representing operations executed to perform the function and a second plurality of nodes representing inputs used to perform the operations, and
the generated neural network comprises a graph including the first plurality of nodes and the second plurality of nodes, and organized in an opposite direction from the knowledge graph such that an output of the knowledge graph comprises an input of the generated neural network;
receiving inputs to perform the function using the knowledge graph;
generating a result of the function based on the received inputs and the knowledge graph;
receiving a request to optimize the generated result of the function;
generating a loss function for the neural network;
adjusting values of parameters in the neural network to optimize the generated result based on the generated loss function and a gradient determination for the parameters in the neural network; and
outputting the adjusted values of the parameters in the neural network in response to the request to optimize the generated result of the function.