US 12,271,498 B2
Graph-based data compliance using natural language text
Donald E. Johnson, Jr., Newton, NJ (US); Somadev Pasala, Montclair, NJ (US); Ravi Kondadadi, Lakeville, MN (US); Hadi D. Halim, Kendall Park, NJ (US); Ramin Anushiravani, San Carlos, CA (US); Ayush Tomar, Morgan Hill, CA (US); Adam Russell, Arlington, MA (US); and Robert K. Rossmiller, Brookfield, WI (US)
Assigned to Optum, Inc., Minnetonka, MN (US)
Filed by Optum, Inc., Minnetonka, MN (US)
Filed on Aug. 21, 2023, as Appl. No. 18/452,768.
Prior Publication US 2025/0068755 A1, Feb. 27, 2025
Int. Cl. G06F 21/62 (2013.01); G06F 16/28 (2019.01)
CPC G06F 21/6218 (2013.01) [G06F 16/288 (2019.01)] 20 Claims
OG exemplary drawing
 
1. A computer-implemented method comprising:
generating, by one or more processors and using a natural language model, entity-relationship data for an access-controlled dataset that identifies a data access constraint for a first dataset entity associated with the access-controlled dataset;
generating, by the one or more processors and based on the entity-relationship data, a knowledge graph that comprises a plurality of vertices and a plurality of edges, wherein (a) a vertex of the plurality of vertices identifies the first dataset entity of the access-controlled dataset and (b) an edge of the plurality of edges identifies a potential interaction between the first dataset entity and a second dataset entity associated with the access-controlled dataset;
identifying, by the one or more processors and using the knowledge graph, a data access condition associated with a data access violation or a data coverage violation;
identifying, by the one or more processors and using the knowledge graph, a violation severity associated with the data access condition based on a size of a graph cycle associated with the first dataset entity and the second dataset entity;
generating, by the one or more processors and using the knowledge graph, a natural language condition description based on the data access condition; and
providing, by the one or more processors and based on the violation severity, a condition alert that identifies the natural language condition description.