US 11,681,726 B2
System for generating specialized phenotypical embedding
Mohamed Ghalwash, Westchester, NY (US); Zijun Yao, Ridgewood, NJ (US); Prithwish Chakraborty, New York, NY (US); James V Codella, Danbury, CT (US); and Daby Mousse Sow, Croton on Hudson, NY (US)
Assigned to INTERNATIONAL BUSINESS MACHINES CORPORATION, Armonk, NY (US)
Filed by International Business Machines Corporation, Armonk, NY (US)
Filed on Dec. 3, 2020, as Appl. No. 17/110,954.
Prior Publication US 2022/0179880 A1, Jun. 9, 2022
Int. Cl. G06F 17/00 (2019.01); G06F 16/28 (2019.01); G06N 20/00 (2019.01); G16H 10/60 (2018.01)
CPC G06F 16/284 (2019.01) [G06N 20/00 (2019.01); G16H 10/60 (2018.01)] 21 Claims
OG exemplary drawing
 
1. A system, comprising:
a memory that stores computer executable components; and
a processor that executes the computer executable components stored in memory, wherein the computer executable components comprise:
a training component that trains a structural pheno-embedding learning model, using a hierarchical knowledge graph of phenotypes and a set of patient data mapped to the hierarchical knowledge graph, to perform cohort expansion based on relationships of the phenotypes of the hierarchical knowledge graph;
a data augmentation component that expands, using the structural pheno-embedding learning model and the hierarchical knowledge graph, a sparse data set of client's patient data of a selected cohort based on a client's cohort selection criteria comprising a phenotype to generate an augmented data set comprising an expanded cohort using a subset of the set of patient data; and
an embedding component that generates, using the structural pheno-embedding learning model and the augmented data set, a filtered graph from the hierarchical knowledge graph, wherein the filtered graph comprises specialized embeddings for a subset of the phenotypes of the hierarchical knowledge graph associated with the phenotype.