US 12,271,832 B2
Artificial intelligence engine for directed hypothesis generation and ranking
Martin Stumpe, Chicago, IL (US); and Alena Harley, Chicago, IL (US)
Assigned to Tempus AI, Inc., Chicago, IL (US)
Filed by Tempus AI, Inc., Chicago, IL (US)
Filed on Feb. 14, 2022, as Appl. No. 17/651,002.
Claims priority of provisional application 63/149,093, filed on Feb. 12, 2021.
Prior Publication US 2022/0261668 A1, Aug. 18, 2022
Int. Cl. G06F 7/02 (2006.01); G06F 16/00 (2019.01); G06F 16/2457 (2019.01); G06F 16/28 (2019.01); G06N 3/045 (2023.01); G06N 5/02 (2023.01); G06N 5/04 (2023.01)
CPC G06N 5/041 (2013.01) [G06F 16/24578 (2019.01); G06F 16/284 (2019.01); G06N 3/045 (2023.01); G06N 5/02 (2013.01)] 30 Claims
OG exemplary drawing
 
1. A method, comprising:
providing a person-centric data storage including biomarker data entities specific to a person and comprising whole transcriptome-derived RNA data, the person-centric data storage also including one or more of clinical findings, disease state diagnostic data, or imaging-derived data, the person-centric data storage further including additional biomarker data and annotations associated with the additional biomarker data, the additional biomarker data related to one or more entities or relations in the person-centric data storage;
determining relationships between biomarker entities and outcomes events within the person-centric data storage;
referencing relationships between biomarker entities and outcomes events within a background knowledge data storage generated from a plurality of structured annotation data sources comprising two or more of scientific literature, guidelines, domain-specific knowledge bases, clinico-genomic data sets, or large-scale EHR data sets, the background knowledge data storage storing the biomarker entities and the outcomes events, as well as the respective relationships between the biomarker entities and the outcomes events;
generating ranked candidate relationships of observational significance between biomarker entities and outcomes events within the person-centric data storage, the generating comprising;
attributing a confidence score to each of the generated candidate relationships, the attributing step comprising, in part, translating the generated candidate relationships into queries against the referenced relationships within the background knowledge data storage and inputting the queries into an artificial intelligence engine configured to determine the confidence scores; and
ranking the generated candidate relationships based on the attributed confidence scores to classify the generated ranked candidate relationships according to predictive accuracy and clinical actionability, wherein the ranking includes decreasing a rank for the generated ranked candidate relationships that match referenced relationships in the background knowledge data storage and increasing a rank for the generated ranked candidate relationships reflecting previously unknown relationships; and
returning one or more of the ranked candidate relationships to a user.