US 12,367,191 B1
Systems and methods for searching database structures using semantically and categorically similar queries
Siddharth Subhash Shenoy, Jersey City, NJ (US); Omar Bari, Brownstown, MI (US); Christian Aaron Hartman, East Greenwich, RI (US); Santosh Vinod Chapaneri, Maharashtra (IN); Atul Ujjwal, Pune (IN); and Dishu Jindal, Waltham, MA (US)
Assigned to UpToDate, Inc.
Filed by UpToDate, Inc., Waltham, MA (US)
Filed on Feb. 29, 2024, as Appl. No. 18/591,607.
Int. Cl. G06F 16/242 (2019.01); G06F 16/2455 (2019.01); G06F 16/248 (2019.01)
CPC G06F 16/2428 (2019.01) [G06F 16/24552 (2019.01); G06F 16/248 (2019.01)] 20 Claims
OG exemplary drawing
 
19. A computer-implemented method for searching a database structure using semantically and categorically similar queries and a hybrid retrieval process, comprising:
training at least one semantic similarity model configured to identify a set of semantically similar queries to a received input query based on semantic similarity criteria;
training at least one hybrid retrieval model configured to apply a hybrid retrieval search process based on one or more search terms;
receiving, at a first user interface, a user query in response to a user query request, wherein the user query includes one or more search terms;
determining, by the at least one semantic similarity model and the user query, one or more semantically similar search queries meeting the semantic similarity criteria;
selecting, using a query parser, one or more search terms of the user query, wherein the one or more search terms satisfy search term criteria including a query category;
searching, using the at least one hybrid retrieval model and based on the one or more search terms of the user query, for information related to the one or more search terms from one or more databases;
generating, based on the information related to the one or more search terms, an output to the user query, wherein the output to the user query includes:
at least a portion of the information related to the one or more search terms; and
one or more citations associated with the information related to the one or more search terms from the one or more databases; and
displaying, at the first user interface, the output to the user query.