US 12,326,891 B2
Generating a product ontology based upon queries in a search engine log
Manoj Kumar Agarwal, Hyderabad (IN); and Jatin Khurana, Uttar Pradesh (IN)
Assigned to MICROSOFT TECHNOLOGY LICENSING, LLC, Redmond, WA (US)
Filed by Microsoft Technology Licensing, LLC, Redmond, WA (US)
Filed on Nov. 22, 2021, as Appl. No. 17/533,036.
Claims priority of provisional application 63/250,994, filed on Sep. 30, 2021.
Prior Publication US 2023/0118171 A1, Apr. 20, 2023
Int. Cl. G06F 40/205 (2020.01); G06F 16/35 (2019.01); G06F 16/36 (2019.01); G06F 18/231 (2023.01); G06F 40/289 (2020.01); G06F 40/40 (2020.01)
CPC G06F 16/35 (2019.01) [G06F 16/367 (2019.01); G06F 18/231 (2023.01); G06F 40/40 (2020.01)] 20 Claims
OG exemplary drawing
 
1. A computing system that is configured to construct a computer-implemented product ontology based upon queries submitted to a search engine, the computing system comprising:
a processor; and
memory storing instructions that, when executed by the processor, cause the processor to perform acts comprising:
obtaining a plurality of queries submitted to the search engine, wherein a query of the plurality of queries comprises a sequence of terms;
identifying, based upon a sequence of terms of a first query, a first phrase and a second phrase;
extracting the second phrase from the first phrase, wherein the second phrase includes fewer terms than the first phrase and is at an end of the first phrase; and
constructing the computer-implemented product ontology based upon an occurrence of the first phrase and an occurrence of the second phrase within the plurality of queries, wherein the product ontology comprises a first product class represented by the first phrase and a second product class represented by the second phrase, wherein the second product class is a child of the first product class in the product ontology based upon the first phrase being the ending phrase of the second phrase;
responsive to receiving an indication that a user query has been received by the search engine, providing the user query as input into a pre-trained machine learning model, wherein the pre-trained machine learning model has been trained upon labeled queries submitted to the search engine, wherein labels assigned to the labeled queries are indicative of whether a query of the labeled queries corresponds to an intent to acquire information about a product; and
responsive to obtaining an output of the pre-trained machine learning model indicative of an intent of the user query to acquire information about a product, causing information to be returned to a client computing device based upon the user query being associated with at least one of the first product class or the second product class in the computer-implemented product ontology.