US 11,874,939 B2
Generating user-specific entity interlinkages of extracted enterprise topic descriptions
Vipindeep Vangala, Hyderabad (IN); Ranganath Kondapally, Hyderabad (IN); Pankaj Vasant Khanzode, Hyderabad (IN); Beethika Tripathi, Hyderabad (IN); Daraksha Parveen, Hyderabad (IN); Madan Gopal Jhanwar, Hyderabad (IN); Jimish Bhayani, Hyderabad (IN); Priyam Bakliwal, Hyderabad (IN); and Jatin Kakkar, Hyderabad (IN)
Assigned to MICROSOFT TECHNOLOGY LICENSING, LLC, Redmond, WA (US)
Filed by MICROSOFT TECHNOLOGY LICENSING, LLC, Redmond, WA (US)
Filed on Jan. 30, 2021, as Appl. No. 17/163,420.
Prior Publication US 2022/0245267 A1, Aug. 4, 2022
Int. Cl. G06F 21/62 (2013.01); G06N 20/00 (2019.01); G06Q 10/101 (2023.01); G06Q 10/105 (2023.01)
CPC G06F 21/6218 (2013.01) [G06N 20/00 (2019.01); G06Q 10/101 (2013.01); G06Q 10/105 (2013.01)] 20 Claims
OG exemplary drawing
 
1. A computer-implemented method, the method comprising:
receiving a plurality of documents extracted from at least one enterprise computing resource storing the plurality of documents for a plurality of user accounts associated with an enterprise;
inputting the plurality of documents into a machine learning (ML) model that is configured to extract a plurality of topics;
receiving the plurality of topics from the ML model;
receiving user profile data that defines, in association with an individual user account of the plurality of user accounts, at least one of:
one or more directory attributes associated with the individual user account, or
user knowledge graph preferences associated with the individual user account;
generating, based on the user profile data, a knowledge graph that defines entity interlinkages associating the plurality of topics;
receiving an access request that corresponds to an individual document accessed via the individual user account;
and
selecting, using the knowledge graph and for exposure to the individual user account in association with the individual document, an individual topic within the knowledge graph that is generated based on the user profile data.