Extracting and surfacing contextually relevant topic descriptions
Techniques for extracting and ranking multiple topic descriptions based on source contexts and subsequently selecting individual topic descriptions to surface based on recipient contexts. More specifically, a mining platform may extract, from a set of source documents making up a corpus, topic descriptions for various topics that are relevant to an enterprise. The mining platform may further rank the extracted topic descriptions based on a source context of those documents from which individual topic descriptions are extracted. Subsequently, when users access enterprise documents including term-usage instances of topics for which one or more topic descriptions have been extracted and ranked, a description serving module may select a topic description that is contextually appropriate for a recipient view the enterprise documents.
1. A computer-implemented method, the method comprising:
receiving a corpus that is extracted from at least one enterprise computing resource that is configured to store a plurality of documents in association with one or more user accounts;
inputting the corpus into a machine learning (ML) model that is configured to:
extract, from the corpus, a plurality of topic descriptions for a plurality of topics, and
generate a ranked listing, of the plurality of topic descriptions, based on one or more source contexts that are associated with individual topic descriptions of the plurality of topic descriptions;
receiving, from the ML model, an output that includes the plurality of topic descriptions and the ranked listing;
identifying an individual topic within an individual document that is accessed from an individual user account of the one or more user accounts;
determining a recipient context of the individual topic in association with the individual user account; and
selecting, from the ranked listing and based on the recipient context, an individual topic description for exposure to the individual user account in association with the individual document.
2. The computer-implemented method of claim 1 , further comprising causing a link to the individual topic description to be generated in association with a term-usage instance of the individual topic within the individual document that is being accessed from the individual user account.
3. The computer-implemented method of claim 1 , wherein the one or more source contexts include an authoritative status of an author, of the individual topic description, in association with the individual topic.
4. The computer-implemented method of claim 1 , wherein the one or more source contexts include a dissemination level, of the individual topic description in association with the individual topic, across the one or more user accounts.
5. The computer-implemented method of claim 1 , wherein the determining the recipient context is based on a directory attribute, of the individual user account, that is indicative of a relationship of the individual user account with respect to the individual topic.
6. The computer-implemented method of claim 1 , wherein the determining the recipient context is based on an indication, within an access control list, of whether the individual user account is restricted access to one or more source documents from which the individual topic description is extracted.
7. The computer-implemented method of claim 1 , wherein the receiving the corpus includes:
receiving metadata corresponding to the plurality of documents that are stored in association with the one or more user accounts;
receiving corpus exclusion criteria for excluding a subset of the plurality of documents from the corpus; and
identifying the corpus based on the metadata and the corpus exclusion criteria.
8. The computer-implemented method of claim 1 , wherein the corpus is a user-specific corpus that uniquely corresponds to the individual user account.
9. The computer-implemented method of claim 1 , wherein the corpus is a tenant-wide corpus that corresponds to multiple user accounts that each have access to the at least one enterprise computing resource.
10. A system, comprising:
at least one processor; and
at least one memory in communication with the at least one processor, the at least one memory having computer-readable instructions stored thereupon that, when executed by the at least one processor, cause the at least one processor to:
generate a corpus that includes a plurality of documents that are stored in association with at least one enterprise computing resource associated with one or more user accounts;
receive an output that is generated by a machine learning (ML) model based on the corpus, wherein the output includes a ranked listing of a plurality of topic descriptions;
identify an individual topic within an individual document that is accessed from an individual user account of the one or more user accounts;
determine a recipient context of the individual topic in association with the individual user account; and
select, from the ranked listing and based on the recipient context, an individual topic description for exposure to the individual user account in association with the individual document.
11. The system of claim 10 , wherein:
the output further indicates one or more source contexts that are associated with individual topic descriptions of the plurality of topic descriptions, and
the individual topic description is further selected based on a correspondence level between the recipient context and an individual source context that is associated with the individual topic description.
12. The system of claim 10 , wherein the recipient context is determined based on a disambiguation of the individual topic that is determined from a term-usage instance of the individual topic within the individual document.
13. The system of claim 10 , wherein the recipient context is determined based on a directory attribute, of the individual user account, that is indicative of a relationship of the individual user account with respect to an author of the individual topic description.
14. The system of claim 10 , wherein the corpus is a user-specific corpus that uniquely corresponds to the individual user account.
15. The system of claim 10 , wherein generating the corpus is based on:
metadata corresponding to the plurality of documents that are stored in association with the one or more user accounts, and
corpus exclusion criteria for excluding a subset of the plurality of documents from the corpus.
16. A system, comprising:
at least one processor; and
at least one memory in communication with the at least one processor, the at least one memory having computer-readable instructions stored thereupon that, when executed by the at least one processor, cause the at least one processor to:
generate a corpus that includes a plurality of documents that are stored in association with at least one enterprise computing resource associated with one or more user account;
receive an output that is generated by a machine learning (ML) model based on the corpus, wherein the output includes:
a ranked listing of a plurality of topic descriptions, and
one or more source contexts that are associated with individual topic descriptions of the plurality of topic descriptions;
identify an individual topic within an individual document that is accessed from an individual user account of the one or more user accounts; and
select, from the ranked listing and based on the one or more source contexts, an individual topic description for exposure to the individual user account in association with the individual document.
17. The system of claim 16 , wherein the one or more source contexts include an authoritative status of an author, of the individual topic description, in association with the individual topic.
18. The system of claim 16 , wherein the one or more source contexts include a dissemination level, of the individual topic description in association with the individual topic, across the one or more user accounts.
19. The system of claim 16 , wherein the corpus is a tenant-wide corpus that corresponds to multiple user accounts that each have access to the at least one enterprise computing resource.
20. The system of claim 16 , wherein the corpus is a user-specific corpus that uniquely corresponds to the individual user account.