Extracting and surfacing topic descriptions from regionally separated data stores

Patent metadata

Field	Value
Publication number	US-11586662-B2
Application number	US-202117210414-A
Country	US
Kind code	B2
Filing date	Mar 23, 2021
Priority date	Mar 5, 2021
Publication date	Feb 21, 2023
Grant date	Feb 21, 2023

Abstract

Extracting and surfacing information corresponding to individual logical topics from enterprise data stores that are separated across multiple geographic regions. A clustering service creates, by utilizing machine learning toolkits that are agnostic to the region in which data is stored, individual topics that have references to multiple shards of data that are stored in different geographic regions. The clustering service also shards the knowledge base state according to the regions from which pieces of data for the particular logical topic was extracted. For example, a first shard containing information extracted from a first document may be stored in a first region whereas a second shard containing information extracted from a second document may be stored in a second region. Responsive to user activity associated with the topic, a serving platform may identify and reconstitute these shards that are stored in different regions so as to surface the regionally extracted and sharded information on that topic to a user.

First claim

What is claimed is:

1. A computer-implemented method, the method comprising:
receiving a corpus that includes at least: first enterprise data that is stored within at least one first enterprise computing resource that is operating within a first region, and second enterprise data that is stored within at least one second enterprise computing resource that is operating within a second region;
inputting the corpus into a machine learning (ML) model that is configured to output data extracts corresponding to a plurality of topics;
receiving, from the ML model, an output that includes at least: first data extracts that are extracted from the first enterprise data, and second data extracts that are extracted from the second enterprise data;
generating, based on the output, a knowledge base (KB) state that defines references between a particular topic and at least: a first KB shard that is generated based on the first data extracts, and a second KB shard that is generated based on the second data extracts;
causing the first KB shard to be stored within the at least one first enterprise computing resource; and
causing the second KB shard to be stored within the at least one second enterprise computing resource.

2. The computer-implemented method of claim 1 , further comprising:
generating a site-region map that includes at least: an association between the first KB shard and a first data object from which first information in the first KB shard was extracted by the ML model, and an indication that the first data object is stored within the at least one first enterprise computing resource that is operating within the first region.

3. The computer-implemented method of claim 2 , further comprising:
detecting a migration of the first data object from the at least one first enterprise computing resource that is operating within the first region to at least one third enterprise computing resource that is operating within a third region;
responsive to the migration:
causing the first KB shard to be stored within the at least one third enterprise computing resource that is operating within the third region; and
updating a topic-region map to indicate that the first data object is stored within the at least one third enterprise computing resource that is operating within the third region.

4. The computer-implemented method of claim 1 , further comprising:
providing the KB state to at least one serving platform that is configured to surface the particular topic by reconstituting the first KB shard, that is stored within the at least one first enterprise computing resource, and the second KB shard that is stored within the at least one second enterprise computing resource.

5. The computer-implemented method of claim 4 , wherein the at least one serving platform is operating within the first region, and wherein the at least one serving platform is further configured to transmit a region-specific query to the at least one second enterprise computing resource to request the second KB shard.

6. The computer-implemented method of claim 1 , wherein:
the first KB shard is not stored in the at least one second enterprise computing resource; and
the second KB shard is not stored in the at least one first enterprise computing resource.

7. The computer-implemented method of claim 1 , further comprising:
receiving, at the first region, a request that corresponds to the particular topic from a client device;
in response to the request, transmitting a region-specific query to the at least one second enterprise computing resource to request the second KB shard;
receiving, at the first region, the second KB shard from the at least one second enterprise computing resource; and
reconstituting, at the first region, the first KB shard and the second KB shard into the particular topic.

8. A system comprising:
one or more processors; and
at least one computer storage medium having computer executable instructions stored thereon which, when executed by the one or more processors, cause the system to:
receive a corpus that includes at least: first enterprise data that is stored within at least one first enterprise computing resource that is operating within a first region, and second enterprise data that is stored within at least one second enterprise computing resource that is operating within a second region;
input the corpus into a machine learning (ML) model that is configured to output data extracts corresponding to a plurality of topics;
generate, based on the output, a knowledge base (KB) state that defines references between a particular topic and at least: a first KB shard, and a second KB shard;
cause the first KB shard to be stored within the at least one first enterprise computing resource; and
cause the second KB shard to be stored within the at least one second enterprise computing resource.

9. The system of claim 8 , wherein the computer executable instructions further cause the system to:
generate a site-region map that includes at least: an association between the first KB shard and a first data object from which first information in the first KB shard was extracted by the ML model, and an indication that the first data object is stored within the at least one first enterprise computing resource that is operating within the first region.

10. The system of claim 9 , wherein the computer executable instructions further cause the system to:
detect a migration of the first data object from the at least one first enterprise computing resource that is operating within the first region to at least one third enterprise computing resource that is operating within a third region;
responsive to the migration:
cause the first KB shard to be stored within the at least one third enterprise computing resource that is operating within the third region; and
update a topic-region map to indicate that the first data object is stored within the at least one third enterprise computing resource that is operating within the third region.

11. The system of claim 8 , wherein the computer executable instructions further cause the system to:
provide the KB state to at least one serving platform that is configured to surface the particular topic by reconstituting the first KB shard, that is stored within the at least one first enterprise computing resource, and the second KB shard that is stored within the at least one second enterprise computing resource.

12. The system of claim 11 , wherein the at least one serving platform is operating within the first region, and wherein the at least one serving platform is further configured to transmit a region-specific query to the at least one second enterprise computing resource to request the second KB.

13. The system of claim 8 , wherein the computer executable instructions further cause the system to:
receive, from the ML model, an output that includes at least: first data extracts that are extracted from the first enterprise data, and second data extracts that are extracted from the second enterprise data, wherein:
the first KB shard is generated based on the first data extracts, and
the second KB shard is generated based on the second data extracts.

14. The system of claim 8 , wherein the computer executable instructions further cause the system to:
receive, at the first region, a request that corresponds to the particular topic from a client device;
in response to the request, transmit a region-specific query to the at least one second enterprise computing resource to request the second KB shard;
receiving, at the first region, the second KB shard from the at least one second enterpri

Assignees

• Microsoft Technology Licensing Llc

Inventors

• Meyerzon Dmitriy
• Voronkov Nikita
• Gvozdev Vladimir
• Miao Kaixiang

Classifications

• G06N3/09

  Supervised learning · CPC title

• G06N5/022 Primary

  Knowledge engineering; Knowledge acquisition · CPC title

• G06N3/045

  Combinations of networks · CPC title

• G06N20/00

  Machine learning · CPC title

• G06F16/214

  Database migration support · CPC title