Systems and methods for identifying similar electronic content items

Patent metadata

Field	Value
Publication number	US-11995088-B2
Application number	US-202017137136-A
Country	US
Kind code	B2
Filing date	Dec 29, 2020
Priority date	Dec 29, 2020
Publication date	May 28, 2024
Grant date	May 28, 2024

Abstract

Techniques for identifying similarity between a primary content item and candidate content items are disclosed. The techniques include receiving a similarity request from a client device, the similarity request including contextual data about a primary content item; determining a vector number for the primary content item using the contextual data; identifying candidate content items based on the contextual data; retrieving vector numbers for the identified candidate content items; computing a distance between the vector number of the primary content item and the vector number of each of the candidate content items; ranking the candidate content items based on their computed distance from the primary content item; and selecting at least a subset of the ranked candidate content items as similar content items and communicating the selected subset of the ranked candidate content items to the client device for display on a display of the client device.

First claim

What is claimed is:

1. A computer-implemented method, comprising:
at a semantic engine:
receiving a similarity request from a client device, the similarity request associated with a primary issue request submitted to an issue tracking system, the primary issue request associated with a primary issue having a primary issue title, a primary issue description, and a primary issue status;
determining a vector number for the primary issue request based on a semantic analysis of the primary issue title, the primary issue description, and the primary issue status;
submitting a query to a database of the issue tracking system to obtain a set of candidate issues, the query including one or more of the primary issue title, the primary issue description, or the primary issue status, each candidate issue including a respective issue title, a respective issue description, or a respective issue status that corresponds to the query;
in accordance with a determination that a particular issue of the set of candidate issues is an unprocessed record, determining a particular vector number for the particular issue using a particular issue title, a particular issue description and a particular issue status of the particular issue;
retrieving a set of stored vector numbers for candidate issues of the set of candidate issues having processed records;
computing a respective distance between the vector number of the primary issue request and the vector number of each candidate issue of the set of candidate issues;
in response to a subset of the candidate issues having the respective distance satisfying an incident criteria, generating an alert message regarding a potential incident, the alert message including incident content generated using one or more of the subset of the candidate issues; and
causing transmission of the alert message to an incident dashboard of an incident management system, the incident dashboard configured to display the incident content.

2. The computer-implemented method of claim 1 , wherein the query includes a similarity request type, the similarity request type being at least one of a request for identifying similar issue requests to an issue request being created by a user, or a request for identifying similar issue requests to an issue request currently being tracked by the user.

3. The computer-implemented method of claim 2 , wherein selecting the set of candidate issue requests comprises selecting the set of candidate issue requests based on rules associated with the similarity request type.

4. The computer-implemented method of claim 1 , wherein the query includes at least one of an issue identifier, a project identifier, or a tenant identifier.

5. The computer-implemented method of claim 4 , wherein at least a portion of the set of candidate issues are selected based on the project identifier and the tenant identifier.

6. The computer-implemented method of claim 4 , wherein the query further includes a field indicating a status of the primary issue.

7. The computer-implemented method of claim 6 , wherein at least a portion of the set of candidate issues are based on the status of the primary issue.

8. The computer-implemented method of claim 1 , wherein determining the vector number for the primary issue request comprises:
determining whether the primary issue request has to be vectorized;
upon determining that the primary issue request has to be vectorized, generating the vector number for the primary issue request by applying a word embedding model on the contextual data in the similarity request.

9. The computer-implemented method of claim 1 , wherein determining the vector number for the primary issue request comprises:
determining whether the primary issue request has to be vectorized;
upon determining that the primary issue request does not have to be vectorized, retrieving the vector number for the primary issue request from a vector database.

10. The computer-implemented method of claim 1 , wherein the incident criteria is satisfied if the subset of candidate issues includes a threshold number of candidate issue that are within a predetermined distance from the primary issue request.

11. Non-transitory computer readable media storing instructions, which when executed by a processor cause the processor to perform a method, comprising:
receiving a similarity request from a client device, the similarity request associated with a primary issue request submitted to an issue tracking system, the primary issue request associated with a primary issue having a primary issue title, a primary issue description, and a primary issue status;
determining a vector number for the primary issue request based on a semantic analysis of the primary issue title, the primary issue description, and the primary issue status;
submitting a query to a database of the issue tracking system to obtain a set of candidate issues, the query including one or more of the primary issue title, the primary issue description, or the primary issue status, each candidate issue including a respective issue title, a respective issue description, or a respective issue status that corresponds to the query;
in accordance with a determination that a particular issue of the set of candidate issues is an unprocessed record, determining a particular vector number for the particular issue using a particular issue title, a particular issue description and a particular issue status of the particular issue;
retrieving a set of stored vector numbers for candidate issues of the set of candidate issues having processed records;
computing a respective distance between the vector number of the primary issue request and the vector number of each candidate issue of the set of candidate issues;
in response to a subset of the candidate issues having the respective distance satisfying an incident criteria, generating an alert message regarding a potential incident, the alert message including incident content generated using one or more of the subset of the candidate issues; and
causing transmission of the alert message to an incident dashboard of an incident management system, the incident dashboard configured to display the incident content.

12. The non-transitory computer readable media of claim 11 , wherein the query includes a similarity request type, the similarity request type being at least one of a request for identifying similar issue requests to an issue request being created by a user, or a request for identifying similar issue requests to an issue request currently being tracked by the user.

13. The non-transitory computer readable media of claim 12 , wherein at least a portion of the set of candidate issues are based on rules associated with the similarity request type.

14. The non-transitory computer readable media of claim 11 , wherein the query includes at least one of an issue identifier, a project identifier, or a tenant identifier.

15. The non-transitory computer readable media of claim 14 , wherein at least a portion of the set of candidate issues are based on the project identifier and the tenant identifier.

16. The non-transitory computer readable media of claim 14 , wherein the query includes a field indicating a status of the primary issue, and at least a portion of the set of candidate issues are selected based on the status of the primary issue.

17. The non-transitory computer readable media of claim 11 , wherein to determine the vector number for the primary issue request, the instructions further cause the processor to:
determine whether the primary issue request has to be vect

Assignees

• Atlassian Pty Ltd
• Atlassian Inc
• Atlassian Us Inc

Inventors

• Awadhwal Gaurav
• Wondo Addo
• Sims Geoff
• Rho Gyeun Gyun
• Saad Ahmed
• Robilliard Simon Hugo

Classifications

• G06N3/0455

  Auto-encoder networks; Encoder-decoder networks · CPC title

• G06N3/0499

  Feedforward networks · CPC title

• G06F16/24578 Primary

  using ranking · CPC title

• G06F17/16

  Matrix or vector computation {, e.g. matrix-matrix or matrix-vector multiplication, matrix factorization (matrix transposition G06F7/78)} · CPC title

• G06F16/35

  Clustering; Classification · CPC title