Invention Grant
- Patent Title: Relevance tuning for online search
-
Application No.: US16718159Application Date: 2019-12-17
-
Publication No.: US11360979B2Publication Date: 2022-06-14
- Inventor: Patrick Robert William Teen
- Applicant: Atlassian Pty Ltd , Atlassian, Inc.
- Applicant Address: AU Sydney; US CA San Francisco
- Assignee: Atlassian Pty Ltd,Atlassian, Inc.
- Current Assignee: Atlassian Pty Ltd,Atlassian, Inc.
- Current Assignee Address: AU Sydney; US CA San Francisco
- Agency: Brownstein Hyatt Farber Schreck, LLP
- Main IPC: G06F16/00
- IPC: G06F16/00 ; G06F16/2453 ; G06F16/953

Abstract:
In an embodiment, the disclosed technologies are capable of creating and storing a digital repository capable of storing digital work items; receiving a query; determining historical alpha values and historical beta values for each of at least two candidate search algorithms; for each of the at least two candidate search algorithms, generating a beta distribution using the historical alpha values and the historical beta values as parameters; for each of the at least two candidate search algorithms, computing, using values sampled from the beta distribution, a score that indicates a likelihood of the candidate search algorithm producing a search result that satisfies a relevance criterion; selecting, from the at least two candidate search algorithms, an algorithm that has a score that satisfies a selection criterion; and executing the query using the selected algorithm.
Public/Granted literature
- US20210182294A1 RELEVANCE TUNING FOR ONLINE SEARCH Public/Granted day:2021-06-17
Information query