Generating candidates for search using scoring/retrieval architecture

Patent metadata

Field	Value
Publication number	US-2020004835-A1
Application number	US-201816021667-A
Country	US
Kind code	A1
Filing date	Jun 28, 2018
Priority date	Jun 28, 2018
Publication date	Jan 2, 2020
Grant date	—

Abstract

Techniques for generating candidates for search using a scoring and retrieval architecture and deep semantic features are disclosed herein. In some embodiments, a computer system generates a profile vector representation for user profiles based profile data, stores the profile vector representations, receives a query subsequent to the storing of the profile vector representations, generates a query vector representation for the query, retrieves the stored profile vector representations of the user profiles based on the receiving of the query, generates a corresponding score for pairings of the user profiles and the query based on a determined level of similarity between the profile vector representation of the user profiles and the query vector representation, and causes an indication of at least a portion of the user profiles to be displayed as search results for the query based on the generated scores of the user profiles.

First claim

What is claimed is:

1 . A computer-implemented method comprising:
for each one of a plurality of user profiles stored on a database of an online service, retrieving, by a first neural network, profile data of the one of the plurality of user profiles from the database of the online service; for each one of the plurality of user profiles, generating, by the first neural network, a profile vector representation based on the retrieved profile data of the one of the plurality of user profiles; storing the profile vector representations of the plurality of user profiles in the database of the online service; receiving, by a computer system having a memory and at least one hardware processor, a query from a computing device of a querying user subsequent to the storing of the profile vector representations, the query comprising query data, the query data comprising at least one of query text or facet selection data; generating, by a second neural network distinct from the first neural network, a query vector representation for the query based on the query data of the query in response to the receiving of the query; retrieving, by the computer system, the stored profile vector representations of the plurality of user profiles from the database of the online service based on the receiving of the query; for each one of the plurality of user profiles, generating, by a third neural network distinct from the first neural network and the second neural network, a corresponding score for a pairing of the one of the plurality of user profiles and the query based on a determined level of similarity between the profile vector representation of the one of the plurality of user profiles and the query vector representation; and causing, by the computer system, an indication of at least a portion of the plurality of user profiles to be displayed on the computing device as search results for the query based on the generated scores of the plurality of user profiles.

2 . The computer-implemented method of claim 1 , wherein the first neural network, the second neural network, and the third neural network are implemented on separate physical computer systems, each one of the separate physical computer systems having its own set of one or more hardware processors separate from the other separate physical computer systems.

3 . The computer-implemented method of claim 1 , wherein the first neural network, the second neural network, and the third neural network each comprise a deep neural network.

4 . The computer-implemented method of claim 1 , wherein the first neural network comprises a convolutional neural network.

5 . The computer-implemented method of claim 1 , wherein the causing the indication of at least a portion of the plurality of user profiles to be displayed on the computing device as search results for the query comprises:
ranking the plurality of user profiles based on their corresponding scores; and causing the at least a portion of the plurality of user profiles to be displayed on the computing device as search results for the query based on the ranking of the plurality of user profiles.

6 . The computer-implemented method of claim 1 , wherein the profile data comprises at least one of a job title, a company, a skill, a school, a degree, and an educational major.

7 . The computer-implemented method of claim 1 , wherein the third neural network determines the level of similarity between the profile vector representation of the one of the plurality of user profiles and the query vector representation based on a cosine similarity calculation.

8 . The computer-implemented method of claim 1 , wherein the third neural network determines the level of similarity between the profile vector representation of the one of the plurality of user profiles and the query vector representation based on a dot product calculation.

9 . The computer-implemented method of claim 1 , further comprising:
selecting the plurality of user profiles in response to the receiving of the query based on a comparison of the query data and the corresponding profile data of the user profiles, wherein the retrieving of the stored profile vector representations of the plurality of user profiles from the database of the online service is further based on the selecting of the plurality of user profiles.

10 . A system comprising:
at least one hardware processor; and a non-transitory machine-readable medium embodying a set of instructions that, when executed by at least one hardware processor, cause the processor to perform operations comprising:
for each one of a plurality of user profiles stored on a database of an online service, retrieving, by a first neural network, profile data of the one of the plurality of user profiles from the database of the online service;
for each one of the plurality of user profiles, generating, by the first neural network, a profile vector representation based on the retrieved profile data of the one of the plurality of user profiles;
storing the profile vector representations of the plurality of user profiles in the database of the online service;
receiving, by a computer system having a memory and at least one hardware processor, a query from a computing device of a querying user subsequent to the storing of the profile vector representations, the query comprising query data, the query data comprising at least one of query text or facet selection data;
generating, by a second neural network distinct from the first neural network, a query vector representation for the query based on the query data of the query in response to the receiving of the query;
retrieving, by the computer system, the stored profile vector representations of the plurality of user profiles from the database of the online service based on the receiving of the query;
for each one of the plurality of user profiles, generating, by a third neural network distinct from the first neural network and the second neural network, a corresponding score for a pairing of the one of the plurality of user profiles and the query based on a determined level of similarity between the profile vector representation of the one of the plurality of user profiles and the query vector representation; and
causing, by the computer system, an indication of at least a portion of the plurality of user profiles to be displayed on the computing device as search results for the query based on the generated scores of the plurality of user profiles.

11 . The system of claim 10 , wherein the first neural network, the second neural network, and the third neural network are implemented on separate physical computer systems, each one of the separate physical computer systems having its own set of one or more hardware processors separate from the other separate physical computer systems.

12 . The system of claim 10 , wherein the first neural network, the second neural network, and the third neural network each comprise a deep neural network.

13 . The system of claim 10 , wherein the first neural network comprises a convolutional neural network.

14 . The system of claim 10 , wherein the causing the indication of at least a portion of the plurality of user profiles to be displayed on the computing device as search results for the query comprises:
ranking the plurality of user profiles based on their corresponding scores; and causing the at least a portion of the plurality of user profiles to be displayed on the computing device as search results for the query based on the ranking of the plurality of user profiles.

15 . The system of claim 10 , wherein the profil

Assignees

• Microsoft Technology Licensing Llc

Inventors

• Ramanath Rohan
• Polatkan Gungor
• Guo Qi
• Ozcaglar Cagri
• Kenthapadi Krishnaram
• Geyik Sahin Cem

Classifications

• G06F16/9535

  Search customisation based on user profiles and personalisation · CPC title

• G06F16/24578 Primary

  using ranking · CPC title

• G06F16/2237

  Vectors, bitmaps or matrices · CPC title

• G06N5/022

  Knowledge engineering; Knowledge acquisition · CPC title

• G06N3/088

  Non-supervised learning, e.g. competitive learning · CPC title