Generating personalized query suggestions

Patent metadata

Field	Value
Publication number	US-11475085-B2
Application number	US-202016801725-A
Country	US
Kind code	B2
Filing date	Feb 26, 2020
Priority date	Feb 26, 2020
Publication date	Oct 18, 2022
Grant date	Oct 18, 2022

Abstract

Machine learning based method for generating personalized query suggestions is described. Different users may have different search intent even when they are inputting the same search query. The technical problem of personalizing search query suggestions produced by a machine learning model is addressed by extending the sequence to sequence machine learning model framework to be able to take into consideration additional, personalized features of the user, such as, e.g., profile industry, language, geographic location, etc. This methodology includes an offline model training framework as well as an online serving framework.

First claim

The invention claimed is:

1. A computer implemented method comprising:
training a sequence to sequence machine learning model using historical data associated with searches in an on-line communication network system;
detecting a search request submitted by a user via a user interface provided by the on-line communication network system, the search request comprising a set of word embeddings;
determining a personalization feature associated with the user, the personalization feature comprising at least one of an item of metadata associated with the user, a feature of a user profile, or an item of context information from output of upstream intent detection;
generating search results based on the search request;
prepending the personalization feature to each word embedding from the set of word embeddings;
providing the set of word embeddings, with and the prepended personalization feature, to an encoder of the sequence to sequence machine learning model as input and, in parallel with the generating of the search results, generating a personalized search query suggestion for the user by executing the sequence to sequence machine learning model;
including the search query suggestion, together with the search results generated based on the search request submitted by the user, into a search results user interface for presentation on a display device of the user.

2. The method of claim 1 , comprising causing presentation of the search results user interface on a display device of the user.

3. The method of claim 1 , wherein the providing, as input to the sequence to sequence machine learning model, the set of word embeddings and the personalization feature, comprises adding the personalization feature into the set of word embeddings.

4. The method of claim 1 , comprising generating an expanded data set for training the sequence to sequence machine learning model by adding the personalization feature to the historical data associated with searches in the on-line communication network system.

5. The method of claim 4 , comprising training the sequence to sequence machine learning model on the expanded data set.

6. The method of claim 1 , wherein the personalization feature is derived from a member characteristic obtained from a member profile that represents the user in the on-line communication network system.

7. The method of claim 6 , wherein the member characteristic obtained from the member profile is an industry identification, a skill, or a professional title.

8. The method of claim 1 , wherein the personalization feature represents a language derived from an interface provided by a computer system of the user.

9. The method of claim 1 , comprising:
detecting a selection of a suggestion from the suggested queries presented on the display device of the user; and
processing the selected suggestion to produce a further set of search results to be presented to the user.

10. A system comprising:
one or more processors; and
a non-transitory computer readable storage medium comprising instructions that when executed by the one or processors cause the one or more processors to perform operations comprising:
training a sequence to sequence machine learning model using historical data associated with searches in an on-line communication network system;
detecting a search request submitted by a user via a user interface provided by the on-line communication network system, the search request comprising a set of word embeddings;
determining a personalization feature associated with the user, the personalization feature comprising at least one of an item of metadata associated with the user, a feature of a user profile, or an item of context information from output of upstream intent detection;
generating search results based on the search request;
prepending the personalization feature to each word embedding from the set of word embeddings;
providing the set of word embeddings, with the prepended personalization feature, to an encoder of the sequence to sequence machine learning model as input and, in parallel with the generating of the search results, generating a personalized search query suggestion for the user by executing the sequence to sequence machine learning model;
including the search query suggestion, together with the search results generated based on the search request submitted by the user, into a search results user interface for presentation on a display device of the user.

11. The system of claim 10 , comprising causing presentation of the search results user interface on a display device of the user.

12. The system of claim 10 , wherein the providing, as input to the sequence to sequence machine learning model, the set of word embeddings and the personalization feature, comprises adding the personalization feature into the set of word embeddings.

13. The system of claim 10 , comprising generating an expanded data set for training the sequence to sequence machine learning model by adding the personalization feature to the historical data associated with searches in the on-line communication network system.

14. The system of claim 13 , comprising training the sequence to sequence machine learning model on the expanded data set.

15. The system of claim 10 , wherein the personalization feature is derived from a member characteristic obtained from a member profile that represents the user in the on-line communication network system.

16. The system of claim 15 , wherein the member characteristic obtained from the member profile is an industry identification, a skill, or a professional title.

17. The system of claim 10 , wherein the personalization feature represents a language derived from an interface provided by a computer system of the user.

18. A machine-readable non-transitory storage medium having instruction data executable by a machine to cause the machine to perform operations comprising:
training a sequence to sequence machine learning model using historical data associated with searches in an on-line communication network system;
detecting a search request submitted by a user via a user interface provided by the on-line communication network system, the search request comprising a set of word embeddings;
determining a personalization feature associated with the user, the personalization feature comprising at least one of an item of metadata associated with the user, a feature of a user profile, or an item of context information from output of upstream intent detection;
generating search results based on the search request;
prepending the personalization feature to each word embedding from the set of word embeddings;
providing the set of word embeddings, with the prepended personalization feature, to an encoder of the sequence to sequence machine learning model as input and, in parallel with the generating of the search results, generating a personalized search query suggestion for the user by executing the sequence to sequence machine learning model;
including the search query suggestion, together with the search results generated based on the search request submitted by the user, into a search results user interface for presentation on a display device of the user.

Assignees

• Microsoft Technology Licensing Llc

Inventors

• Zhong Jianling
• Guo Weiwei
• Guo Lin
• Gao Huiji
• Long Bo

Classifications

• G06N3/045

  Combinations of networks · CPC title

• G06F16/9535 Primary

  Search customisation based on user profiles and personalisation · CPC title

• G06N20/00

  Machine learning · CPC title

• G06F16/9538 Primary

  Presentation of query results · CPC title

• G06N3/08

  Learning methods · CPC title