Method and apparatus for real-time interactive recommendation

Patent metadata

Field	Value
Publication number	US-11741072-B2
Application number	US-201716622253-A
Country	US
Kind code	B2
Filing date	Jun 14, 2017
Priority date	Jun 14, 2017
Publication date	Aug 29, 2023
Grant date	Aug 29, 2023

Abstract

An interactive recommendation system is provided. The system comprises an information generation module configured to: provide a set of information contents for outputting in an information interface, wherein each of the set of information contents is associated with an information item; receive a selection of one of the set of information contents that signals the user's interest or disinterest in an information item; determine, based on a model and history data related to prior activities of the user including a plurality of selections from the set of information contents, one or more information items to be included in recommendation information for the user; provide the recommendation information for outputting in a recommendation information interface; and update, based on the history data, the set of information contents for outputting in the information interface.

First claim

What is claimed is:

1. An interactive recommendation system comprising an information generation module configured to:
provide a set of information contents for outputting in an information interface, each of the set of information contents being associated with an information item;
receive a selection of one of the set of information contents that signals the user's interest or disinterest in the information item and an input of a search query to a search engine, wherein in receiving the input of the search query to the search engine, the information generation module is further configured to:
extract a set of keywords of the input and a timestamp for the information item; and
determine, based on a model and history data related to prior activities of the user including a plurality of selections from the set of information contents and the input of the search query, one or more information items that are to be included in recommendation information for the user and that are associated with the set of keywords based on a mapping table that maps the set of keywords to the one or more information items, wherein each of the prior activities is associated with the timestamp; and
provide the recommendation information for outputting in a recommendation information interface.

2. The system of claim 1 , wherein the information generation module is further configured to:
update, based on the history data, the set of information contents for outputting in the information interface.

3. The system of claim 1 , wherein the information generation module is further configured to:
determine, based on the selection of the one of the set of information contents, disinterest in an information item associated with the one of the set of information contents; and
remove the information item from the recommendation information.

4. The system of claim 2 , wherein the information generation module is further configured to:
determine, based on the selection of the one of the set of information contents, interest in an information item associated with the one of the set of information contents; and
update the set of information contents for outputting in the information interface based on the information item.

5. The system of claim 1 , wherein the history data comprises a log of a set of activities within a predetermined period; wherein the system further comprises a feature generation module configured to:
determine a set of features for an information item of the one or more information items, wherein each feature of the set of features is associated with one of the set of activities, and
determine, for each feature of set of features, a value representing a number of occurrences of the information item in the log of an activity of the set of activities associated with the each feature;
wherein the system further comprises a user interest prediction module configured to determine a score that represents a likelihood of the user's interest in the information item based on the values; and
wherein the information generation module is configured to provide the information item in the recommendation information interface based on the score.

6. The system of claim 5 , wherein the information generation module is further configured to:
determine the score for each of a set of candidate information items;
rank the set of candidate information items based on the scores; and
select the one or more information items from the candidate information items based on the ranking.

7. The system of claim 6 , wherein the information generation module is further configured to:
obtain a set of candidate information contents, wherein each of the set of candidate information contents is associated with one of the candidate information items;
rank the set of candidate information contents based on the scores; and
provide one or more of the set of candidate information contents as a part of the set of information contents for outputting in the information interface.

8. The system of claim 7 , wherein the information generation module is further configured to control the information interface to output the one or more of the set of candidate information contents in an order based on the ranking of the set of candidate information contents.

9. A method for generating recommendation information, the method comprising:
providing a set of information contents for outputting in an information interface, each of the set of information contents being associated with an information item;
receiving, from the user, a selection of one of the set of information contents that signals the user's interest or disinterest in the information item and an input of a search query to a search engine, wherein receiving the input of the search query to the search engine further comprises:
extracting a set of keywords of the input and a timestamp for the information item; and
determining, based on a model and history data related to prior activities of the user including a plurality of selections from the set of information contents and the input of the search query, one or more information items that are to be included in recommendation information for the user and that are associated with the set of keywords based on a mapping table that maps the set of keywords to the one or more information items, wherein each of activities is associated with the timestamp; and
providing, in a recommendation information interface, the recommendation information.

10. The method of claim 9 , further comprising:
updating, based on the history data, the set of information contents for outputting in the information interface.

11. The method of claim 9 , further comprising:
determining, based on the selection of the one of the set of information contents, disinterest in an information item associated with the one of the set of information contents; and
removing the information item from the recommendation information.

12. The method of claim 10 , further comprising:
determining, based on the selection of the one of the set of information contents, interest in an information item associated with the one of the set of information contents; and
updating the set of information contents for outputting in the information interface based on the information item.

13. The method of claim 9 , wherein the history data comprises a log of a set of activities within a predetermined period; wherein the method further comprises:
determining a set of features for an information item of the one or more information items, wherein each feature of the set of features is associated with one of the set of activities;
determining, for each feature of set of features, a value representing a number of occurrences of the information item in the log of an activity of the set of activities associated with the each feature;
determining a score that represents a likelihood of the user's interest in the information item based on the values; and
providing the information item in the recommendation information interface based on the score.

14. The method of claim 13 , further comprising:
determining the score for each of a set of candidate information items;
ranking the set of candidate information items based on the scores; and
selecting the one or more information items from the candidate information items based on the ranking.

15. The method of claim 14 , further comprising
obtaining a set of candidate information contents, wherein each of the set of candidate information contents is associated with one of the candidate information items;
ranking the set of candidate information co

Assignees

• Alibaba Group Holding Ltd

Inventors

• Tao Jun
• Ai Shuyu
• Xie Xu
• Zhu Keyao
• Liu Yun
• Wen Jianhua
• Yan Yuliang
• Lang Jun
• Si Luo

Classifications

• G06F16/23 Primary

  Updating · CPC title

• G06N20/00

  Machine learning · CPC title

• G06F16/9535 Primary

  Search customisation based on user profiles and personalisation · CPC title