US 11,954,162 B2
Recommending information to present to users without server-side collection of user data for those users
Dongjin Lee, San Jose, CA (US); Dongyun Jin, San Jose, CA (US); Rezwana Karim, Sunnyvale, CA (US); Zhi Nie, Sunnyvale, CA (US); Rishikesh Ghewari, Sunnyvale, CA (US); Yinan Li, Mountain View, CA (US); and Mingyang Wang, Santa Clara, CA (US)
Assigned to SAMSUNG ELECTRONICS CO., LTD., Gyeonggi-Do (KR)
Filed by SAMSUNG ELECTRONICS CO., LTD., Gyeonggi-Do (KR)
Filed on Sep. 30, 2020, as Appl. No. 17/039,305.
Prior Publication US 2022/0100809 A1, Mar. 31, 2022
Int. Cl. G06F 16/9535 (2019.01); G06F 16/28 (2019.01); G06F 16/9538 (2019.01)
CPC G06F 16/9535 (2019.01) [G06F 16/287 (2019.01); G06F 16/9538 (2019.01)] 15 Claims
OG exemplary drawing
 
1. A method, comprising:
identifying, by a client device, user data pertaining to use of the client device by a user;
storing, by the client device, the user data without sharing with a server-side data processing system at least a portion of the user data that is determined to remain private;
determining, by the client device, at least one persona trait of the user based on the user data pertaining to the use of the client device by the user;
receiving, by the client device and from the server-side data processing system, persona categorization data including a machine learning model trained by the server-side data processing system, the machine learning model being executable to predict at least one persona category from a plurality of persona categories;
assigning, by the client device executing the machine learning model, the user to at least one persona category of the plurality of persona categories based on the at least one determined persona trait of the user;
determining, by the client device, at least one context trait of present use of the client device by the user based on the user data pertaining to the use of the client device by the user, the user data pertaining to the use of the client device by the user comprising at least one trait type pertaining to an application category usage and a score for the application category usage;
receiving, by the client device and from the server-side data processing system, context categorization data including a further machine learning model trained by the server-side data processing system, the further machine learning model being executable to predict a context categorization from a plurality of context categorizations;
assigning, by the client device executing the further machine learning model, the present use of the client device by the user to at least one context category selected from the plurality of context categories based, at least in part, on the at least one determined context trait of the present use of the client device by the user;
sending the at least one persona category and the at least one context category to the server-side data processing system and identifying, by the server-side data processing system, based on the data received from the client device, information to present to users who are assigned to the at least one persona category and the at least one context category to which the user is assigned;
receiving the information from the server-side data processing system, wherein the information specifies a list of a plurality of applications recommended to install on the client device of the user, wherein each application of the plurality of applications is ranked by the server-side data processing system;
adjusting the rankings of one or more of the plurality of applications of the list based on recent user activities detected within a selected threshold period of time;
determining, by interacting with an operating system of the client device, applications installed on the client device and filtering the list to remove the applications installed on the client device from the list; and
recommending to the user, by a user interface of the client device, the information as filtered and ranked post adjusting.