Personalized mechanisms to resolve explore-exploit dilemma with dynamically shared learnings

Patent metadata

Field	Value
Publication number	US-11030634-B2
Application number	US-201916263873-A
Country	US
Kind code	B2
Filing date	Jan 31, 2019
Priority date	Jan 30, 2018
Publication date	Jun 8, 2021
Grant date	Jun 8, 2021

Abstract

A method including displaying content elements on one or more websites to users. The classification of the users into segments can be based on each impression of content elements being displayed on the one or more websites to a user of the users, tracking impression response data comprising (a) a response of the user to the content element of the content elements displayed on the one or more websites, and (b) one or more segments of the segments in which the user is classified. The method can also include receiving a request from a first user. The method can also include generating a mixture distribution for the first segment for the first content element based on the impression response data for the first segment and generating the webpage to comprise the selected content element. Other embodiments are disclosed.

First claim

What is claimed is:

1. A system comprising:
one or more processors; and
one or more non-transitory computer-readable media storing computing instructions configured to run on the one or more processors and perform:
displaying content elements on one or more websites to users;
performing a classification of the users into segments, each of the users being classified into one or more segments of the segments;
for each impression of a content element of the content elements being displayed on the one or more websites to a user of the users, tracking impression response data comprising (a) a response of the user to the content element of the content elements displayed on the one or more websites, and (b) the one or more segments of the segments in which the user is classified;
receiving a request from a first user of the users to display a webpage of the one or more websites, the first user being classified into one or more first segments of the segments;
determining an amount of information in a respective posterior distribution of each of the one or more first segments for each of the content elements based on the impression response data;
for each first segment of the one or more first segments and each first content element of the content elements in which the amount of information for the first segment for the first content element is below a predetermined informational threshold:
determining a nearest segment to the first segment for the first content element; and
generating a mixture distribution for the first segment for the first content element based on the impression response data for the first segment for the first content element and the impression response data for the nearest segment for the first content element;

determining weightings of the content elements for the one or more first segments, based at least in part on the mixture distributions;
selecting a selected content element from among the content elements based on the weightings of the content elements for the one or more first segments; and
generating the webpage to comprise the selected content element.

2. The system of claim 1 , wherein determining the amount of information in the respective posterior distribution of each of the one or more first segments for each of the content elements comprises one of:
generating a standard deviation of the respective posterior distribution; or
generating a coefficient of variation of the respective posterior distribution.

3. The system of claim 1 , wherein determining the nearest segment to the first segment for the first content element further comprises:
determining second segments of the segments in which an amount of information in a respective posterior distribution of each of the second segments is equal to or greater than the predetermined informational threshold;
computing, for each second segment of the second segments, a respective correlation between a posterior means of common content elements in the first segment and the second segment; and
selecting the nearest segment from among the second segments based on a maximum of the correlations for the second segments.

4. The system of claim 3 , wherein:
the common content elements in the first segment and the second segment are restricted to content elements that belong to an identical content element class.

5. The system of claim 1 , wherein:
an amount of information from the impression response data for the nearest segment for the first content element that is used in the mixture distribution for the first segment for the first content element is adapted based on an information parameter in the mixture distribution for the first segment for the first content element.

6. The system of claim 5 , wherein:
the information parameter in the mixture distribution for the first segment for the first content element is based on a variance of a posterior distribution of the first segment for the first content element.

7. The system of claim 1 , wherein:
the segments classify the users based on one of:
a respective state of each of the users; or
a respective zip code of each of the users.

8. The system of claim 1 , wherein:
each of the users is assigned a respective channel classification;
the impression response data further comprises, for each impression of the content element, (a) the respective channel classification of the user of the users, and (b) respective responses of the user in each of two channels; and
the weightings are further determined based on the respective channel classification of the first user and the respective responses of the users in each of the two channels in the impression response data.

9. The system of claim 1 , wherein:
the impression response data further comprises, for each impression of the content element, a respective time of the response of the user; and
the weightings are further determined based on posterior distributions using the impression response data, as adjusted by a temporal decay factor, based on the respective times of the impression response data.

10. The system of claim 9 , wherein:
each of the users is assigned a respective channel classification;
the impression response data further comprises, for each impression of the content element, (a) the respective channel classification of the user of the users, and (b) respective responses of the user in each of two channels; and
the weightings are further determined based on the respective channel classification of the first user and the respective responses of the users in each of the two channels in the impression response data.

11. A method being implemented via execution of computing instructions configured to run at one or more processors and stored at one or more non-transitory computer-readable media, the method comprising:
displaying content elements on one or more websites to users;
performing a classification of the users into segments, each of the users being classified into one or more segments of the segments;
for each impression of a content element of the content elements being displayed on the one or more websites to a user of the users, tracking impression response data comprising (a) a response of the user to the content element of the content elements displayed on the one or more websites, and (b) the one or more segments of the segments in which the user is classified;
receiving a request from a first user of the users to display a webpage of the one or more websites, the first user being classified into one or more first segments of the segments;
determining an amount of information in a respective posterior distribution of each of the one or more first segments for each of the content elements based on the impression response data;
for each first segment of the one or more first segments and each first content element of the content elements in which the amount of information for the first segment for the first content element is below a predetermined informational threshold:
determining a nearest segment to the first segment for the first content element; and
generating a mixture distribution for the first segment for the first content element based on the impression response data for the first segment for the first content element and the impression response data for the nearest segment for the first content element;

determining weightings of the content elements for the one or more first segments, based at least in part on the mixture distributions;
selecting a selected content element from among the content elements based on the weightings of the content elements for the one or more first segments; and
generating the webpage to comprise the selected c

Assignees

• Walmart Apollo Llc

Inventors

• Mitra Abhimanyu
• Achan Kannan
• Ali Afroza
• Deshpande Omkar
• Kashyap Abhay

Classifications

• H04L67/535

  Tracking the activity of the user (network monitoring arrangements H04L43/00; recording of computer activity G06F11/34) · CPC title

• G06F16/9535

  Search customisation based on user profiles and personalisation · CPC title

• G06F16/958

  Organisation or management of web site content, e.g. publishing, maintaining pages or automatic linking · CPC title

• G06Q30/0205 Primary

  based on location or geographical consideration · CPC title

• G06F16/906

  Clustering; Classification · CPC title