Dynamic evaluation and use of global and contextual personas

Patent metadata

Field	Value
Publication number	US-11367105-B2
Application number	US-202016748371-A
Country	US
Kind code	B2
Filing date	Jan 21, 2020
Priority date	Jan 21, 2020
Publication date	Jun 21, 2022
Grant date	Jun 21, 2022

Abstract

A system and method for content selection and presentation is disclosed. A plurality of content elements configured for presentation in at least one content container is received and one of the plurality of content elements is selected for presentation in the at least one content container. The one of the plurality of content elements is selected by a trained selection model configured to select one of an individual context or a global context. An interface including the selected one of the plurality of content elements is generated.

First claim

What is claimed is:

1. A system for content selection and presentation, comprising:
a communications interface that communicates with a computing device of each of a plurality of users of an online platform to obtain user data;
a database storing a plurality of content elements;
a memory resource storing instructions; and
a processor coupled to the communications interface, the database and the memory resource, the processor being configured to execute the instructions to:
receive, from a first device of a user of the plurality of users, user data indicating one or more instances the user interacted with one or more content elements associated with the online platform on an interface presented by the first device of the user, the interface including a plurality of containers;
obtain, from the database, the plurality of content elements configured for presentation in at least a first content container of the plurality of containers;
iteratively implement a machine learning process that generates a trained selection model by utilizing a reinforcement learning mechanism and at least an individual explore-exploit mechanism;
based on the user data, implement, by utilizing the trained selection model, a selection process to fill at least the first content container of the interface with a content element of the plurality of content elements, the selection process comprising:
selecting a context of a plurality of contexts including an individual context or a global context, the selection being based on a comparison of an expected future reward value of the individual context and an expected future reward value of the global context, wherein the expected future reward value of the global context (CTR global ) is calculated according to:

CT
⁢

R

g
⁢
l
⁢
o
⁢
b
⁢
a
⁢
l

=

∑

k
=
1

n

⁢

⁢

(

q
k

*

p
k

∑

j
=
1

n

⁢

p
j

)

where p k is a past click-through rate, p j is a reward value for a j-th interaction, and q k is a future click-through rate;
based at least on the selected context, selecting a first content element associated with the selected context; and

generate and present the selected first content element in the first content container of the interface on a display of the first device of the user.

2. The system of claim 1 , wherein the expected future reward value is a click-through rate.

3. The system of claim 1 , wherein the expected future reward value is estimated using Thompson sampling.

4. The system of claim 1 , wherein the expected future reward value of the individual context (CTR contextual ) is calculated according to:

CT
⁢

R
contextual

=

∑

l
=
1

L

⁢

[

pct
l

*

(

∑

k
=
1

n
l

⁢

(

q
lk

*

p

l
⁢
k

∑

j
=
1

n
l

⁢

p

l
⁢
j

)

]

where p lk is a sampled click-through rate, q lk is a future click-through rate for each item I lk associated with the individual context C l , p ij , is a reward value for a j-th interaction associated with the individual context C l and pct l is a proportion of impressions to customers for the individual context C l where l=1,2, . . . , L, k=1,2, . . . , n l , and Σ L 1 pct l =1.

5. The system of claim 1 , wherein the individual context is two or more individual contexts combined.

6. The system of claim 5 , wherein the two or more individual contexts are combined based on a pairwise distance.

7. A non-transitory computer readable medium having instructions stored thereon, wherein the instructions, when executed by a processor cause a device to perform ope

Assignees

• Walmart Apollo Llc

Inventors

• Deng Yang
• Deshpande Omkar
• Ali Afroza
• Mitra Abhimanyu
• Achan Kannan

Classifications

• G06Q30/0255 Primary

  based on user history · CPC title

• G06Q30/0268

  at point-of-sale [POS] · CPC title

• G06Q30/0246

  Traffic · CPC title