Systems and methods for updating creatives generation models

Patent metadata

Field	Value
Publication number	US-11736422-B2
Application number	US-202217889211-A
Country	US
Kind code	B2
Filing date	Aug 16, 2022
Priority date	Mar 4, 2020
Publication date	Aug 22, 2023
Grant date	Aug 22, 2023

Abstract

Disclosed embodiments provide systems and methods related to updating creatives generation models. The system may include at least one memory unit for storing instructions and at least one processor configured to execute the instructions to perform operations. The operations may include receiving a feedback indication concerning an email message sent to a target, the email message constructed using a first template and associated with a first context, the feedback indication including a recommendation identifier; updating, in response to receiving the feedback indication, a feedback value for the email message stored in a delay buffer; obtaining the updated feedback value upon satisfaction of a time delay condition; updating, using the updated feedback data and the recommendation identifier, a machine learning model configured to recommend templates based on contexts; and constructing and providing a second email message using a second template recommended by the updated machine learning model for a second context.

First claim

What is claimed is:

1. A computer cluster architecture comprising at least one computer cluster including at least one computing node, wherein the at least one computing node comprises:
an interface database containing a plurality of graphical user interfaces (GUIs), the plurality of GUIs including a creation interface, an editor interface, an upload interface, a plurality of approval interfaces, one or more preview interfaces, and one or more save interfaces,
a self-serve tool in data communication with the interface database and configured to be accessed by a user through a portal, the self-serve tool comprising one or more of the plurality of GUIs and provide the one or more of the plurality of GUIs for the user upon receipt of an input by the user through the portal,
a multi-armed bandit application programming interface (API) in data communication with the self-serve tool, the multi-armed bandit API comprising rules for analyzing creatives published by the self-serve tool and for determining a winning creative from the creatives, and
a listener continually or periodically polling feedback data of the creatives for processing by the multi-armed bandit API,
wherein:
the self-serve tool is configured to send to a device a creative configured for display on the device,
the multi-armed bandit API is configured to receive from the device feedback data associated with the creative, the feedback data indicating a degree of user interaction with the creative,
the multi-armed bandit API is configured to check the feedback data for a baked time associated with the creative, the baked time being a time when an age of the feedback data is equal to or greater than a delay time associated with the feedback data,
the listener is configured to detect the feedback data based on the baked time,
the multi-armed bandit API is configured to process the feedback data by comparing the feedback data to one or more performance criteria of the creative upon the feedback data being detected by the listener, and
the multi-armed bandit API is configured to be dynamically updated based on the processing of the feedback data.

2. The computer cluster architecture of claim 1 , wherein the multi-armed bandit API is configured to:
determine a difference between the feedback data and expected feedback data, and
be dynamically updated based on the determined difference.

3. The computer cluster architecture of claim 1 , wherein the multi-armed bandit API is configured to:
determine the feedback data does not have a baked time, and
wait a predetermined amount of time before determining the feedback data has a baked time.

4. The computer cluster architecture of claim 1 , wherein the creative includes at least one of an email, webpage advertisement, or a webpage, for a vehicle finance marketing campaign.

5. The computer cluster architecture of claim 1 , wherein:
the at least one computing node is a first computing node,
the at least one computer cluster includes a second computing node, and
the multi-armed bandit API is configured to:
send the feedback data to the second computing node before sending the feedback data to the first computing node, and
determine the feedback data cannot be processed by the second computing node.

6. The computer cluster architecture of claim 5 , wherein the listener is configured to detect when a node is available and direct the feedback data to the available node.

7. The computer cluster architecture of claim 6 , wherein the feedback data is sent from a distributed delay queue associated with the at least one computer cluster.

8. The computer cluster architecture of claim 1 , wherein the feedback data is sent to the multi-armed bandit API after determining the feedback data has a baked time.

9. The computer cluster architecture of claim 1 , wherein the associated delay time is (i) calculated based on a distribution channel associated with the feedback data and (ii) stored in a memory component of the at least one computer cluster.

10. The computer cluster architecture of claim 1 , wherein:
the listener detects the feedback data based on the delay time, and
the listener is configured to direct the feedback data to the at least one computing node for processing while preventing another computing node of the at least one computer cluster from receiving the feedback data such that the at least one computer cluster is enabled to work efficiently by reducing or eliminating number of operations.

11. The computer cluster architecture of claim 1 , wherein the multi-armed bandit API is configured to store the feedback data at a memory component of the at least one computer cluster.

12. The computer cluster architecture of claim 1 , wherein the multi-armed bandit API includes a machine learning model.

13. The computer cluster architecture of claim 12 , wherein the machine-learning model comprises at least one of a long short-term memory (LSTM) network, attention network, sequence-to-sequence (seq2seq) network, or autoencoder.

14. The computer cluster architecture of claim 1 , wherein the feedback data is associated with a distribution channel and the creative is associated with the distribution channel.

15. The computer cluster architecture of claim 4 , wherein the multi-armed bandit API includes a first model comprising parameters associated with the distribution channel and the feedback data.

16. The computer cluster architecture of claim 15 , wherein the feedback data indicates a time associated with a user interaction with the creative.

17. The computer cluster architecture of claim 15 , wherein:
the distribution channel is a first distribution channel,
the creative is a first creative, and
the multi-armed bandit API is further configured to:
determine a second model associated with (i) a second creative associated with the first creative or (ii) a second distribution channel associated with the first creative, and
update the second model based on the received feedback data.

18. The computer cluster architecture of claim 1 , wherein the multi-armed bandit API is configured to update the creative.

19. A method implemented by a computer cluster architecture, wherein the computer cluster architecture comprises at least one computer cluster including at least one computing node, and wherein the at least one computing node comprises:
an interface database containing a plurality of graphical user interfaces (GUIs), the plurality of GUIs including a creation interface, an editor interface, an upload interface, a plurality of approval interfaces, one or more preview interfaces, and one or more save interfaces,
a self-serve tool in data communication with the interface database and configured to be accessed by a user through a portal, the self-serve tool comprising one or more of the plurality of GUIs and provide the one or more of the plurality of GUIs for the user upon receipt of an input by the user through the portal,
a multi-armed bandit application programming interface (API) in data communication with the self-serve tool, the multi-armed bandit API comprising rules for analyzing creatives published by the self-serve tool and for determining a winning creative from the creatives, and
a listener continually or periodically polling feedback data of the creatives for processing by the multi-armed bandit API,

the method comprising:
sending, by the self-serve tool, to a device a creative configured for display on the device;
receiving, by the multi-armed bandit API, from the device feedback data associa

Assignees

• Capital One Services Llc

Inventors

• Kulkarni Kirankumar
• Darivemula Savio Joseph
• Konduru Anil
• Patel Gunjan
• Shen Jie
• Birch Kelly L
• Manion Patrick James

Classifications

• H04L51/02 Primary

  using automatic reactions or user delegation, e.g. automatic replies or chatbot-generated messages · CPC title

• G06F16/9535

  Search customisation based on user profiles and personalisation · CPC title

• G06N5/04

  Inference or reasoning models · CPC title

• G06N20/00

  Machine learning · CPC title

• G06N7/01

  Probabilistic graphical models, e.g. probabilistic networks · CPC title