Response prediction for electronic communications

Patent metadata

Field	Value
Publication number	US-11907862-B2
Application number	US-202318195653-A
Country	US
Kind code	B2
Filing date	May 10, 2023
Priority date	Jun 4, 2020
Publication date	Feb 20, 2024
Grant date	Feb 20, 2024

Abstract

Systems, methods, and apparatuses are described herein for performing sentiment analysis on electronic communications relating to one or more image-based communications methods, such as emoji. Message data may be received. The message data may correspond to a message that is intended to be sent but has not yet been sent to an application. Using a first machine learning model, one or more subsets of the plurality of emoji may be determined. The one or more subsets of the plurality of emoji may comprise one or more different types and quantities of emoji, and may each correspond to the same or a different sentiment. Using a second machine learning model, one or more emojis may be selected from the one or more subsets. The one or more emojis selected may correspond to responses to the message.

First claim

What is claimed is:

1. A first computing device comprising:
one or more processors; and
memory storing instructions that, when executed by the one or more processors, cause the first computing device to:
determine a first trained machine learning model, wherein the first trained machine learning model was trained, using first training data to modify one or more first weights of a first artificial neural network, to determine subsets comprising at least two different emojis of a plurality of emojis;
determine a second trained machine learning model, wherein the second trained machine learning model was trained, using second training data different from the first training data to modify one or more second weights of a second artificial neural network, to select emojis, from the plurality of emojis, that correspond to one or more predicted responses to one or more messages;
receive data corresponding to a message that has not yet been sent to an application executing on a second computing device;
process, using the first trained machine learning model, the message to determine one or more subsets comprising at least two different emojis of the plurality of emojis, wherein at least one of the one or more subsets corresponds to at least two different emoji types;
select, using the second trained machine learning model and based on the message, one or more second emojis from the one or more subsets of the plurality of emojis; and
transmit the one or more second emojis.

2. The first computing device of claim 1 , wherein the application is a messaging application, wherein the message is intended to be posted in the messaging application.

3. The first computing device of claim 1 , wherein the instructions, when executed by the one or more processors, cause the first computing device to process the message by causing the first computing device to:
weight each of the plurality of emojis based on:
a quantity of the plurality of emojis that correspond to an emoji type; and
a sentiment corresponding to the emoji type.

4. The first computing device of claim 1 , wherein a first emoji type of the at least two different emoji types corresponds to a positive reaction, and wherein a second emoji type of the at least two different emoji types corresponds to a negative reaction.

5. The first computing device of claim 1 , wherein at least one of the one or more subsets further corresponds to:
a first quantity of a first emoji type of the at least two different emoji types, and
a second quantity of a second emoji type of the at least two different emoji types.

6. The first computing device of claim 1 , wherein the first training data comprises a history of messages in one or more of a plurality of applications that each comprise a plurality of different emoji corresponding to one or more different sentiments.

7. The first computing device of claim 1 , wherein the second training data comprises a history of responses to one or more past messages of a history of messages in one or more of a plurality of applications.

8. The first computing device of claim 1 , wherein the instructions, when executed by the one or more processors, cause the first computing device to transmit the one or more second emojis by transmitting a confidence value associated with each of the one or more second emojis.

9. The first computing device of claim 1 , wherein the instructions, when executed by the one or more processors, cause the first computing device to process the message by causing the first computing device to:
determine that a first emoji of the plurality of emojis belongs to a first subset of the one or more subsets based on one or more second emojis of the plurality of emojis.

10. The first computing device of claim 1 , wherein the instructions, when executed by the one or more processors, cause the first computing device to transmit the one or more second emojis by causing the first computing device to:
transmit a count of the one or more second emojis.

11. The first computing device of claim 1 , wherein the message comprises one or more emojis.

12. A method comprising:
determining a first trained machine learning model, wherein the first trained machine learning model was trained, using first training data to modify one or more first weights of a first artificial neural network, to determine subsets comprising at least two different emojis of a plurality of emojis;
determining a second trained machine learning model, wherein the second trained machine learning model was trained, using second training data different from the first training data to modify one or more second weights of a second artificial neural network, to select emojis, from the plurality of emojis, that correspond to one or more predicted responses to one or more messages;
receiving, by a first computing device, data corresponding to a first message that has not yet been sent to a first application, of a plurality of applications, executing on a second computing device;
processing, by the first computing device and using the first trained machine learning model, the first message to determine one or more subsets comprising at least two different emojis of the plurality of emojis, wherein at least one of the one or more subsets corresponds to at least two different emoji types;
selecting, by the first computing device and using the second trained machine learning model and based on the first message, one or more second emojis from the one or more subsets of the plurality of emojis; and
transmitting, by the first computing device, the one or more second emojis.

13. The method of claim 12 , wherein the first application is a messaging application, wherein the first message is intended to be posted in the first application.

14. The method of claim 12 , wherein processing the first message comprises:
weighting each of the plurality of emojis based on:
a quantity of the plurality of emojis that correspond to an emoji type; and
a sentiment corresponding to the emoji type.

15. The method of claim 12 , wherein a first emoji type of the at least two different emoji types corresponds to a positive reaction, and wherein a second emoji type of the at least two different emoji types corresponds to a negative reaction.

16. The method of claim 12 , wherein at least one of the one or more subsets further corresponds to:
a first quantity of a first emoji type of the at least two different emoji types, and
a second quantity of a second emoji type of the at least two different emoji types.

17. The method of claim 12 , wherein transmitting the one or more second emojis comprises transmitting a confidence value associated with each of the one or more second emojis.

18. A method comprising:
determining a first trained machine learning model, wherein the first trained machine learning model was trained, by a first computing device and using first training data to modify one or more first weights of a first artificial neural network, to determine subsets comprising at least two different emojis of a plurality of emojis that correspond to at least two different emoji types;
determining a second trained machine learning model, wherein the second trained machine learning model was trained, by the first computing device and using second training data different from the first training data to modify one or more second weights of a second artificial neural network, to select emojis, from the plurality of emojis, that correspond to one or more predicted responses to one or more messages;
receiving, by the first comput

Assignees

• Capital One Services Llc

Inventors

• Osborn Kevin
• Loucks Eric
• Edwards Joshua
• Bergeron George
• Johnson Kyle
• Lee Brian

Classifications

• G06N3/0464

  Convolutional networks [CNN, ConvNet] · CPC title

• G06N3/09

  Supervised learning · CPC title

• G06N5/04 Primary

  Inference or reasoning models · CPC title

• G06F40/30 Primary

  Semantic analysis · CPC title

• G06N20/00

  Machine learning · CPC title