Feedback control for automated messaging adjustments

Patent metadata

Field	Value
Publication number	US-11755656-B2
Application number	US-202217838349-A
Country	US
Kind code	B2
Filing date	Jun 13, 2022
Priority date	Aug 9, 2021
Publication date	Sep 12, 2023
Grant date	Sep 12, 2023

Abstract

A processor may receive data and generate a quantified representation of the data by processing the data using at least one machine learning (ML) algorithm, the quantified representation of the data indicating a sentiment of content of the data. The processor may automatically revise the content of the communications data. The revising may include determining a reaction to the content of the communications data, generating a quantified representation of the reaction, determining a difference between the quantified representation of the reaction and the quantified representation of the communications data, identifying, based on the difference, a portion of the content having an unintended sentiment, and replacing the portion of the content with different content.

First claim

What is claimed is:

1. A method comprising:
receiving, by a processor, text data;
obtaining, by the processor, feedback responsive to replacement content from at least one public source;
generating, by the processor, a quantified representation of the feedback;
generating, by the processor, a quantified representation of the text data by processing the text data using at least one natural language processing (NLP) algorithm, the quantified representation of the text data indicating a sentiment of content of the text data; and
automatically revising, by the processor, the content of the text data, the revising comprising:
determining a difference between quantified data indicative of a reaction and the quantified representation of the text data;
identifying, based on the difference, a portion of the content having an unintended sentiment;
determining which one of the quantified data indicative of the reaction and the quantified representation of the feedback has a preferred sentiment; and
replacing the portion of the content with different content, the replacing comprising inserting the different content within a user interface to thereby automatically adjust a sentiment associated with the user interface, wherein the different content comprises untested content in the text data in response to the quantified representation of the reaction having the preferred sentiment or the different content comprises the replacement content in the text data in response to the quantified representation of the feedback having the preferred sentiment.

2. The method of claim 1 , wherein the revising further comprises:
comparing the quantified representation of the text data to quantified reference data;
identifying, based on the comparing, a second portion of the content having an unintended sentiment; and
replacing the second portion of the content with second different content.

3. The method of claim 1 , wherein the obtaining the feedback comprises:
replacing a third portion of the content with untested replacement content; and
publishing the text data including the untested replacement content and obtaining feedback on the untested replacement content from the at least one public source.

4. The method of claim 1 , wherein the replacing comprises generating at least a portion of the different content by executing a natural language generation algorithm, identifying at least a portion of the different content by performing a thesaurus lookup, receiving at least a portion of the different content as user input, or a combination thereof.

5. The method of claim 1 , further comprising generating the quantified data indicative of the reaction by publishing the text data and obtaining feedback from at least one public source.

6. The method of claim 1 , further comprising generating the quantified data indicative of the reaction by performing machine learning processing of the quantified representation of the text data and a model of past reactions, future reactions, or a combination thereof.

7. The method of claim 6 , further comprising training the model, the training comprising creating future reaction training data through extrapolation of past reaction training data.

8. A system comprising:
a processor; and
a non-transitory computer-readable medium in communication with the processor and storing instructions that, when executed by the processor, cause the processor to perform processing comprising:
receiving text data;
obtaining feedback responsive to replacement content from at least one public source;
generating a quantified representation of the feedback;
generating a quantified representation of the text data by processing the text data using at least one natural language processing (NLP) algorithm, the quantified representation of the text data indicating a sentiment of content of the text data; and
automatically revising the content of the text data, the revising comprising:
determining a difference between quantified data indicative of a reaction and the quantified representation of the text data;
identifying, based on the difference, a portion of the content having an unintended sentiment;
determining which one of the quantified data indicative of the reaction and the quantified representation of the feedback has a preferred sentiment; and
replacing the portion of the content with different content, the replacing comprising inserting the different content within a user interface to thereby automatically adjust a sentiment associated with the user interface, wherein the different content comprises untested content in the text data in response to the quantified representation of the reaction having the preferred sentiment or the different content comprises the replacement content in the text data in response to the quantified representation of the feedback having the preferred sentiment.

9. The system of claim 8 , wherein the revising further comprises:
comparing the quantified representation of the text data to quantified reference data;
identifying, based on the comparing, a second portion of the content having an unintended sentiment; and
replacing the second portion of the content with second different content.

10. The system of claim 8 , wherein the obtaining the feedback comprises:
replacing a third portion of the content with untested replacement content; and
publishing the text data including the untested replacement content and obtaining feedback on the untested replacement content from the at least one public source.

11. The system of claim 8 , wherein the replacing comprises generating at least a portion of the different content by executing a natural language generation algorithm, identifying at least a portion of the different content by performing a thesaurus lookup, receiving at least a portion of the different content as user input, or a combination thereof.

12. The system of claim 8 , wherein the processing further comprises generating the quantified data indicative of the reaction by publishing the text data and obtaining feedback from at least one public source.

13. The system of claim 8 , wherein the processing further comprises generating the quantified data indicative of the reaction by performing machine learning processing of the quantified representation of the text data and a model of past reactions, future reactions, or a combination thereof.

14. The system of claim 13 , wherein the processing further comprises training the model, the training comprising creating future reaction training data through extrapolation of past reaction training data.

15. A method comprising:
receiving, by a processor, communications data;
obtaining, by the processor, feedback responsive to replacement content from at least one public source;
generating, by the processor, a quantified representation of the feedback;
generating, by the processor, a quantified representation of the communications data by processing the communications data using at least one machine learning (ML) algorithm, the quantified representation of the communications data indicating a sentiment of content of the communications data; and
automatically revising, by the processor, the content of the communications data, the revising comprising:
determining a difference between a quantified representation of a reaction to the content of the communications data and the quantified representation of the communications data;
identifying, based on the difference, a portion of the content having an unintended sentiment;
determining which one of the quantified data indicative of the reaction and the quantified repre

Assignees

• Capital One Services Llc

Inventors

• Atkins Phoebe
• Miracolo Max
• Edwards Joshua
• Mcclanahan Brian
• Lin Alexander
• Cheng Lin Ni Lisa
• Vargas Cruz

Classifications

• G06Q30/02 Primary

  Marketing; Price estimation or determination; Fundraising · CPC title

• G06F16/90332 Primary

  Natural language query formulation or dialogue systems · CPC title

• G06F40/30

  Semantic analysis · CPC title

• G06F40/247

  Thesauruses; Synonyms · CPC title

• G06F40/211

  Syntactic parsing, e.g. based on context-free grammar [CFG] or unification grammars · CPC title