Electronic system for automatically generating resource distributions based on SMS-based instructions using machine learning

Patent metadata

Field	Value
Publication number	US-12093919-B2
Application number	US-202117188979-A
Country	US
Kind code	B2
Filing date	Mar 1, 2021
Priority date	Mar 1, 2021
Publication date	Sep 17, 2024
Grant date	Sep 17, 2024

Abstract

Systems, computer program products, and methods are described herein for automatically generating resource distributions. The present invention may be configured to receive a text-based instruction and parse, using a machine learning model, the text-based instruction to generate a structured resource distribution including predicted distribution elements. The present invention may be configured to generate, based on the structured resource distribution, a resource distribution. In some embodiments, the text-based instruction may include an email message, an SMS message, recorded speech converted to text, text input to a chat function, text recognized in an image, and/or the like.

First claim

What is claimed is:

1. A system for automatically generating resource distributions, the system comprising:
a parser comprising a convolution neural network, a named entity recognition component, an entity identifier component, an entity sorter component, and an assembler component;
a decision engine;
a historical user information data structure comprising historical user data;
a historical resource distribution information data structure comprising historical resource distribution data;
at least one non-transitory storage device; and
at least one processing device coupled to the parser, the decision engine, the historical user information data structure, the historical resource distribution information data structure, and the at least one non-transitory storage device, wherein the at least one processing device is configured to:
train the convolution neural network using annotated linguistic data;
train the convolutional neural network using historical text-based instruction training data comprising historical text-based instructions and historical resource distributions generated in response to receiving the historical text-based instructions;
receive a text-based instruction;
preprocess, using the parser, the text-based instruction into text;
predict, using the named entity recognition component, the convolution neural network, the historical user information data structure, and the historical resource distribution information data structure, predicted distribution entities and predicted distribution elements based on the text;
identify, using the entity identifier component, the historical user information data structure, and the historical resource distribution information data structure, distribution entities and distribution elements from the predicted distribution entities and the predicted distribution elements;
sort, using the entity sorter component and the convolutional neural network, the distribution entities and the distribution elements into pre-defined distribution fields;
assemble, using the assembler component and the convolutional neural network, the distribution entities and the distribution elements in the pre-defined distribution fields into a structured resource distribution;
determine, using the decision engine, whether the structured resource distribution comprises actual distribution elements;
determine, in response to determining that the structured resource distribution does not comprise the actual distribution elements, the actual distribution elements by accessing the historical user data in the historical user information data structure and the historical resource distribution data in the historical resource distribution information data structure; and
generate, based on the actual distribution elements, a resource distribution.

2. The system of claim 1 , wherein the text-based instruction comprises at least one of an email message, an SMS message, recorded speech converted to text, text input to a chat function, or text recognized in an image.

3. The system of claim 1 , wherein the at least one processing device is configured to:
when sorting the distribution entities, sort, based on surrounding words in the text-based instruction and syntactic roots, the distribution entities and the distribution elements into the pre-defined distribution fields.

4. The system of claim 1 , wherein the predicted distribution elements comprise at least one of a predicted sending source retainer, a predicted receiving source retainer, a predicted amount of resources to be distributed, a predicted type of resource distribution, a predicted type of resources to be distributed, a predicted date on which resources are to be distributed, a predicted frequency at which resources are to be distributed, a predicted type of sending source retainer, or a predicted type of receiving source retainer.

5. The system of claim 1 , wherein the at least one processing device is further configured to parse, using the convolutional neural network, the text-based instruction to determine additional information associated with the text-based instruction.

6. The system of claim 5 , wherein the additional information comprises at least one of a sender alias from which the text-based instruction was sent, a recipient alias to which the text-based instruction was sent, an entity associated with the resource distribution, a name of a user, an address of the user, or content of the text-based instruction.

7. The system of claim 1 , wherein the at least one processing device is configured to:
identify multiple resource distributions within the text-based instruction; and
generate, for each resource distribution of the multiple resource distributions, a structured resource distribution.

8. The system of claim 1 , wherein the at least one processing device is configured to provide, to another user, the resource distribution for authorization.

9. The system of claim 1 , wherein the at least one processing device is configured to perform the resource distribution.

10. A computer program product for automatically generating resource distributions comprising a non-transitory computer-readable medium comprising code causing a first apparatus to:
train a convolution neural network using annotated linguistic data, wherein the first apparatus comprises (i) a parser comprising the convolution neural network, a named entity recognition component, an entity identifier component, an entity sorter component, and an assembler component, (ii) a decision engine, (iii) a historical user information data structure comprising historical user data, and (iv) a historical resource distribution information data structure comprising historical resource distribution data;
train the convolutional neural network using historical text-based instruction training data comprising historical text-based instructions and historical resource distributions generated in response to receiving the historical text-based instructions;
receive a text-based instruction;
preprocess, using the parser, the text-based instruction into text;
predict, using the named entity recognition component, the convolution neural network, the historical user information data structure, and the historical resource distribution information data structure, predicted distribution entities and predicted distribution elements based on the text;
identify, using the entity identifier component, the historical user information data structure, and the historical resource distribution information data structure, distribution entities and distribution elements from the predicted distribution entities and the predicted distribution elements;
sort, using the entity sorter component and the convolutional neural network, the distribution entities and the distribution elements into pre-defined distribution fields;
assemble, using the assembler component and the convolutional neural network, the distribution entities and the distribution elements in the pre-defined distribution fields into a structured resource distribution;
determine, using the decision engine, whether the structured resource distribution comprises actual distribution elements;
determine, in response to determining that the structured resource distribution does not comprise the actual distribution elements, the actual distribution elements by accessing the historical user data in the historical user information data structure and the historical resource distribution data in the historical resource distribution information data structure; and
generate, based on the actual distribution elements, a resource distribution.

11. The computer program product of claim 10 , wherein the text-based in

Assignees

• Bank Of America

Inventors

• Roongta Ashwin
• Blogg Aaron
• Li Yvonne Y
• Osborne Leslieann
• Shah Anuj
• Sodano Thomas A
• Zhang Zhexiao

Classifications

• G06F40/30

  Semantic analysis · CPC title

• G06Q20/385

  using an alias or single-use codes · CPC title

• G06N20/00

  Machine learning · CPC title

• G06Q20/386

  using messaging services or messaging apps (using mobile network messaging services for payment, e.g. SMS G06Q20/3255) · CPC title

• G06F40/295 Primary

  Named entity recognition · CPC title