Training an artificial intelligence engine for real-time monitoring to eliminate false positives

Patent metadata

Field	Value
Publication number	US-12417277-B2
Application number	US-202418760564-A
Country	US
Kind code	B2
Filing date	Jul 1, 2024
Priority date	Apr 29, 2022
Publication date	Sep 16, 2025
Grant date	Sep 16, 2025

Abstract

A system and method for training an artificial intelligence engine for real-time monitoring to eliminate false positives is disclosed. The system includes at least one processor, a communication interface coupled to the processor, and a memory device storing executable code. Executing the executable code causes the processor to receive data from an AI security model, receive data from a false positive database, and correlate both sets of data. The correlated data is used to generate a training dataset and a test dataset used to train a false positive identification model. After evaluating the false positive identification model, an AI engine is applied to user registration. The AI engine includes an AI security model and the false positive identification model. Additionally, a system for evaluating the security of user registration utilizing the false positive identification model is disclosed.

First claim

What is claimed is:

1. A system for training an AI engine for real-time monitoring of false positives during user registration comprising:
at least one processor;
a communication interface communicatively coupled to the at least one processor; and
a memory device storing executable code that, when executed, causes the processor to:
receive a first set of data from at least one AI security model, wherein the data comprises input data from the at least one AI security model and output data from the at least one AI security model, wherein the input data comprises data indicative of user registration, wherein the output data comprises a classification;
receive a second set of data from a false positive database, wherein the false positive database comprises a list of false positive instances;
correlate the first set of data with the second set of data to create a correlated dataset;
from the correlated dataset, create a training dataset and a test dataset;
use the training dataset to train a false positive identification model to identify false positives;
use the test dataset to perform an evaluation of the false positive identification models ability to identify false positives; and
based on the results of the evaluation, apply an AI engine to a user registration system, wherein the AI engine comprises an AI security model and the false positive identification model;

wherein the at least one AI security model is configured to compare user information to personally identifiable information from one or more public or private databases.

2. The system for training an AI engine according to claim 1 , wherein applying the AI engine comprises:
receive the user information from a user attempting to register a new account;
upon receiving user information, extract data from the user information;
apply the at least one AI security model to the extracted data;
receive a result from the at least one AI security model, wherein the result is either a negative result or a positive result;
upon receiving the negative result from the AI security model, allow user registration;
upon receiving the positive result from the AI security model, apply the false positive identification model;
receive a result from the false positive identification model, wherein the result may be a false result or a true result;
upon receiving the false result from the false positive identification model, allow user registration; and
upon receiving the true result from the false positive identification model, block user registration.

3. The system for training an AI engine according to claim 2 , wherein applying the AI engine further comprises
determining a false positive rate,
based on the false positive rate, triggering an alert to modify the at least one AI security model.

4. The system for training an AI engine according to claim 3 , wherein modifying the AI security model comprises:
retraining the AI security model, or
selecting a new AI security model.

5. The system for training an AI engine according to claim 1 , wherein the false positive identification model comprises a neural network.

6. The system for training an AI engine according to claim 5 , wherein the neural network is selected from the group consisting of a convolution neural network (CNN), a recurrent neural network (RNN), and a feed-forward network.

7. The system for training an AI engine according to claim 1 , wherein the false positive identification model comprises a Bayesian machine learning algorithm.

8. A method for training an AI engine for real-time monitoring of false positives during user registration, the method comprising:
receiving a first set of data from at least one AI security model, wherein the data comprises input data from the at least one AI security model and output data from the at least one AI security model, wherein the input data comprises data indicative of user registration, wherein the output data comprises a classification,
receiving a second set of data from a false positive database, wherein the false positive database comprises a list of false positive instances;
correlating the first set of data with the second set of data to create a correlated dataset;
creating a training dataset and a test dataset from the correlated dataset;
training a false positive identification model to identify false positives using the training dataset;
evaluating the false positive identification model using the test dataset;
altering the false positive identification model, based on results of the evaluation; and
applying an AI engine to user registration, based on the results of the evaluation, wherein the AI engine comprises an AI security model and the false positive identification model;
wherein the at least one AI security model is configured to compare user information to personally identifiable information from one or more public or private databases.

9. The method for training an AI engine according to claim 8 , wherein applying the false positive identification model comprises:
receiving the user information;
extracting content from user information;
applying the AI security model to the extracted data;
receiving a result from the AI security model, wherein the result is either a positive result or a negative result;
allowing user registration, based on receiving the negative result;
applying the false positive identification model, based on receiving the positive result;
receiving a result from the false positive identification model, wherein the result is either a false result or a true result;
allowing user registration, based on receiving the false result; and
blocking user registration, based on receiving a true result.

10. The method for training an AI engine according to claim 9 , wherein applying the false positive identification model further comprises:
determining a false positive rate, and
triggering an alert to modify the AI security model.

11. The method for training an AI engine according to claim 8 , wherein the false positive identification model comprises a neural network.

12. The method for training an AI engine according to claim 11 , wherein the neural network is selected from the group consisting of a convolution neural network (CNN), a recurrent neural network (RNN), and a feed-forward network.

13. The method for training an AI engine according to claim 8 , wherein the false positive identification model comprises a Bayesian machine learning algorithm.

14. A system for eliminating false positives comprising:
at least one processor;
a communication interface communicatively coupled to the at least one processor; and
a memory device storing executable code that, when executed, causes the processor to:
receive user information from a user attempting to register a account;
apply an AI security model to the user information, wherein the security model is configured to determine whether or not the user is attempting fraud;
receive a result from the AI security model, wherein the result is either a positive result or a negative result;
apply a false positive identification model, based on receiving the positive result;
receive a result from the false positive identification model, wherein the result is either a false result or a true result; and
allow user registration, based on receiving the false result;

wherein the at least one AI security model is configured to compare user information to personally identifiable information from one or more public or private databases.

15. The system for eliminating false positives according to claim 14 , wherein the memory device furt

Assignees

• Truist Bank

Inventors

• Wright David
• Pham David
• Lewis Adam Thomas
• Cluff Kenneth William

Classifications

• G06N3/047

  Probabilistic or stochastic networks · CPC title

• G06F2221/2117

  User registration · CPC title

• G06N3/09

  Supervised learning · CPC title

• G06N3/088

  Non-supervised learning, e.g. competitive learning · CPC title

• G06N3/0895

  Weakly supervised learning, e.g. semi-supervised or self-supervised learning · CPC title