# US 2026/0303625 A1 · Recurrent neural network identifying anomalies in instructions data

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2026/0303625 A1
- Document ID: US20260303625A1
- Status: Published application
- Kind: A1
- Application number: 19/089,161
- Filed: 2025-03-25
- Published: 2026-10-01
- Earliest claimed: 2025-03-25
- Assignees: Truist Bank
- Applicants: Truist Bank
- Inventors: Matthew Laine Donlan; Michael Joseph Senuta
- Web page: https://patentgrep.com/patent/US20260303625A1

## CPC

- H04L 63/1425: {Traffic logging, e.g. anomaly detection}
- H04L 41/0609: {based on severity or priority}
- H04L 41/16: using machine learning or artificial intelligence
- H04L: TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION (arrangements common to telegraphic and telephonic communication H04M)

## Abstract

A method and a computing system use a recurrent neural network to identify anomalies in instructions data. Transfer instructions data is received and stored by the computing system. The data includes information to enable an electronic data transfer from a source to a destination. A recurrent neural network is trained to build a machine learning model using previously stored data related to the source. The data is evaluated using the model to identify any anomality representing a predetermined risk level associated with the data. An alert message is generated when an anomality representing the predetermined risk level is identified by the model and the electronic data transfer is processed when no anomality identified by the model exceeds a predetermined risk level.

## Claims

### Claim 1

A computing system using a recurrent neural network to identify anomalies in instructions data, the computing system comprising:
a processor; and
a memory device operatively coupled with the processor, the memory device storing a data validation application having executable code that, when executed, causes the processor to perform steps of:
store in the memory device a first risk level, a second risk level and a third risk level;
receive and store transfer instructions data, the transfer instructions data including information to enable an electronic data transfer from a source to a destination;
train a recurrent neural network to build a machine learning model using previously stored data related to the source;
evaluate the transfer instructions data using the model to identify any anomality representing one of the first through third risk levels associated with the transfer instructions data;
process the electronic data transfer when the first risk level is identified by the model;
process the electronic data transfer when the second risk level is identified by the model and the stored data includes preauthorization by the source; and
generate an alert message when the third risk level is identified by the model.

### Claim 2 (depends on 1)

The computing system according to claim 1 including a secure communication interface in communication with the processor for receiving the data related to the source and the transfer instructions data.

### Claim 3 (depends on 1)

The computing system according to claim 1 wherein the train step includes using previously stored data related to the destination included in the transfer instructions data.

### Claim 4 (depends on 1)

The computing system according to claim 1 including a secure communication interface in communication with the processor for receiving the data related to the destination.

### Claim 5 (depends on 1)

The computing system according to claim 1 wherein the evaluate step includes checking for information related to the destination in at least one external database.

### Claim 6 (depends on 1)

The computing system according to claim 1 wherein the evaluate step includes checking at least one internal compliance rule.

### Claim 7 (depends on 1)

The computing system according to claim 1 wherein the evaluate step includes checking for information related to the destination on networks external to the computing system.

### Claim 8 (depends on 1)

The computing system according to claim 1 including assigning a first transfer fee to the first risk level and a second transfer fee to the second risk level, the second transfer fee being a higher amount than the first transfer fee.

### Claim 9 (depends on 1)

The computing system according to claim 1 including a display device operatively coupled with the processor and displaying the alert message with the assigned risk level.

### Claim 10 (depends on 1)

The computing system according to claim 1 wherein the processor sends the alert message to the source.

### Claim 11

A method of using a recurrent neural network to identify anomalies in instructions data, the method comprising steps of:
storing a first risk level, a second risk level and a third risk level;
receiving and storing transfer instructions data, the transfer instructions data including information to enable an electronic data transfer from a source to a destination;
training a recurrent neural network to build a machine learning model using previously stored data related to the source;
evaluating the transfer instructions data using the model to identify an anomality representing one of the first through third risk levels;
processing the electronic data transfer when the identified anomality represents the first risk level;
processing the electronic data transfer when the identified anomality represents the second risk level and the stored data includes a preauthorization by the source; and
generate an alert message when the identified anomality represents the third risk level.

### Claim 12 (depends on 11)

The method according to claim 11 including a secure communication interface in communication with the processor for receiving the data related to the source and the transfer instructions data.

### Claim 13 (depends on 11)

The method according to claim 11 wherein the training step includes using previously stored data related to the destination included in the transfer instructions data.

### Claim 14 (depends on 13)

The method according to claim 13 including a secure communication interface in communication with the processor for receiving data related to the destination.

### Claim 15 (depends on 11)

The method according to claim 11 wherein the evaluating step includes checking at least one external database for information related to the destination.

### Claim 16 (depends on 11)

The method according to claim 11 wherein the evaluating step includes checking at least one internal compliance rule.

### Claim 17 (depends on 11)

The method according to claim 11 wherein the evaluating step includes checking information on networks external to the computing system for information related to the destination.

### Claim 18 (depends on 11)

The method according to claim 11 wherein the alert message includes the identified anomality.

### Claim 19 (depends on 18)

The computing system according to claim 18 including a display device operatively coupled with the processor and displaying the alert message with the associated risk level.

### Claim 20 (depends on 11)

The method according to claim 11 including assigning a first transfer fee to the first risk level and a second transfer fee to the second risk level, the second transfer fee being a higher amount than the first transfer fee.
