IP Library Granted Patent US 12683784
Granted Patent B2
US 12683784 · App. 18/228,059 · Granted Jul 14, 2026

Data analysis systems and methods for detecting anomalies in tokenized datasets

Inventors: Aaron Lee Buchanan (Mint Hill, NC); Saumya Agrawal (Dallas, TX); Bryant Coughlin (Green Bay, WI); DeAngelo Floyd (Charlotte, NC); John Raymond Omernik (Wittenberg, WI)
Assignee: BANK OF AMERICA CORPORATION
H04L9/3213
View Patent ↗
Loading inventors, assignments & file history…
Monitor This Case
Get email alerts when status or documents change.
Order Certified Copies
Most orders are placed with the USPTO same day — all within 24 business hours.
Order via The Patent Place →
Pre-filled with this patent's details
Quick Facts
Patent No.
US 12683784
App. No.
18/228,059
Granted
Jul 14, 2026
Kind
B2
Abstract

Systems, methods, and computer program products are provided herein for detecting anomalies in tokenized datasets. An example method includes receiving a user input associated with a first account credential that defines a sequence of one or more alphanumeric characters and/or symbols and determining one or more credential characteristics of the sequence defined by the first account credential. The example method also includes determining a first token for the first account credential based on the one or more determined credential characteristics and detecting an anomaly for the first token implicating a validity of the first account credential. The method may further include modifying one or more metadata entries of the first token based on the detected anomaly. The anomaly for the first token may be determined based upon the frequency at which the first token is associated with account credential inputs during a time period.

Claims (57)

1 . A system for detecting anomalies in tokenized datasets, the system comprising:

at least one non-transitory storage device; and

at least one processor coupled to the at least one non-transitory storage device, wherein the at least one processor is configured to:

receive a user input associated with a first account credential, wherein the first account credential defines at least a sequence of one or more alphanumeric characters and/or symbols;

determine one or more credential characteristics of the sequence defined by the first account credential, wherein the one or more credential characteristics of the sequence defined by the first account credential define a pattern for the alphanumeric characters and/or symbols of the first account credential;

determine a first token for the first account credential based on the one or more determined credential characteristics, wherein the determined first token is associated with a first pattern of alphanumeric characters and/or symbols that is generated based on one or more account credentials other than the first account credential;

generate a credential length parameter for the first account credential;

generate a character frequency parameter for the first account credential indicative of a probability of the first account credential having one or more repeated alphanumeric characters and/or symbols;

detect an anomaly for the first token implicating a validity of the first account credential, wherein the anomaly for the first token is determined in an instance in which the credential length parameter of the first account credential exceeds a common length probability associated with the first token and in an instance in which the character frequency parameter of the first account credential exceeds a probability of repeated alphanumeric characters and/or symbols associated with the first token;

modify one or more metadata entries of the first token based on the detected anomaly; and

preclude access to one or more system applications for the first account credential.

2 . The system of claim 1 , wherein the processor is further configured to, in response to the modification of the one or more metadata entries, generate a user notification for presentation to an operator associated with the system.

3 . The system of claim 1 , wherein, in detecting the anomaly for the first token, the processor is further configured to:

determine a frequency at which the first token is associated with account credential inputs during a time period; and

detect the anomaly for the first token in an instance in which the determined frequency satisfies a prevalence threshold.

4 . The system of claim 1 , wherein the processor is further configured to:

determine a second token different from the first token for the first account credential based on the one or more determined credential characteristics;

detect an anomaly for the second token implicating a validity of the first account credential; and

modify one or more metadata entries of the second token based on the detected anomaly.

5 . The system of claim 1 , wherein the processor is further configured to generate a word accuracy parameter for the first account credential, and wherein the anomaly for the first token is determined at least in part based upon the word accuracy parameter.

6 . The system of claim 5 , wherein the processor is further configured to generate a symbol position parameter for the first account credential, and wherein the anomaly for the first token is determined at least in part based upon the symbol position parameter.

7 . A computer program product for detecting anomalies in tokenized datasets, the computer program product comprising a non-transitory computer-readable medium comprising code that, when executed, causes an apparatus to:

receive a user input associated with a first account credential, wherein the first account credential defines at least a sequence of one or more alphanumeric characters and/or symbols;

determine one or more credential characteristics of the sequence defined by the first account credential, wherein the one or more credential characteristics of the sequence defined by the first account credential define a pattern for the alphanumeric characters and/or symbols of the first account credential;

determine a first token for associated with the first account credential based on the one or more determined credential characteristics, wherein the determined first token is associated with a first pattern of alphanumeric characters and/or symbols that is generated based on one or more account credentials other than the first account credential;

generate a credential length parameter for the first account credential;

generate a character frequency parameter for the first account credential indicative of a probability of the first account credential having one or more repeated alphanumeric characters and/or symbols;

detect an anomaly for the first token implicating a validity of the first account credential, wherein the anomaly for the first token is determined in an instance in which the credential length parameter of the first account credential exceeds a common length probability associated with the first token and in an instance in which the character frequency parameter of the first account credential exceeds a probability of repeated alphanumeric characters and/or symbols associated with the first token;

modify one or more metadata entries of the first token based on the detected anomaly; and

preclude access to one or more system applications for the first account credential.

8 . The computer program product of claim 7 , further comprising code that, when executed, causes the apparatus to:

determine a frequency at which the first token is associated with account credential inputs during a time period; and

detect the anomaly for the first token in an instance in which the determined frequency satisfies a prevalence threshold.

9 . The computer program product of claim 7 , further comprising code that, when executed, causes the apparatus to:

determine a second token different from the first token for the first account credential based on the one or more determined credential characteristics;

detect an anomaly for the second token implicating a validity of the first account credential; and

modify one or more metadata entries of the second token based on the detected anomaly.

10 . The computer program product of claim 7 , further comprising code that, when executed, causes the apparatus to generate a character frequency parameter for the first account credential indicative of a probability of the first account credential having one or more repeated alphanumeric characters and/or symbols, and wherein the anomaly for the first token is determined in an instance in which the character frequency parameter of the first account credential exceeds a probability of repeated alphanumeric characters and/or symbols associated with the first token.

11 . The computer program product of claim 7 , further comprising code that, when executed, causes the apparatus to generate a word accuracy parameter for the first account credential, and wherein the anomaly for the first token is determined at least in part based upon the word accuracy parameter.

12 . A method for detecting anomalies in tokenized datasets, the method comprising:

receiving a user input associated with a first account credential, wherein the first account credential defines at least a sequence of one or more alphanumeric characters and/or symbols;

determining one or more credential characteristics of the sequence defined by the first account credential, wherein the one or more credential characteristics of the sequence defined by the first account credential define a pattern for the alphanumeric characters and/or symbols of the first account credential;

determining a first token for the first account credential based on the one or more determined credential characteristics, wherein the determined first token is associated with a first pattern of alphanumeric characters and/or symbols that is generated based on one or more account credentials other than the first account credential;

generating a credential length parameter for the first account credential;

generating a character frequency parameter for the first account credential indicative of a probability of the first account credential having one or more repeated alphanumeric characters and/or symbols;

detecting an anomaly for the first token implicating a validity of the first account credential, wherein the anomaly for the first token is determined in an instance in which the credential length parameter of the first account credential exceeds a common length probability associated with the first token and in an instance in which the character frequency parameter of the first account credential exceeds a probability of repeated alphanumeric characters and/or symbols associated with the first token; and

modifying one or more metadata entries of the first token based on the detected anomaly; and

precluding access to one or more system applications for the first account credential.

13 . The method of claim 12 , further comprising:

determining a frequency at which the first token is associated with account credential inputs during a time period; and

detecting the anomaly for the first token in an instance in which the determined frequency satisfies a prevalence threshold.

14 . The method of claim 12 , further comprising:

determining a second token different from the first token for the first account credential based on the one or more determined credential characteristics;

detecting an anomaly for the second token implicating a validity of the first account credential; and

modifying one or more metadata entries of the second token based on the detected anomaly.

15 . The method of claim 12 , further comprising generating a credential length parameter for the first account credential, and wherein the anomaly for the first token is determined in an instance in which the credential length parameter of the first account credential exceeds a common length probability associated with the first token.

16 . The method of claim 12 , further comprising generating a word accuracy parameter for the first account credential, and wherein the anomaly for the first token is determined at least in part based upon the word accuracy parameter.