Granted Patent B2

US 12,536,329 · App. 18/264,019 · Granted Jan 27, 2026

Inventors: Kevin Jones (Atlanta, GA); Saravanan Pitchaimani (Atlanta, GA)

Assignee: OneTrust, LLC

G06F21/6254G06F16/285G06N20/20

Abstract

Claims (61)

1 . A method comprising:

generating, by one or more hardware processors utilizing a machine-learning classification model, a label identifying a type of target data associated with a data element used by a data source in handling the target data;

determining, by the one or more hardware processors, a plurality of data samples comprising real occurrences of the type of the target data associated with the data element from the data source;

determining, by the one or more hardware processors and based on the label identifying the type of the target data associated with the data element, that the plurality of data samples is to be anonymized;

generating, by the one or more hardware processors and based on determining that the plurality of data samples is to be anonymized, a plurality of supplemental anonymizing data samples associated with the label and comprising a fictitious occurrence of the type of the target data handled by the data source involving the data element as indicated by the label;

generating, by the one or more hardware processors, a review dataset comprising the plurality of supplemental anonymizing data samples intermingled with the plurality of data samples; and

sending, by the one or more hardware processors, the review dataset over a network to a review computing system.

2 . The method of claim 1 , wherein generating the plurality of supplemental anonymizing data samples comprises:

identifying a subset of supplemental anonymizing data samples from a pool of supplemental anonymizing data samples for a plurality of types of the target data, wherein each supplemental anonymizing data sample of the subset of supplemental anonymizing data samples is associated with the type of the target data; and

selecting the plurality of supplemental anonymizing data samples from the subset of supplemental anonymizing data samples.

3 . The method of claim 1 , wherein generating the plurality of supplemental anonymizing data samples comprises generating each of the plurality of supplemental anonymizing data samples via a random generator.

4 . The method of claim 1 wherein:

generating the label assigned to the data element comprises processing, at a first server, at least one of the plurality of data samples using the machine-learning classification model to generate the label; and

generating the plurality of supplemental anonymizing data samples comprises generating, at a second server, the plurality of supplemental anonymizing data samples.

5 . The method of claim 1 further comprising scanning metadata for the data source to identify the data element as being used to handle the target data.

6 . The method of claim 1 further comprising modifying each of the plurality of data samples and the plurality of supplemental anonymizing data samples to mask the plurality of data samples and the plurality of supplemental anonymizing data samples.

7 . The method of claim 1 , wherein the label comprises data identifying the type of the target data as at least one of a first name, a last name, a telephone number, a social security number, a credit card number, an account number, or an email address.

8 . A non-transitory computer-readable medium having program code that is stored thereon, the program code executable by one or more processing devices for performing operations comprising:

generating, utilizing a machine-learning classification model, labels identifying types of target data associated with data elements used by a data source in handling the target data, the labels comprising a first label of a first type of the target data and a second label of a second type of the target data;

determining, according to the labels, a first plurality of data samples comprising real occurrences of the first type of the target data associated with a first data element from the data source and a second plurality of data samples comprising real occurrences of the second type of the target data associated with a second data element from the data source;

determining, based on the first label assigned to the first data element by the machine-learning classification model, that the first plurality of data samples does not need to be anonymized, wherein the first label identifies the first type of the target data associated with the first data element;

determining, based on the second label assigned to the second data element by the machine-learning classification model, that the second plurality of data samples is to be anonymized, wherein the second label identifies a second type of the target data associated with the second data element;

generating, based on determining that the second plurality of data samples is to be anonymized, a plurality of supplemental anonymizing data samples associated with the second label and comprising a fictitious occurrence of the second type of the target data handled by the data source involving the second data element as indicated by the second label; and

generating a review dataset comprising the first plurality of data samples, the second plurality of data samples, and the plurality of supplemental anonymizing data samples, wherein the plurality of supplemental anonymizing data samples is intermingled with the second plurality of data samples within the review dataset.

9 . The non-transitory computer-readable medium of claim 8 ,

wherein the operations further comprise:

sending the review dataset over a network to a review computing system;

receiving one or more corrected labels for one or more data elements in the review dataset; and

performing one or more additional computing operations utilizing the review dataset with the one or more corrected labels.

10 . The non-transitory computer-readable medium of claim 8 ,

wherein generating the plurality of supplemental anonymizing data samples comprises:

identifying a subset of supplemental anonymizing data samples from a pool of supplemental anonymizing data samples for a plurality of types of the target data, wherein each supplemental anonymizing data sample of the subset of supplemental anonymizing data samples is associated with the second type of the target data; and

selecting the plurality of supplemental anonymizing data samples from the subset of supplemental anonymizing data samples.

11 . The non-transitory computer-readable medium of claim 8 , wherein generating the plurality of supplemental anonymizing data samples comprises generating each of the plurality of supplemental anonymizing data samples via a random generator according to a format based on the second plurality of data samples.

12 . The non-transitory computer-readable medium of claim 8 , wherein the first label and the second label comprise data identifying the first type of the target data and the second type of the target data as at least one of a first name, a last name, a telephone number, a social security number, a credit card number, an account number, or an email address.

13 . A system comprising:

first computing hardware configured to perform a first set of operations comprising:

identifying a data element used by a data source in handling target data;

generating, utilizing a machine-learning classification model, a label identifying a type of target data being handled by the data element by processing a plurality of data samples comprising real occurrences of data for the data element; and

second computing hardware communicatively coupled to the first computing hardware, wherein the second computing hardware is configured to perform a second set of operations comprising:

receiving the label and the plurality of data samples from the first computing hardware;

determining, based on the label generated by the machine-learning classification model, that the plurality of data samples is to be anonymized;

generating, based on determining that the plurality of data samples is to be anonymized, a plurality of supplemental anonymizing data samples, wherein each of the plurality of supplemental anonymizing data samples is associated with the label and comprises a fictitious occurrence of data for the data element; and

generating a review dataset comprising the plurality of supplemental anonymizing data samples intermingled with the plurality of data samples.

14 . The system of claim 13 further comprising:

third computing hardware configured to perform a third set of operations comprising:

receiving the review dataset from the second computing hardware;

processing the review dataset using a rules-based model to determine the data element is not handling the type of target data indicated by the label; and

responsive to determining the data element is not handling the type of target data indicated by the label, generate a notification that the label is incorrect.

15 . The system of claim 13 further comprising:

third computing hardware configured to perform a third set of operations comprising:

providing the review dataset for viewing on a graphical user interface;

receiving an indication via the graphical user interface that the label for the data element is correct; and

responsive to receiving the indication, storing data indicating the data element is used by the data source for handling the type of target data.

16 . The system of claim 13 , wherein the machine-learning classification model comprises a multi-label classification model, wherein the first computing hardware is configured to utilize the multi-label classification model to assign a plurality of labels to the data element based on a plurality of separate portions of the data element.

17 . The system of claim 13 , wherein the first computing hardware is configured for identifying the data element used by the data source in handling the target data by scanning metadata for the data source to identify the data element as being used to handle the target data.

18 . The system of claim 13 , wherein the second computing hardware generates the plurality of supplemental anonymizing data samples by performing the second set of operations further comprising:

identifying a subset of supplemental anonymizing data samples from a pool of supplemental anonymizing data samples for a plurality of types of the target data, each supplemental anonymizing data sample of the subset of supplemental anonymizing data samples is associated with the type of target data; and

selecting the plurality of supplemental anonymizing data samples from the subset of supplemental anonymizing data samples.

19 . The system of claim 13 , wherein the second computing hardware generates the plurality of supplemental anonymizing data samples by performing the second set of operations further comprising generating each of the plurality of supplemental anonymizing data samples via a random generator according to a format based on the plurality of data samples.

20 . The system of claim 13 , wherein the second computing hardware is configured to perform the second set of operations further comprising modifying each of the plurality of data samples and the plurality of supplemental anonymizing data samples to mask the plurality of data samples and the plurality of supplemental anonymizing data samples.

Assignments (2)

Continuity (2)

References Cited (400)