Back to results
US 12,608,503 B2Granted
Systems and methods for sanitizing sensitive data and preventing data leakage using on-demand artificial intelligence models
AssigneeCapital One Services, LLC
InventorsGiorgio Natili, Sudheendra Kumar Kaanugovi
Life of this patent
Estimating the term…
Abstract
Methods and systems are described herein for leveraging artificial intelligence to sanitize sensitive data and prevent the data from leaving the mobile device and/or be exposed to unauthorized third parties. More specifically, methods and systems are described for a novel and unconventional architecture for a data sanitization application, a novel and unconventional delivery format for the data sanitization model, and a novel and unconventional output format of the data sanitization model.
Claims20
A mobile device for local data desensitization involving a data-size-threshold-based desensitization model download and an accuracy-based desensitization model download for performing the local data desensitization, the mobile device comprising:
- one or more processors and memories storing instructions that, when executed by the one or more processors, cause operations comprising:
- in response to a data size of a first dataset accessible on the mobile device exceeding a threshold data size, performing a first desensitization of the first dataset involving permanent local sensitive data deletion by (i) executing a first query of a remote server for a desensitization model to obtain a first machine learning model version trained to classify data instances in locally-stored datasets in order to desensitize the locally-stored datasets and (ii) using the first machine learning model version to identify a first data instance of the first dataset for permanent deletion from the mobile device;
- in response to the mobile device (i) transmitting, to the remote server, classification information related to an accuracy of the first desensitization performed with the first machine learning model version in classifying the first dataset and (ii) executing a second query of the remote server for a desensitization model, obtaining, from the remote server, a second machine learning model version for replacement of the first machine learning model version, wherein the remote server transmits the second machine learning model version in response to determining that the second machine learning model version to the mobile device for the second query in response to the remote server determining, based on the classification information related to the accuracy of the first desensitization in classifying the first dataset, that the second machine learning model version is required to replace the first machine learning model version;
- and
- performing a second desensitization of the first dataset using the second machine learning model version.
2 more elements
A method for desensitizing data using on demand artificial intelligence models, the method comprising:
- in response to a data size of a first dataset accessible on a mobile device exceeding a threshold data size, performing a first desensitization of the first dataset by (i) executing a first query of a remote server for a desensitization model to obtain a first machine learning model version trained to classify data instances in locally-stored datasets in order to desensitize the locally-stored datasets and (ii) using the first machine learning model version to identify a first data instance of the first dataset for permanent deletion from the mobile device;
- in response to the mobile device (i) transmitting, to the remote server, classification information related to an accuracy of the first desensitization performed with the first machine learning model version and (ii) executing a second query of the remote server for a desensitization model, obtaining, from the remote server, a second machine learning model version for replacement of the first machine learning model version, wherein the remote server transmits the second machine learning model version to the mobile device for the second query in response to the remote server determining, based on the classification information related to the accuracy of the first desensitization, that the second machine learning model version is required to replace the first machine learning model version;
- and
- performing a second desensitization of the first dataset using the second machine learning model version.
1 more element
One or more non-transitory, computer-readable media comprising instructions that, when executed by one or more processors, cause operations comprising:
- based on a data size of a first dataset accessible on a mobile device exceeding a threshold data size, performing a first desensitization of the first dataset by (i) executing a first query of a remote server for a desensitization model to obtain a first version of a machine learning model trained to classify data instances in locally-stored datasets in order to desensitize the locally-stored datasets and (ii) using the first version of the machine learning model to identify a first data instance of the first dataset for permanent deletion from the mobile device;
- based on the mobile device (i) transmitting, to the remote server, classification information related to an accuracy of the first desensitization performed with the first version of the machine learning model and (ii) executing a second query of the remote server for a desensitization model, obtaining, from the remote server, a second version of the machine learning model for replacement of the first version of the machine learning model, wherein the remote server transmits the second version of the machine learning model to the mobile device for the second query based on the remote server determining, using the classification information related to the accuracy of the first desensitization, that the second version of the machine learning model is required to replace the first version of the machine learning model;
- and
- performing a second desensitization of the first dataset using the second version of the machine learning model.
1 more element
Dependencies are inferred from the claim text.