Back to results
US 12,646,067 B2Granted
Threat identification and mitigation system using machine learning
AssigneeT-Mobile USA, Inc.
InventorsYasmin Karimli, Jan Olav Opedal
Life of this patent
Estimating the term…
Abstract
Various implementations generally relate to systems and methods for collecting application use data generated based on interactions by a user with one or more applications of an enterprise computer system and upon receiving a request from the user to perform an action related to a customer account, applying a trained machine learning model to the application use data and the request to predict a likelihood that the request is fraudulent.
Claims20
A system comprising:
- at least one hardware processor;
- and
- at least one non-transitory memory storing instructions that, when executed by the at least one hardware processor, cause the system to:
- process application use data generated based on interactions by a user with one or more applications of an enterprise computer system;
- generate, based on the application use data, a fraud score that indicates a likelihood that the user will commit fraud against an enterprise associated with the enterprise computer system;
- receive, subsequent to generating the fraud score, a request from the user to perform an action related to a customer account of the enterprise,
- wherein the request comprises a request for a SIM swap;
- generate an anomaly score for the action that indicates a degree of deviation of the action from expected activity associated with the customer account,
- wherein the anomaly score is generated independently from the fraud score;
- apply a trained machine learning model to the fraud score and the anomaly score to predict a likelihood that the action is fraudulent;
- and
- disable access of the user to the one or more applications of the enterprise computer system when the likelihood is greater than a threshold.
9 more elements
A method comprising:
- processing, by an enterprise computer system, application use data that is generated based on interactions by a user with one or more applications of the enterprise computer system to generate a fraud score,
- wherein the fraud score indicates a likelihood that the user will commit fraud against an enterprise associated with the enterprise computer system;
- generating, by the enterprise computer system, an anomaly score for an action requested by the user,
- wherein the anomaly score is generated independently of the fraud score,
- wherein the action comprises purchasing a mobile phone;
- applying, by the enterprise computer system, a trained machine learning model to the fraud score and the anomaly score to output a prediction indicating whether the requested action is fraudulent;
- and
- denying, by the enterprise computer system, execution of the action when the prediction indicates the requested action is fraudulent.
5 more elements
A non-transitory computer-readable storage medium storing executable instructions, the instructions when executed by one or more processors causing the one or more processors to:
- process application use data associated with a customer account of an enterprise, the first application use data generated based on interactions with one or more applications of a computer system associated with the enterprise;
- generate a vulnerability score for the customer account based on the first application use data,
- wherein the vulnerability score indicates a likelihood that the customer account will be a target of fraud;
- receive a request to perform an action associated with the customer account;
- generate an anomaly score for the action,
- wherein the anomaly score is generated independently from the vulnerability score;
- apply a trained machine learning model to the vulnerability score and the anomaly score to predict a likelihood that the action is fraudulent;
- and
- configure the customer account to block the action when the likelihood is greater than a threshold.
6 more elements
Dependencies are inferred from the claim text.