Artificial intelligence-driven data recovery and cybersecurity
AssigneeGE Precision Healthcare LLC
InventorsNivedha Srinivasan, Mouleeswaran Kumar
Life of this patent
Estimating the term…
Abstract
One or more systems, devices, computer program products and/or computer-implemented methods of use provided herein relate to AI-driven data recovery and cybersecurity. In various embodiments, a system can be provided. The system can comprise a memory that can store computer executable components. The system can further comprise a processor that can execute the computer executable components stored in the memory, where the computer executable components can comprise a detection component that can detect an anomaly caused by a cyberattack. The computer executable components can further comprise an isolation component that can isolate, based on detection of the anomaly, one or more computing systems affected by the cyberattack.
Claims20
A system, comprising:
- a memory that stores computer executable components;
- and
- a processor that executes the computer executable components stored in the memory, wherein the computer executable components comprise a multi-stage artificial intelligence (AI) model, the multi-stage AI model comprising:
- a detection component comprising large language model (LLM) agents and machine learning algorithms, wherein the LLM agents:
- perform a security audit by continuously monitoring network traffic, system logs, and system access activities associated with one or more computing systems;
- and
- generate structured audit outputs comprising semantic indicators of system behavior and potential security events based on the monitored data, and the machine learning algorithms detect an anomaly caused by a cyberattack by analyzing data generated during the security audit;
- and
- an isolation component that isolates, based on detection of the anomaly, one or more computing systems affected by the cyberattack.
A computer-implemented method, comprising:
- performing, by a system operatively coupled to a processor, a security audit of one or more computing systems using large language model (LLM) agents, the security audit comprising continuously monitoring network traffic, system logs, and system access activities associated with the one or more computing systems;
- generating structured audit outputs comprising semantic indicators of system behavior and potential security events based on the monitored data;
- detecting, by the system, using one or more machine learning algorithms that analyze data generated during the security audit, an anomaly caused by a cyberattack;
- and
- isolating, by the system, based on the detecting, one or more computing systems affected by the cyberattack from additional computing systems in a network.
A computer program product comprising a non-transitory computer readable memory having program instructions embodied therewith, the program instructions executable by a processor to cause the processor to:
- perform a security audit of one or more computing systems using large language model (LLM) agents that continuously monitor network traffic, system logs, and system access activities associated with the one or more computing systems;
- generate structured audit outputs comprising semantic indicators of system behavior and potential security events based on the monitored data;
- detect, based on analysis of data generated during the security audit using one or more machine learning algorithms, an anomaly caused by a cyberattack;
- and
- isolate, based on detection of the anomaly, one or more computing systems affected by the cyberattack from additional computing systems in a network.
Dependencies are inferred from the claim text.