AI-Driven Cybersecurity System for Detecting Advanced Persistent Threats in Hybrid Cloud Networks

Abstract

The present invention relates to an AI-Driven Cybersecurity System for Detecting Advanced Persistent Threats in Hybrid Cloud Networks. The AI-Driven Cybersecurity System detects Advanced Persistent Threats (APTs) in hybrid cloud networks by combining federated learning, anomaly detection, and real-time threat intelligence. Federated learning enables local model training, preserving privacy by only aggregating model weights. Anomaly detection engines use LSTM and GNN algorithms to detect APT patterns, while predictive analytics enhance proactive defense. The system employs automated incident response and compliance monitoring with Explainable AI (XAI), ensuring transparency and adaptability in hybrid cloud environments. This system is ideal for industries requiring secure, privacy-conscious cybersecurity solutions. Accompanied Drawing [FIG. 1]

Specification

Description:[001] This invention pertains to cybersecurity, focusing on artificial intelligence (AI) applications for detecting and mitigating cyber threats, specifically Advanced Persistent Threats (APTs), in hybrid cloud environments. By integrating federated learning, anomaly detection, behavioral analysis, and real-time threat intelligence, this invention addresses critical challenges in securing hybrid cloud networks that span on-premises, public cloud, and private cloud infrastructures. The system is especially applicable in domains requiring high levels of security and privacy, including finance, healthcare, smart cities, and enterprise data protection.
BACKGROUND OF THE INVENTION
[002] The following description provides the information that may be useful in understanding the present invention. It is not an admission that any of the information provided herein is prior art or relevant to the presently claimed invention, or that any publication specifically or implicitly referenced is prior art.
[003] The growth of , Claims:1. A cybersecurity system for detecting Advanced Persistent Threats in hybrid cloud networks, comprising:
a federated learning module for local model training and global model aggregation across cloud segments,
an anomaly detection engine utilizing LSTM and GNN models,
a threat intelligence integration module for real-time threat detection and predictive analytics.
2. The system of claim 1, wherein the federated learning module trains models locally, periodically aggregating model weights to create a global model without centralizing data and an anomaly detection module that combines LSTM and GNN networks to identify sequential and relationship-based anomalies in hybrid cloud activity.
3. The system of claim 2, further comprising adaptive thresholding for detection sensitivity adjustment based on network activity and previous detection outcomes, comprising:
Predictive analytics module for identifying potential attack vectors based on threat intelligence and internal network behavior, assigning risk scores to

Documents