Context-sensitive, adaptive, and federated system for threat intelligence

DE202026106097U1Undetermined Publication Date: 2026-11-26GUPTA GAURAV KUMAR +1
0 Cites 0 Cited by

Patent Information

Application Number
DE202026106097
Authority / Receiving Office
DE · DE
Patent Type
Utility models
Current Assignee / Owner
Filing Date
2026-09-25
Publication Date
2026-11-26
Estimated Expiration
2036-09-30
Patent Text Reader

Abstract

A context-aware, adaptive, federated threat intelligence system (100), comprising: a variety of local threat intelligence nodes (LTINs) (102), each LTIN (102) being deployed within a corresponding organization and communicatively coupled to one or more local cybersecurity sources, the one or more local cybersecurity sources comprising: at least one of the following: a security information and event management (SIEM) system, an intrusion detection / prevention (IDS / IPS) system, an endpoint detection and response (EDR) system, a firewall, a NetFlow platform, and an external cyber threat intelligence feed, each LTIN (102) being configured to: locally capture a variety of heterogeneous security events; and preprocess the variety of heterogeneous security events by performing at least one of the following steps: data cleansing, timestamp synchronization,Duplicate exclusion and event normalization, wherein each LTIN (102) is further configured to: correlate the preprocessed security events to one or more attack sessions; and extract from each attack session a variety of cyber threat entities that include at least one of the following: Indicators of Compromise (IoCs), affected assets, vulnerabilities, attack behavior, malware artifacts, user activity, network connections, and MITRE-ATT&CK mappings; a Semantic Threat Context Profile (STCP) (104) created by each LTIN (102) for each attack session, wherein the STCP (104) represents an operational context of a corresponding cyberattack and includes an entity layer, a behavior layer, a time layer, and a context layer, the entity layer including information associated with IoCs, affected assets, vulnerabilities, and malware artifacts.the behavioral layer includes information related to attack behavior, attack sequences, and MITRE-ATT&CK tactics and techniques; the temporal layer includes the chronological sequence and relationships between security events; and the context layer includes the severity of the threat, the reliability of the source, and the environmental context; a context-aware threat intelligence network (CATIN) (106) used by each LTIN (102) to process the STCPs (104) and generate an optimized local threat model by learning structural relationships and temporal evolution related to the relevant cyberattacks, with the multitude of LTINs (102) configured to generate privacy-preserving local model updates from the optimized local threat models.without transmitting raw cybersecurity data beyond the appropriate organizational trust boundaries; and a federated collaboration server (108) that is communicatively coupled to the multitude of LTINs (102), wherein the federated collaboration server (108) is configured to receive the privacy-preserving local model updates, evaluate a Threat Intelligence Contribution Score associated with each local model update, and perform a weighted secure aggregation of the local model updates, wherein the federated collaboration server (108) is further configured to apply Federated Adaptive Moment Optimization (FedAdam) to the weighted aggregated local model updates to generate a global threat intelligence model (110), wherein the federated collaboration server (108) is configured toto distribute the global threat intelligence model (110) to the multitude of LTINs (102), each LTIN (102) being configured to use the global threat intelligence model (110) to perform threat prediction, contextual threat interpretation, adaptive threat prioritization, risk assessment, and a recommendation for countermeasures based on characteristics of a detected threat, each LTIN (102) being further configured to collect validated operational feedback related to the detected threat and to use this validated operational feedback to refine the STCPs (104), update threat reliability information, and retrain the CATIN (106) for a subsequent round of federated learning, thereby enabling continuous development of threat intelligence across the multitude of distributed organizations.
Need to check novelty before this filing date? Find Prior Art