EUVIMEDEUVIMED entdecken
Crossref

Multimodal AI for Cloud Security: Intelligent Correlation of Network Traffic, Audit Logs, and System Metrics

Adeel Ali

Research Corridor Journal of Engineering Science · 01.02.2026 · 1 (2) · 265-280

DOI: 10.66320/k6evjc57 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Research Corridor Journal of Engineering Science, online erschienen am 01.02.2026. Verfasst von Adeel Ali.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

The rapid evolution of hyper-scale cloud infrastructures has created a security paradox: while observability data is abundant, actionable intelligence is hindered by fragmentation. Heterogeneous telemetry—including network flows, audit logs, and system metrics—is typically analyzed in isolation by traditional Intrusion Detection Systems (IDS) and SIEM platforms. This "fragmented observability" fails to detect sophisticated, multi-stage attacks that manifest as subtle, sub-threshold indicators across multiple domains simultaneously.This article presents a novel Multimodal Artificial Intelligence architecture designed to bridge these gaps by intelligently correlating heterogeneous data through a unified representation learning framework. Unlike simplistic ensemble methods, the proposed system integrates specialized encoders—CNN-Transformers for traffic, BERT-based models for logs, and Bidirectional LSTMs for metrics—with a novel "Cross-Source Intelligence Fabric." This core innovation enforces behavioral consistency, performs adaptive temporal alignment for asynchronous streams, and models deep contextual dependencies.Experimental evaluation on complex attack datasets demonstrates that this multimodal approach achieves a detection effectiveness score of 0.94 F1-score, representing a 41.48% improvement over best-in-class unimodal baselines. Furthermore, the model significantly reduces operational friction by lowering false positive rates by 35% and decreasing detection latency by 28%. The architecture effectively reveals stealthy attack patterns, such as "low-and-slow" data exfiltration and cryptojacking, which remain invisible to isolated monitoring systems. These findings establish that intelligent cross-modal correlation represents a necessary paradigm shift in cloud security analytics, moving the industry beyond siloed, reactive detection toward holistic, context-aware defense.

Zitieren

APA

Ali, A. (2026). Multimodal AI for Cloud Security: Intelligent Correlation of Network Traffic, Audit Logs, and System Metrics. Research Corridor Journal of Engineering Science, 1(2), 265-280. https://doi.org/10.66320/k6evjc57

Vancouver

Ali A. Multimodal AI for Cloud Security: Intelligent Correlation of Network Traffic, Audit Logs, and System Metrics. Research Corridor Journal of Engineering Science. 2026;1(2):265-280. doi:10.66320/k6evjc57

BibTeX
@article{Ali2026,
  author = {Ali, Adeel},
  title = {Multimodal AI for Cloud Security: Intelligent Correlation of Network Traffic, Audit Logs, and System Metrics},
  journal = {Research Corridor Journal of Engineering Science},
  year = {2026},
  volume = {1},
  number = {2},
  pages = {265--280},
  doi = {10.66320/k6evjc57}
}
Kostenlos starten →