EUVIMEDEUVIMED entdecken
Crossref

HYBRID AI ARCHITECTURE FOR PREDICTIVE CYBER RISK INTELLIGENCE

KOLAGANI SOWJANYA, K.SHILPIKA, Dr.K SRIDHAR

JDEBM · 2026

DOI: 10.65713/jdebmv2i401 ↗

Kurz gefasst

Fachartikel in der Zeitschrift JDEBM, 2026. Verfasst von KOLAGANI SOWJANYA und 2 weiteren Autor:innen. Die Arbeit nennt 20 Literaturangaben.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Organizations must decide which of thousands of known vulnerabilities and alerts to act on first, yet common practice ranks them by static severity scores that do not estimate how likely an attack is or what it would cost. Data-driven methods have improved individual tasks: machine learning predicts which vulnerabilities will be exploited, deep anomaly detectors flag suspicious network and host behavior, and attack graphs model how an adversary can move between hosts. These capabilities are rarely combined, so their outputs remain separate scores that do not add up to a forecast of business risk. This paper proposes HyCRI, a hybrid AI architecture for predictive cyber risk intelligence that couples symbolic and neural components. A security knowledge graph and rule-based attack-graph generation supply structure; a calibrated exploit-likelihood model, an autoencoder-ensemble anomaly detector, and a relational graph convolutional network supply learned probabilities; and a Bayesian attack graph fuses both into the probability that each critical asset is compromised within a planning horizon, expressed as expected loss. A decision layer ranks mitigations by risk reduction per unit cost and explains each recommendation through attack paths and feature attributions. We formalize each layer, give the risk-update algorithm and a reference implementation stack, and define a leakage-aware evaluation protocol. Empirical results are not yet reported; the protocol and results template specify how the architecture should be validated.

Zitieren

APA

SOWJANYA, K., K.SHILPIKA, & SRIDHAR, D. (2026). HYBRID AI ARCHITECTURE FOR PREDICTIVE CYBER RISK INTELLIGENCE. JDEBM. https://doi.org/10.65713/jdebmv2i401

Vancouver

SOWJANYA K, K.SHILPIKA, SRIDHAR D. HYBRID AI ARCHITECTURE FOR PREDICTIVE CYBER RISK INTELLIGENCE. JDEBM. 2026. doi:10.65713/jdebmv2i401

BibTeX
@article{SOWJANYA2026,
  author = {SOWJANYA, KOLAGANI and K.SHILPIKA and SRIDHAR, Dr.K},
  title = {HYBRID AI ARCHITECTURE FOR PREDICTIVE CYBER RISK INTELLIGENCE},
  journal = {JDEBM},
  year = {2026},
  doi = {10.65713/jdebmv2i401}
}
Kostenlos starten →