Guardado en:
| Autor principal: | |
|---|---|
| Formato: | Recurso digital |
| Lenguaje: | |
| Publicado: |
Zenodo
2025
|
| Acceso en línea: | https://doi.org/10.5281/zenodo.17589250 |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
Tabla de Contenidos:
- <p><span>With the development of network technology, network attack methods are becoming increasingly complex and intelligent. Traditional rule-based defense methods are no longer sufficient to cope with zero-day attacks, advanced persistent threats (APTs), and complex distributed denial-of-service attacks. Artificial intelligence (AI) has shown great application potential in the field of cybersecurity, enabling adaptive, intelligent, and predictive defense. This paper proposes a novel Neural-Heuristic Adaptive Cybersecurity Defense (NeuDef) algorithm. Its core innovation lies in combining a dynamic attack vector prediction mechanism, a heuristic-generative hybrid strategy mechanism, and a multimodal adaptive reinforcement mechanism to achieve forward-looking prediction and intelligent defense against network attacks. This paper systematically constructs the mathematical model and algorithmic framework of NeuDef, elaborates on the fusion method of graph neural networks, generative models, and multi-objective reinforcement learning in the algorithm, and introduces an adaptive weight mechanism to optimize the comprehensive defense utility function. This research provides a theoretical basis and algorithmic design ideas for AI-driven cybersecurity defense.</span></p>