• español
    • English
  • Login
  • español 
    • español
    • English

UniversidaddeCádiz

Área de Biblioteca, Archivo y Publicaciones
Comunidades y colecciones
Ver ítem 
  •   RODIN Principal
  • Producción Científica
  • Artículos Científicos
  • Ver ítem
  •   RODIN Principal
  • Producción Científica
  • Artículos Científicos
  • Ver ítem
JavaScript is disabled for your browser. Some features of this site may not work without it.

LLM-Based Cyberattack Detection Using Network Flow Statistics

Thumbnail
Identificadores

URI: http://hdl.handle.net/10498/38738

DOI: 10.3390/app15126529

ISSN: 2076-3417

Ficheros
OA_2025_0714.pdf (668.5Kb)
Estadísticas
Ver estadísticas
Métricas y Citas
 
Compartir
Exportar a
Exportar a MendeleyRefworksEndNoteBibTexRIS
Metadatos
Mostrar el registro completo del ítem
Autor/es
Gutiérrez Galeano, Leopoldo JesúsAutoridad UCA; Domínguez Jiménez, Juan JoséAutoridad UCA; Schäfer, Jörg; Medina Bulo, María InmaculadaAutoridad UCA
Fecha
2025-06-10
Departamento/s
Ingeniería Informática
Fuente
Applied Sciences (Switzerland), Vol. 15, Núm. 12, 2025
Resumen
Cybersecurity is a growing area of research due to the constantly emerging new types of cyberthreats. Tools and techniques exist to keep systems secure against certain known types of cyberattacks, but are insufficient for others that have recently appeared. Therefore, research is needed to design new strategies to deal with new types of cyberattacks as they arise. Existing tools that harness artificial intelligence techniques mainly use artificial neural networks designed from scratch. In this paper, we present a novel approach for cyberattack detection using an encoder–decoder pre-trained Large Language Model (T5), fine-tuned to adapt its classification scheme for the detection of cyberattacks. Our system is anomaly-based and takes statistics of already finished network flows as input. This work makes significant contributions by introducing a novel methodology for adapting its original task from natural language processing to cybersecurity, achieved by transforming numerical network flow features into a unique abstract artificial language for the model input. We validated the robustness of our detection system across three datasets using undersampling. Our model achieved consistently high performance across all evaluated datasets. Specifically, for the CIC-IDS-2017 dataset, we obtained an accuracy, precision, recall, and F-score of more than 99.94%. For CSE-CIC-IDS-2018, these metrics exceeded 99.84%, and for BCCC-CIC-IDS-2017, they were all above 99.90%. These results collectively demonstrate superior performance for cyberattack detection, while maintaining highly competitive false-positive rates and false-negative rates. This efficacy is achieved by relying exclusively on real-world network flow statistics, without the need for synthetic data generation.
Materias
large language model; machine learning; fine-tuning; deep learning; cybersecurity; network security; cyberthreat; attack detection; multiclass classification; attack classification
Colecciones
  • Artículos Científicos [12034]
  • Articulos Científicos Ing. Inf. [314]
Attribution-NonCommercial-NoDerivatives 4.0 Internacional
Esta obra está bajo una Licencia Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 Internacional

Listar

Todo RODINComunidades y ColeccionesPor fecha de publicaciónAutoresTítulosMateriasEsta colecciónPor fecha de publicaciónAutoresTítulosMaterias

Mi cuenta

AccederRegistro

Estadísticas

Ver Estadísticas de uso

Información adicional

Acerca de...Deposita en RODINPolíticasNormativasDerechos de autorEnlaces de interésEstadísticasNovedadesPreguntas frecuentes

RODIN está accesible a través de

OpenAIREOAIsterRecolectaHispanaEuropeanaBaseDARTOATDGoogle Académico

Enlaces de interés

Sherpa/RomeoDulcineaROAROpenDOARCreative CommonsORCID

RODIN está gestionado por el Área de Biblioteca, Archivo y Publicaciones de la Universidad de Cádiz

ContactoSugerenciasAtención al Usuario