RT info:eu-repo/semantics/article
T1 Network intrusion detection based on extended RBF neural network with offline reinforcement learning
A1 López Martín, Manuel
A1 Sánchez Esguevillas, Antonio Javier
A1 Arribas Sánchez, Juan Ignacio
A1 Carro Martínez, Belén
K1 Communication system security
K1 intrusion detection
K1 Neural networks
K1 Radial basis function networks
K1 Network intrusion detection
K1 33 Ciencias Tecnológicas
K1 3325 Tecnología de las Telecomunicaciones
AB Network intrusion detection focuses on classifying network traffic as either normal or attackcarrier. The classification is based on information extracted from the network flow packets. This is a complexclassification problem with unbalanced datasets and noisy data. This work extends the classic radial basisfunction (RBF) neural network by including it as a policy network in an offline reinforcement learningalgorithm. With this approach, all parameters of the radial basis functions (along with the network weights)are learned end-to-end by gradient descent without external optimization. We further explore how additionaldense hidden-layers, and the number of radial basis kernels influence the results. This novel approach isapplied to five prominent intrusion detection datasets (NSL-KDD, UNSW-NB15, AWID, CICIDS2017and CICDDOS2019) achieving better performance metrics than alternative state-of-the-art models. Eachdataset provides different restrictions and challenges allowing a better validation of results. Analysis ofthe results shows that the proposed architectures are excellent candidates for designing classifiers with theconstraints imposed by network intrusion detection. We discuss the importance of dataset imbalance andhow the proposed methods may be critically important for unbalanced datasets.
PB Institute of Electrical and Electronics Engineers
SN 2169-3536
YR 2021
FD 2021
LK https://uvadoc.uva.es/handle/10324/54210
UL https://uvadoc.uva.es/handle/10324/54210
LA eng
NO IEEE Access, 2021, vol. 9, p. 153153-153170
NO Producción Científica
DS UVaDOC
RD 14-nov-2024