An intrusion detection approach based on improved deep belief network

Tian, Qiuting; Han, Dong‐Guk; Li, Kuan‐Ching; Liu, Xingao; Duan, Letian; Castiglione, Arcangelo

doi:10.1007/s10489-020-01694-4

Cited by 87 publications

(34 citation statements)

References 33 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Compared with the existing method MCAN, DCAN can make use of the complex correlation between multimodal features in a more effective way and extract more discriminative features for images and questions. This exploration of modeling dense intra- and inter-modality interactions has been applied to intelligent transportation [ 42 ], intelligent robot [ 43 ], and other fields [ 44 , 45 , 46 ]. Applying it to a wider range of scenarios will be an inevitable trend in the future.…”

Section: Discussionmentioning

confidence: 99%

An Effective Dense Co-Attention Networks for Visual Question Answering

Han

2020

Sensors

Self Cite

View full text Add to dashboard Cite

At present, the state-of-the-art approaches of Visual Question Answering (VQA) mainly use the co-attention model to relate each visual object with text objects, which can achieve the coarse interactions between multimodalities. However, they ignore the dense self-attention within question modality. In order to solve this problem and improve the accuracy of VQA tasks, in the present paper, an effective Dense Co-Attention Networks (DCAN) is proposed. First, to better capture the relationship between words that are relatively far apart and make the extracted semantics more robust, the Bidirectional Long Short-Term Memory (Bi-LSTM) neural network is introduced to encode questions and answers; second, to realize the fine-grained interactions between the question words and image regions, a dense multimodal co-attention model is proposed. The model’s basic components include the self-attention unit and the guided-attention unit, which are cascaded in depth to form a hierarchical structure. The experimental results on the VQA-v2 dataset show that DCAN has obvious performance advantages, which makes VQA applicable to a wider range of AI scenarios.

show abstract

Section: Discussionmentioning

confidence: 99%

An Effective Dense Co-Attention Networks for Visual Question Answering

Han

2020

Sensors

Self Cite

View full text Add to dashboard Cite

show abstract

“…x y z y (8) However, due to the use of weight clipping to force the Lipschitz constraint to be satisfied, CWGAN still produces poor samples or does not converge in some cases. Therefore, this paper add a gradient penalty term to its input to replace the original weight reduction method [31].…”

Section: Proposed Methodology 1) Cwganmentioning

confidence: 99%

“…Intrusion detection models based on shallow machine learning algorithms exhibit problems such as difficulty in processing large-scale network intrusion data, poor recognition of various new types of attacks, high false alarm rates, and excessive reliance on researchers for feature design and feature selection [4][5]. In recent years, deep learning technologies such as GAN (Generate Adversarial Network) [6], SAE (Stacked Autoencoder) [7], DBN (Deep Belief Network) [8], DNN (Deep Neural Network) [9], LSTM (Long-Term Short-Term Memory) [4] are widely used in the field of intrusion detection. Deep learning technologies can automatically extract high-level abstract features of network data and realize accurate identification of network attacks, which can overcome the limitations of shallow learning algorithms.…”

Section: Introductionmentioning

confidence: 99%

Network Intrusion Detection Based on Conditional Wasserstein Generative Adversarial Network and Cost-Sensitive Stacked Autoencoder

Zhang

Wang

et al. 2020

IEEE Access

View full text Add to dashboard Cite

In the field of intrusion detection, there is often a problem of data imbalance, and more and more unknown types of attacks make detection difficult. To resolve above issues, this paper proposes a network intrusion detection model called CWGAN-CSSAE, which combines improved conditional Wasserstein Generative Adversarial Network (CWGAN) and cost-sensitive stacked autoencoders (CSSAE). First of all, the CWGAN network that introduces gradient penalty and L2 regularization is used to generate specified minority attack samples to reduce the class imbalance of the training dataset. Secondly, the stacked autoencoder is used to intelligently extract the deep abstract features of the network data. Finally, a cost-sensitive loss function is constructed to give a large misclassification cost to a minority of attack samples. Thus, effective detection of network intrusion attacks can be realized. The experimental results based on KDDTest+, KDDTest-21, and UNSW-NB15 datasets show that the CWGAN-CSSAE network intrusion detection model improves the detection accuracy of minority attacks and unknown attacks. In addition, the method in this paper is compared with other existing intrusion detection methods, excellent results have been achieved in performance indicators such as accuracy and F1 score. The accuracy on the above datasets reached 90.34%, 80.78% and 93.27% respectively. The accuracy of U2R on the KDDTest+ and KDDTest-21 datasets both reached 42.50%. The accuracy of R2L on the KDDTest+ and KDDTest-21 datasets reached 54.39% and 52.51%, respectively. And the F1 score on the above datasets reached 91.01%, 87.18% and 93.99% respectively.

show abstract

“…Tian et al [37] firstly preprocessed the data features with probabilistic mass function (PMF) and min-max normalization, and then improved the likelihood function of DBN in the unsupervised training stage, that is, introducing the penalty term combination based on kullback leibler (KL) divergence and non-mean Gaussian distribution to resolve the feature homogenization and over fitting.…”

Section: B Intrusion Detection Technologymentioning

confidence: 99%

Research on Intrusion Detection Based on Particle Swarm Optimization in IoT

et al. 2021

View full text Add to dashboard Cite

With the advent of the "Internet plus" era, the Internet of Things (IoT) is gradually penetrating into various fields, and the scale of its equipment is also showing an explosive growth trend. The age of the "Internet of Everything" is coming. The integration and diversification of IoT terminals and applications make IoT more vulnerable to various intrusion attacks. Therefore, it is particularly important to design an intrusion detection model that guarantees the security, integrity and reliability of the IoT. Traditional intrusion detection technology has the disadvantages of low detection rate and poor scalability, which cannot adapt to the complex and changeable IoT environment. In this paper, we propose a particle swarm optimization-based gradient descent (PSO-LightGBM) for the intrusion detection. In this method, PSO-LightGBM is used to extract the features of the data and inputs it into one-class SVM (OCSVM) to discover and identify malicious data. The UNSW-NB15 dataset is applied to verify the intrusion detection model. The experimental results show that the model we propose is very robust in detecting either normal or various malicious data, especially small sample data such as Backdoor, Shellcode and Worms. INDEX TERMS Intrusion detection, Internet of Things (IoT), particle swarm optimization (PSO), oneclass SVM(OCSVM)

show abstract

An intrusion detection approach based on improved deep belief network

Cited by 87 publications

References 33 publications

An Effective Dense Co-Attention Networks for Visual Question Answering

An Effective Dense Co-Attention Networks for Visual Question Answering

Network Intrusion Detection Based on Conditional Wasserstein Generative Adversarial Network and Cost-Sensitive Stacked Autoencoder

Research on Intrusion Detection Based on Particle Swarm Optimization in IoT

Contact Info

Product

Resources

About