The growth of the Internet of Things (IoT) presents challenges in the field of security. The Intrusion Detection System is an alternative to protecting the internet of things. In this study, we propose an intrusion detection system model that combines unsupervised algorithm and a deep neural network. Autoencoder as unsupervised learning algorithm has a function as a feature extractor that speeds up the learning process on a deep neural network. The performance of a deep learning model depends heavily on the selection of hyperparameters of neural network architecture. In this case, we used Bayesian Hyperparameter Optimization to perform hyperparameter tuning of deep learning models with various activation and weight initialization techniques. The accumulation result is useful to help determine the correct activation function and weight initialization and the hyperparameters that most influence the deep learning model. The results of this study show that Bayesian hyperparameter optimization can improve classification results significantly. Evaluation using the BoT-IoT dataset, the classification accuracy results in deep learning model can reach 99.99%.
Given the tremendous success of the Internet of Things in interconnecting consumer devices, we observe a natural trend to likewise interconnect devices in industrial settings, referred to as Industrial Internet of Things or Industry 4.0. While this coupling of industrial components provides many benefits, it also introduces serious security challenges. Although sharing many similarities with the consumer Internet of Things, securing the Industrial Internet of Things introduces its own challenges...
This paper is an overview of the most used activation functions, classic functions and current functions as well. When we say classic, we mean the first activation functions, the most popular and used in the past. But due to their disadvantages appeared other new activation functions that we refer them as current. These most influential functions are among the most known artificial intelligence activation functions in the research of Machine learning and Deep Learning as well. With each function...
Bayesian Optimization (BO) methods are useful for optimizing functions that are expen- sive to evaluate, lack an analytical expression and whose evaluations can be contaminated by noise. These methods rely on a probabilistic model of the objective function, typically a Gaussian process (GP), upon which an acquisition function is built. The acquisition function guides the optimization process and measures the expected utility of performing an evaluation of the objective at a new point. GPs assume...
Last. Helge Janicke(DMU: De Montfort University)H-Index: 20
view all 4 authors...
Abstract In this paper, we present a survey of deep learning approaches for cyber security intrusion detection, the datasets used, and a comparative study. Specifically, we provide a review of intrusion detection systems based on deep learning approaches. The dataset plays an important role in intrusion detection, therefore we describe 35 well-known cyber datasets and provide a classification of these datasets into seven categories; namely, network traffic-based dataset, electrical network-based...
#2Nour Moustafa(UNSW: University of New South Wales)H-Index: 20
Last. Benjamin Turnbull(UNSW: University of New South Wales)H-Index: 13
view all 4 authors...
Abstract The proliferation of IoT systems, has seen them targeted by malicious third parties. To address this challenge, realistic protection and investigation countermeasures, such as network intrusion detection and network forensic systems, need to be effectively developed. For this purpose, a well-structured and representative dataset is paramount for training and validating the credibility of the systems. Although there are several network datasets, in most cases, not much information is giv...
The Internet of Things (IoT) has been rapidly evolving towards making a greater impact on everyday life to large industrial systems. Unfortunately, this has attracted the attention of cybercriminals who made IoT a target of malicious activities, opening the door to a possible attack to the end nodes. Due to the large number and diverse types of IoT devices, it is a challenging task to protect the IoT infrastructure using a traditional intrusion detection system. To protect IoT devices, a novel e...
#1Youngjun Yoo(POSTECH: Pohang University of Science and Technology)H-Index: 4
Abstract This paper proposes a method to find the hyperparameter tuning for a deep neural network by using a univariate dynamic encoding algorithm for searches. Optimizing hyperparameters for such a neural network is difficult because the neural network that has several parameters to configure; furthermore, the training speed for such a network is slow. The proposed method was tested for two neural network models; an autoencoder and a convolution neural network with the Modified National Institu...
The cloud computing paradigm is changing how businesses operate, providing greater efficiency, tolerance, elasticity and flexibility in computing workloads. Underpinning these changes are multiple data centers, operated by different entities and distributed globally. Despite these benefits, cloud computing presents new classes of cyber-attack, opportunities to attack and processes to subvert. One of the primary strategies to defend against cyber-attacks is the migration process. A secure Virtual...
Intrusion Detection is crucial in cybersecurity. So is the ability to identify the myriad of attacks. Artificial Neural Networks are an established and proven method of accurate classification. There are approaches to make ANN models faster by applying Principal Component Analysis as a feature extractor. However, ANNs are extremely versatile, a wide range of setups can achieve significantly different classification results. The main contribution of this paper is the evaluation of the way the hyp...
In this paper, we introduce a deep learning approach, based on feed-forward neural networks, for big data time series forecasting with arbitrary prediction horizons. We firstly propose a random search to tune the multiple hyper-parameters involved in the method performance. There is a twofold objective for this search: firstly, to improve the forecasts and, secondly, to decrease the learning time. Next, we propose a procedure based on moving averages to smooth the predictions obtained by the dif...
