Deep Learning with Dense Random Neural Networks for Detecting Attacks Against IoT-Connected Home Environments

Brun, Olivier; Yin, Yilong; Gelenbe, Erol; Kadioglu, Yasin Murat; Augusto-Gonzalez, Javier; Ramos, Manuel

doi:10.1007/978-3-319-95189-8_8

Cited by 70 publications

(51 citation statements)

References 16 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Using ML, a variety of research efforts resulted in remarkable achievements in the area of speech recognition [5] [6] and image recognition [7] [8]. Likewise, it is also an accepted fact that the advancements in machine learning have started a new era for artificial intelligence as well as paved the way for the development of intelligent intrusion detection systems [3] [9] . Although, the detection of anomalous patterns from a given data seems straight forward but there are several challenges which make this task difficult to achieve, such as:…”

Section: Introductionmentioning

confidence: 99%

A Novel Random Neural Network Based Approach for Intrusion Detection Systems

Qureshi

Larijani

Ahmad

et al. 2018

2018 10th Computer Science and Electronic Engineering (CEEC)

View full text Add to dashboard Cite

Computer security and privacy of user specific data is a prime concern in day to day communication. The mass use of internet connected systems has given rise to many vulnerabilities which includes attacks on smart devices. Regular occurrence of such events has made the availability of scalable Intrusion Detection System (IDS) a perilous challenge. An intelligent IDS should be able to stop the malicious activity before it destabilizes the core network and to achieve this goal we propose a novel Random Neural Network based Intrusion Detection System (RNN-IDS) in this paper. The performance is evaluated by training different numbers of input and hidden layer neurons with learning rates on benchmark NSL-KDD dataset for binary classification. To validate the feasibility of proposed scheme, results were compared with existing systems and its performance was evaluated by the detection of novel attacks while obtaining an accuracy of 94.50%.

show abstract

Section: Introductionmentioning

confidence: 99%

A Novel Random Neural Network Based Approach for Intrusion Detection Systems

Qureshi

Larijani

Ahmad

et al. 2018

2018 10th Computer Science and Electronic Engineering (CEEC)

View full text Add to dashboard Cite

show abstract

“…Recent advancements have implemented RNN models to predict the toxicity levels of chemical compounds based on their physicalchemical structure [30]. In [31] a deep learning model based on dense random neural networks for the detection of denial-of-service network attacks against Internet of Things gateways is presented. Our proposal utilises the original single cell architecture of RNN [30], however with a hybrid Multichannel RNN and MLP network for improved accuracy.…”

Section: Related Workmentioning

confidence: 99%

“…Our proposal utilises the original single cell architecture of RNN [30], however with a hybrid Multichannel RNN and MLP network for improved accuracy. Furthermore, we achieve high accuracy comparable to CNN without the dense RNN clusters and the proposed learning algorithms in [24,27,31]. Our work utilises RMSProp algorithm introduced by Hinton [32], which adapts resilient propagation (RProp) [33] algorithm for stochastic gradient descent (SGD) [34].…”

Section: Related Workmentioning

confidence: 99%

Hrnn4f: Hybrid Deep Random Neural Network for Multi-Channel Fall Activity Detection

Tahir¹,

Ahmad

Morison

et al. 2019

Prob. Eng. Inf. Sci.

View full text Add to dashboard Cite

Falls are a major health concern in older adults. Falls lead to mortality, immobility and high costs to social and health care services. Early detection and classification of falls is imperative for timely and appropriate medical aid response. Traditional machine learning models have been explored for fall classification. While newly developed deep learning techniques have the ability to potentially extract high-level features from raw sensor data providing high accuracy and robustness to variations in sensor position, orientation and diversity of work environments that may skew traditional classification models. However, frequently used deep learning models like Convolutional Neural Networks (CNN) are computationally intensive. To the best of our knowledge, we present the first instance of a Hybrid Multichannel Random Neural Network (HMCRNN) architecture for fall detection and classification. The proposed architecture provides the highest accuracy of 92.23% with dropout regularization, compared to other deep learning implementations. The performance of the proposed technique is approximately comparable to a CNN yet requires only half the computation cost of the CNN-based implementation. Furthermore, the proposed HMCRNN architecture provides 34.12% improvement in accuracy on average than a Multilayer Perceptron.

show abstract

“…There are some significant research attempts towards utilizing neural network for strengthening the security system in MANET. Most recently, the work carried out by Brun et al [53] have make use of the advanced version of the neural network i.e. deep learning approach for the purpose of identifying the security threat in MANET.…”

Section: Ii) Neural Network Based Security Solutionmentioning

confidence: 99%

Insights on Security Improvements and Implications of Artificial Intelligence in MANET

T.M.¹,

Shivakumar²

2019

CAE

View full text Add to dashboard Cite

Mobile Adhoc Network (MANET) is claimed to be an integral part of all futuristic network and communication system e.g. cloud, reconfigurable network, Internet-of-Things (IoT), interdomain routing, etc. However, the security problems of MANET is not yet been mitigated irrespective of archives of security approaches and solution. For a MANET system to be integrated with upcoming technologies it is required that their routing system should offer higher level of resistance against all potential threats as well as they should also offer a better flexibility to tailor the security scheme based on demand of the environment, where the mobile nodes are operating. It is also seen that Artificial Intelligence is another big boon which is currently being proliferating at faster speed and there are some segment of research where it is also claimed to contribute towards security system in MANET. Therefore, this paper offers insights to the current state of security system using conventional as well as using Artificial Intelligence to explore about its research gap.

show abstract

Deep Learning with Dense Random Neural Networks for Detecting Attacks Against IoT-Connected Home Environments

Cited by 70 publications

References 16 publications

A Novel Random Neural Network Based Approach for Intrusion Detection Systems

A Novel Random Neural Network Based Approach for Intrusion Detection Systems

Hrnn4f: Hybrid Deep Random Neural Network for Multi-Channel Fall Activity Detection

Insights on Security Improvements and Implications of Artificial Intelligence in MANET

Contact Info

Product

Resources

About