ACM Transactions on Internet Technology最新文献

As Internet of Things devices become widely used in the real world, it is crucial to protect them from malicious intrusions. However, the data scarcity of IoT limits the applicability of traditional intrusion detection methods, which are highly data-dependent. To address this, in this paper we propose the Open-Set Dandelion Network (OSDN) based on unsupervised heterogeneous domain adaptation in an open-set manner. The OSDN model performs intrusion knowledge transfer from the knowledge-rich source network intrusion domain to facilitate more accurate intrusion detection for the data-scarce target IoT intrusion domain. Under the open-set setting, it can also detect newly-emerged target domain intrusions that are not observed in the source domain. To achieve this, the OSDN model forms the source domain into a dandelion-like feature space in which each intrusion category is compactly grouped and different intrusion categories are separated, i.e., simultaneously emphasising inter-category separability and intra-category compactness. The dandelion-based target membership mechanism then forms the target dandelion. Then, the dandelion angular separation mechanism achieves better inter-category separability, and the dandelion embedding alignment mechanism further aligns both dandelions in a finer manner. To promote intra-category compactness, the discriminating sampled dandelion mechanism is used. Assisted by the intrusion classifier trained using both known and generated unknown intrusion knowledge, a semantic dandelion correction mechanism emphasises easily-confused categories and guides better inter-category separability. Holistically, these mechanisms form the OSDN model that effectively performs intrusion knowledge transfer to benefit IoT intrusion detection. Comprehensive experiments on several intrusion datasets verify the effectiveness of the OSDN model, outperforming three state-of-the-art baseline methods by (16.9% ). The contribution of each OSDN constituting component, the stability and the efficiency of the OSDN model are also verified.

The Medical Internet of Things (MIoT) requires extreme information and communication security, particularly for remote consultation systems. MIoT’s integration of physical and computational components creates a seamless network of medical devices providing high-quality care via continuous monitoring and treatment. However, traditional security methods such as cryptography cannot prevent privacy compromise and information leakage caused by security breaches. To solve this issue, this paper proposes a novel Federated Learning Intrusion Detection System (FLIDS). FLIDS combines Generative Adversarial Network (GAN) and Federated Learning (FL) to detect cyber attacks like Denial of Service (DoS), data modification, and data injection using machine learning. FLIDS shows exceptional performance with over 99% detection accuracy and 1% False Positive Rate (FPR). It saves bandwidth by transmitting 3.8 times fewer bytes compared to central data collection. These results prove FLIDS’ effectiveness in detecting and mitigating security threats in Medical Cyber-Physical Systems (MCPS). The paper recommends scaling up FLIDS to use computing resources from multiple mobile devices for better intrusion detection accuracy and efficiency while reducing the burden on individual devices in MIoT.

In this “info-plosion” era, recommendation systems (or recommenders) play a significant role in finding interesting items in the surge of on-line digital activities and e-commerce. Several techniques have been widely applied for recommendation systems, but the cold start and sparsity problems remain a major challenge. The cold start problem occurs when generating recommendations for new users and items without sufficient information. Sparsity refers to the problem of having a large amount of users and items but with few transactions or interactions. In this paper, a novel cross-domain recommendation model, Cross-Domain Evolution Learning Recommendation (abbreviated as CD-ELR), is developed to communicate the information from different domains in order to tackle the cold start and sparsity issues by integrating matrix factorization and recurrent neural network. We introduce an evolutionary concept to describe the preference variation of users over time. Furthermore, several optimization methods are developed for combining the domain features for precision recommendation. Experimental results show that CD-ELR outperforms existing state-of-the-art recommendation baselines. Finally, we conduct experiments on several real-world datasets to demonstrate the practicability of the proposed CD-ELR.

Abstract: The electromyogram (EMG), also known as an EMG, is used to assess nerve impulses in motor nerves, sensory nerves, and muscles. EMS is a versatile tool used in various biomedical applications. It is commonly employed to determine physical health, but it also finds utility in evaluating emotional well-being, such as through facial electromyography. Classification of EMG signals has attracted the interest of scientists since it is crucial for identifying neuromuscular disorders (NMDs). Recent advances in the miniaturization of biomedical sensors enable the development of medical monitoring systems. This paper presents a portable and scalable architecture for machine learning modules designed for medical diagnostics. In particular, we provide a hybrid classification model for NMDs. The proposed method combines two supervised machine learning classifiers with the discrete wavelet transform (DWT). During the online testing phase, the class label of an EMG signal is predicted using the classifiers’ optimal models, which can be identified at this stage. The simulation results demonstrate that both classifiers have an accuracy of over 98%. Finally, the proposed method was implemented using an embedded CompactRIO-9035 real-time controller.

A world of healthcare possibilities has been opened with the development of the Internet of Medical Things and related machine learning, deep learning, and artificial intelligence approaches. It has a broad range of uses: when linked to the Internet, common medical equipment and sensors may gather important data; deep learning and artificial intelligence algorithms use this data to understand symptoms and patterns and allow remote healthcare. There are a large number of people affected by thyroid disorders across the world. The ultrasound-based thyroid nodule detection using traditional methods increased the burden on the expertise. Therefore, alternate methods are required to overcome this problem. In order to facilitate early thyroid disorder detection, this research aims to offer an IoT-based ensemble learning framework. In the proposed ensemble model, three pre-trained models DeiT, Mixer-MLP and Swin Transformer, are used for feature extraction. The mRMR technique is used for relevant feature selection. A total of 24 machine learning models have been trained, and weighted average ensemble learning is employed using the Improved Jaya optimization algorithm and Coronavirus Herd Immunity optimization algorithm. The ensemble model with the improved Jaya optimization algorithm achieved excellent results. The best value for accuracy, precision, sensitivity, specificity, F2-score and ROC-AUC score are 92.83%, 87.76%, 97.66%, 88.89%, 0.9551 and 0.9357, respectively. The main focus of this research is to increase the specificity. A poor value of specificity can lead to a high false positive rate. This situation can increase anxiety and emotionally weaken the patient. The proposed ensemble model with the Improved Jaya optimization algorithm outperformed state-of-the-art techniques and can assist medical experts.

The Internet of Things-enabled Smart Healthcare System (IoT-SHS) is a networked infrastructure of intelligent wearables, software applications, health systems, and services that continuously monitors and transmits patient-sensitive data using an open wireless channel. The conventional security mechanisms are unsuitable for detecting attacks in the dynamic IoT-SHS context due to resource limitations and heterogeneity in low-cost healthcare devices. Deep Learning (DL) solutions for Intrusion Detection System (IDS) and softwarization of the network has the potential to achieve secure network services in the IoT-SHS environment. Motivated by the aforementioned discussion, we propose an intelligent softwarized IDS for protecting the critical infrastructure of the IoT-SHS ecosystem. Specifically, the DL-based IDS is designed using a hybrid cuda Long Short-Term Memory Deep Neural Network (cuLSTM-DNN) algorithm to assist network administrators in efficient decision-making for the generated intrusions. To further bolster the system’s resilience, we suggest a deployment architecture for the proposed CUDA-powered IDS using OpenStack Tacker in a real SDN environment, ensuring that virtual machines can directly utilize the host’s NVIDIA GPU, thereby streamlining and enhancing the network’s operational efficiency. The experimental results using the CICDDoS2019 dataset confirm the effectiveness of the proposed framework over some baseline and recent state-of-the-art techniques.

Advances in blockchain technology have attracted significant attention across the world. The practical blockchain applications emerging in various domains ranging from finance, healthcare, and entertainment, have quickly become attractive targets for adversaries. The novelty of the technology coupled with the high degree of anonymity it provides made malicious activities even less visible in the blockchain environment. This made their robust detection challenging.

This paper presents EtherShield, an novel approach for identifying malicious activity on the Ethereum blockchain. By combining temporal transaction information and contract code characteristics, EtherShield can detect various types of threats and provide insight into the behavior of contracts. The time-interval based analysis used by EtherShield enables expedited detection, achieving comparable accuracy to other approaches with significantly less data. Our validation analysis, which involved over 15,000 Ethereum accounts, demonstrated that EtherShield can significantly expedite the detection of malicious activity while maintaining high accuracy levels (86.52% accuracy with 1 hour of transaction history data and 91.33% accuracy with 1 year of transaction history data).