Search

Anomaly Detection systems based on Machine and Deep learning are the most promising solutions to detect cyberattacks in the industry. However, these techniques are vulnerable to adversarial attacks that downgrade prediction performance. Several techniques have been proposed to measure the robustness of Anomaly Detection in the literature. However, they do not consider that, although a small perturbation in an anomalous sample belonging to an attack, i.e., Denial of Service, could cause it to be misclassified as normal while retaining its ability to damage, an excessive perturbation might also transform it into a truly normal sample, with no real impact on the industrial system. This paper presents a methodology to calculate the robustness of Anomaly Detection models in industrial scenarios. The methodology comprises four steps and uses a set of additional models called support models to determine if an adversarial sample remains anomalous. We carried out the validation using the Tennessee Eastman process, a simulated testbed of a chemical process. In such a scenario, we applied the methodology to both a Long-Short Term Memory (LSTM) neural network and 1-dimensional Convolutional Neural Network (1D-CNN) focused on detecting anomalies produced by different cyberattacks. The experiments showed that 1D-CNN is significantly more robust than LSTM for our testbed. Specifically, a perturbation of 60% (empirical robustness of 0.6) of the original sample is needed to generate adversarial samples for LSTM, whereas in 1D-CNN the perturbation required increases up to 111% (empirical robustness of 1.11).

The effect of electromagnetic radiation on public health is a recurring topic in the societal and political discourse, peaking with the introduction of every new generation of cellular technology. With the introduction of 5G, promising high peak download speeds as well as high-power beams, there is a need to revisit traditional measurement approaches. ICNIRP, a.k.a the International Committee on Non-Ionising Radiation Protection, offers a useful guideline to evaluate the electromagnetic field exposure of living tissues and provides some limits to keep exposure well below the threshold where it is considered harmful. However, modern packet radio technologies such as 5G or Wi-Fi are different from old broadcasting technologies. They deliver high power in very short bursts, spread over a wide band, thus increasing the difficulty of measuring electric fields with traditional instruments, such as spectrum analyzers. In addition, 5G promises a high spatial focusing performance, which means that the field can vary significantly even in a small area. Hence, measurements with a higher spatial density than we can achieve with expensive and bulky spectrum analyzers are urgently needed. Software-defined radios (SDRs), as a size- and cost-efficient alternative, can be used to capture signals in the time domain and thus increase measurement accuracy. However, software-defined radios are not designed to be used as RF power meters. They require accurate calibration and data analysis to ensure the measured power is correct. The aim of this work is to provide a general framework to calibrate SDRs, enabling them to measure RF power and extract the corresponding electric field value. Subsequently, the influence of the SDR parameters on the accuracy of the electric field measurement is investigated. To assess the performance of the proposed calibration framework in a real-life scenario, we rely on our private 5G network with a calibrated SDR to measure the RF power from a 5G network. Our measurements show that the average electric field exposure of 5G networks is well below 1 V/m.

Federated Learning (FL) has emerged as a promising approach to address privacy concerns inherent in Machine Learning (ML) practices. However, conventional FL methods, particularly those following the Centralized FL (CFL) paradigm, utilize a central server for global aggregation, which exhibits limitations such as bottleneck and single point of failure. To address these issues, the Decentralized FL (DFL) paradigm has been proposed, which removes the client–server boundary and enables all participants to engage in model training and aggregation tasks. Nevertheless, as CFL, DFL remains vulnerable to adversarial attacks, notably poisoning attacks that undermine model performance. While existing research on model robustness has predominantly focused on CFL, there is a noteworthy gap in understanding the model robustness of the DFL paradigm. In this paper, a thorough review of poisoning attacks targeting the model robustness in DFL systems, as well as their corresponding countermeasures, are presented. Additionally, a solution called DART is proposed to evaluate the robustness of DFL models, which is implemented and integrated into a DFL platform. Through extensive experiments, this paper compares the behavior of CFL and DFL under diverse poisoning attacks, pinpointing key factors affecting attack spread and effectiveness within the DFL. It also evaluates the performance of different defense mechanisms and investigates whether defense mechanisms designed for CFL are compatible with DFL. The empirical results provide insights into research challenges and suggest ways to improve the robustness of DFL models for future research.