1,721,000 research outputs found
Enabling Lightweight Federated Learning in NextG Wireless Networks
No full textNextG wireless will heavily rely on Federated Learning (FL) applications to learn context-aware AI solutions from the massive amount of generated data. Ensuring the reliability of wireless links for such applications is paramount, especially for FL where packet loss can severely hamper performance and efficiency. Traditional approaches fall short under the high packet loss characteristics of wireless networks. This demo shows how the integration of Fountain Codes (FC) into the FL process can bring notable improvements in packet transmission efficiency, especially under high packet loss conditions
Photonic-accelerated AI for cybersecurity in sustainable 6G networks
No full textThe sixth generation (6G) of mobile communications, expected to be deployed around the year 2030, is predicted to be characterized by ubiquitous connected intelligence. With Artificial Intelligence (AI) operations being deployed in every aspect of future network infrastructure, network security will also evolve from current solutions to intelligent architectures. To meet the massive amount of operations computed by AI models, photonic hardware can be exploited, delivering higher processing speed and computing density and lower power consumption with respect to electronic counterparts.In this paper, we propose a photonic-based Convolutional Neural Network (CNN) solution able to work on real-time traffic, capable of identifying Denial of Service (DoS) Hulk attacks with 99.73 mean F1-score when exploiting 4 bits. We also compared photonic accelerators with their electronic counterparts, showing limited F1-score degradation, especially in the 4 and 8 bit scenarios
Real-time clustering based on deep embeddings for threat detection in 6G networks
Full text linkTrials and deployments of sixth Generation (6G) wireless networks, delivering extreme capacity, reliability, and effi ciency, are expected as early as 2030. Attempts from both industry and academia are trying to defi ne the next generation network infrastructure. 6G will set in motion the fourth industrial revolution, delivering global, integrated intelligence. In this scenario, Artifi cial Intelligence (AI)-assisted architecture for 6G networks will realize knowledge discovery, automatic network adjustment and intelligent service provisioning. The long-term vision for implementing 6G security is to implement an autonomous, self-learning AI-assisted architecture that can perform threat mitigation without disrupting the normal use, enabling the resilience and reliability of the network and fulfi lling security automation. This work proposes
a fi rst implementation of a proactive threat discovery service to be deployed at 6G base stations, paving the way for collective network intelligence in the context of cybersecurity mechanisms. Specifi cally, a fully unsupervised Deep Learning (DL) model is presented, able to recognize both Denial of Service (DoS) Hulk and DoS Goldeneye, with 97 . 0% and 92 . 2% mean F1-score respectively
Real-Time Network Packet Classification Exploiting Computer Vision Architectures
Full text linkForthcoming 6G/NextG networks highlight the need for advanced Artificial Intelligence (AI)-based security mechanisms to identify malicious activities and adapt to emerging threats. In this context, the integration of computer vision techniques into the cybersecurity field is promising due to their potential for sophisticated pattern recognition. In this paper we introduce a computationally efficient classification scheme acting directly on the raw packets collected at base stations and enforcing real-time conversion of packets into images. The innovative points of the proposed solution are the lightweight implementation, aligning well with the demands of future 6G networks, and the operation at network edge, enabling early threat identification as close as possible to the packet origin. We investigate the performance of this approach both in terms of F1-score and prediction time using state-of-the-art computer vision architectures and a customized Convolutional Neural Network (CNN) in an intrusion detection problem using a 5G dataset. Experimental results show the superiority of the CNN architecture over complex models. Across multiple packet window sizes (i.e., 10, 50, 100 packets), the CNN consistently outperforms the other state-of-the-art computer vision models, achieving very high F1-scores (namely, 0.99593, 0.99860, 0.99895). A scalability analysis highlights a trade-off between CNN scalability and performance, where larger values lead to increased prediction time. On the other hand, the other computer vision models exhibit better scalability, enabling an optimal model selection without trade-offs
Towards softwarization in the IoT: Integration and evaluation of t-res in the oneM2M architecture
No full textSoftwarization is a systemic trend which appears under several paradigms impacting networks, services and terminals evolution. Even the Internet of Things (IoT) scenario is going to be affected by this revolution. The efforts that are being made to provide IoT objects with embedded logic reconfiguration capabilities and the architectures that are being defined to support standard Machine-2-Machine (M2M) transactions are clear expressions of this trend. In this direction, a powerful framework to enable softwarization in constrained IoT objects is T-Res, where a virtual-machine based design is used to support execution tasks abstraction. To adopt T-Res in a real scenario in which device software functionalities can be remotely managed in an automatic fashion, its integration in a global M2M architecture must be defined. In this paper, the integration of TRes in the oneM2M architecture is presented, and its feasibility concerning application logic execution time and power consumption is analyzed by taking into account new generation devices. Performance results show that although oneM2M compliant TRes requires greater execution times and energy consumption when compared to a monolithic firmware approach, it results to be a feasible solution to be implemented in current and next generation IoT objects in order to provide advanced application logic reconfigurability features in softwarized ecosystems
Protecting NextG Military Networks with Convolutional Neural Networks
No full textNowadays, defense applications consider 5G networks to meet the requirements of the military communications. However, security aspects must be improved to face the challenges of the operational scenarios. Additionally, advanced AI-based security measures are being investigated in forthcoming NextG networks to detect and adapt to emerging threats. In this context, the integration of computer vision techniques in cybersecurity is very promising.In this paper we present a computationally-efficient real-time approach to convert network packets into images directly at base stations. This lightweight implementation aligns well with NextG real-time demands, allowing for the identification of threats at their source. Then we developed a custom Convolutional Neural Network (CNN) operating on the converted packets and aimed at intrusion detection in current and future wireless networks.We then evaluated the performance of this approach to identify and detect malicious content in network traffic utilizing a recent dataset built on a 5G network. Results demonstrate that the designed CNN can achieve high F1-scores, i.e., 0.99593, 0.99860, and 0.99895, across different packet window sizes (10, 50, and 100 packets), indicating that computer vision techniques are promising for detecting malicious network traffic at their source in wireless networks
Photonic-aware Neural Networks for Packet Classification in Beyond 5G Networks
No full textThe benefits introduced by novel network technologies such as 5G and beyond, including low latency and support for billions of devices, have the potential to transform the lives of people. However, the features promised by these new technologies have also attracted malicious actors, with various motivations for attacking the network infrastructure, from cybercrime-based frauds to political goals. Thus, to enable the full potential of the emerging network technologies, it is necessary to take into accounts these attacks and develop tailored countermeasures. One future direction in mitigating the risks of potential attacks is the automatic classification of malicious packets, with the possibility to drop them if classified in the attack category. Hence, in this context, we propose a solution based on Neural Networks (NNs) to automatically classify packets into two classes, i.e., benign and attack, directly in the Radio Access Network (RAN), specifically inspecting packets when they are relayed at the next generation eNB (gNB)-Central Unit (CU) level. Since NNs can be computationally intensive algorithms, potentially increasing the latency of the network, we decide to leverage Photonic-Aware Neural Network (PANN), photonic accelerators able to perform NN computations in the analog optical domain and with time-of-flight latency. We devised two different PANN architectures, considering different photonic constraints. The classification performance of the two architectures has been assessed on the CICIDS-2017 dataset and compared with electronic counterparts. Results proved that the F1-score loss due to underlying hardware constraints is negligible, paving the way for PANN applications in next generation networks
Going Beyond Counting First Authors in Author Co-citation Analysis
Full text linkThe present study examines one of the fundamental aspects of author co-citation analysis (ACA) - the way co-citation
counts are defined. Co-citation counting provides the data on which all subsequent statistical analyses and mappings
are based, and we compare ACA results based on two different types of co-citation counting - the traditional type that
only counts the first one among a cited work's authors on the one hand and a non-traditional type that takes into
account the first 5 authors of a cited work on the other hand. Results indicate that the picture produced through this non-traditional author co-citation counting contains more coherent author groups and is therefore considerably clearer. However, this picture represents fewer specialties in the research field being studied than that produced through the traditional first-author co-citation counting when the same number of top-ranked authors is selected and analyzed. Reasons for these effects are discussed
- …
