Article
Computer Science, Information Systems
Han Qiu, Tian Dong, Tianwei Zhang, Jialiang Lu, Gerard Memmi, Meikang Qiu
Summary: This article presents a method for attacking DL-based network intrusion detection systems using adversarial attacks, with only black-box access and utilizing techniques such as model extraction and saliency maps to generate AEs, successfully compromising a state-of-the-art NIDS system.
IEEE INTERNET OF THINGS JOURNAL
(2021)
Article
Computer Science, Artificial Intelligence
Tianyu Du, Shouling Ji, Bo Wang, Sirui He, Jinfeng Li, Bo Li, Tao Wei, Yunhan Jia, Raheem Beyah, Ting Wang
Summary: This paper presents DetectSec, a platform for analyzing the robustness of object detection models. It conducts a thorough evaluation of adversarial attacks on 18 standard object detection models and compares the effectiveness of different defense strategies. The findings highlight the differences between adversarial attacks and defenses in object detection tasks compared to image classification tasks, and provide insights for understanding and defending against such attacks.
INTERNATIONAL JOURNAL OF INTELLIGENT SYSTEMS
(2022)
Article
Computer Science, Information Systems
Joao Vitorino, Isabel Praca, Eva Maia
Summary: Machine Learning has great value in Network Intrusion Detection, but it is vulnerable to adversarial attacks. Advances in adversarial learning have allowed the generation of realistic examples for ML development and deployment with real network traffic flows. However, challenges and issues still exist when it comes to using adversarial ML in Network Intrusion Detection.
COMPUTERS & SECURITY
(2023)
Article
Computer Science, Information Systems
Xiaokang Zhou, Wei Liang, Weimin Li, Ke Yan, Shohei Shimizu, Kevin I-Kai Wang
Summary: The study introduces a novel adversarial attack generation method to degrade the classification precision of intelligent intrusion detection in IoT systems by identifying critical feature elements and minimal perturbations. The method also develops a hierarchical node selection algorithm based on random walk with restart to select more vulnerable nodes.
IEEE INTERNET OF THINGS JOURNAL
(2022)
Article
Computer Science, Artificial Intelligence
Dengpan Ye, Chuanxi Chen, Changrui Liu, Hao Wang, Shunzhi Jiang
Summary: This paper discusses the saliency map method for enhancing model interpretability, as well as a novel approach combined with additional noises and inconsistency strategy to detect adversarial examples. Experimental results demonstrate that the proposed method effectively detects adversarial attacks with high success rate across common datasets and models, showing its generality compared to existing state-of-the-art techniques.
INTERNATIONAL JOURNAL OF INTELLIGENT SYSTEMS
(2022)
Article
Computer Science, Information Systems
Angel Luis Perales Gomez, Lorenzo Fernandez Maimo, Felix J. Garcia Clemente, Javier Alejandro Maroto Morales, Alberto Huertas Celdran, Gerome Bovet
Summary: Anomaly detection systems based on machine and deep learning are promising solutions for detecting cyberattacks in the industry. This paper presents a methodology to calculate the robustness of these models in industrial scenarios and shows that 1D-CNN is significantly more robust than LSTM for detecting anomalies in a simulated chemical process.
Article
Multidisciplinary Sciences
Ebtihaj Alshahrani, Daniyal Alghazzawi, Reem Alotaibi, Osama Rabie
Summary: The research conducted experiments on adversarial machine learning, indicating that evasion attacks had a significant impact on the accuracy of machine learning-based network intrusion detection systems.
Article
Computer Science, Artificial Intelligence
Hemant Rathore, Animesh Sasan, Sanjay K. Sahay, Mohit Sewak
Summary: This study validates the vulnerability of machine learning-based malware detection models to adversarial samples and proposes countermeasures to improve their accuracy and resistance. The proposed MalDQN agent achieves a high fooling rate and reduces the accuracy of the malware detection models. The defensive strategies significantly enhance the capability of the models to detect and resist adversarial applications.
PATTERN RECOGNITION LETTERS
(2022)
Review
Computer Science, Information Systems
Huseyin Ahmetoglu, Resul Das
Summary: The development of network technologies and the increasing amount of data transferred on networks have led to a rise in cyber threats and attacks. Machine learning offers tools and techniques for automating the detection and analysis of these attacks. This study discusses the different machine learning approaches used to detect and analyze attacks, including anomaly detection, classification, and analysis. The study also examines the performance and results of different methods, as well as the datasets used in the research.
INTERNET OF THINGS
(2022)
Article
Computer Science, Artificial Intelligence
Yinghua Gao, Dongxian Wu, Jingfeng Zhang, Guanhao Gan, Shu-Tao Xia, Gang Niu, Masashi Sugiyama
Summary: Although adversarial training (AT) has been considered as a potential defense against backdoor attacks, it has not yielded satisfactory results and has even strengthened backdoor attacks in some cases. This motivates a comprehensive evaluation of the effectiveness of AT against backdoor attacks in various settings. The research finds that the type and budget of perturbations used in AT are crucial factors, and common perturbations in AT are only effective for specific backdoor trigger patterns. Based on these findings, practical suggestions for backdoor defense, such as relaxed adversarial perturbation and composite AT, are presented. This work not only enhances confidence in AT's ability to defend against backdoor attacks but also provides valuable insights for future research.
IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS
(2023)
Article
Computer Science, Information Systems
Alper Sarikaya, Banu Gunel Kilic, Mehmet Demirci
Summary: This study proposes a method that uses generative adversarial networks to generate adversarial attack data, and designs a robust IDS model to enhance resistance against adversarial attacks. By training machine learning classifiers with multiple feature sets and autoencoders, the proposed model achieves higher accuracy and F1-score.
COMPUTERS & SECURITY
(2023)
Article
Computer Science, Artificial Intelligence
Jiefei Wei, Luyan Yao, Qinggang Meng
Summary: This paper proposes a novel defence method to improve the adversarial robustness of DNN classifiers without using adversarial training. This method introduces two new loss functions to punish overconfidence and protect the network from non-targeted attacks. It also presents a new robustness diagram to analyze and visualize the network's robustness against adversarial attacks and a Log-Softmax-pattern-based adversarial attack detection method.
Article
Computer Science, Artificial Intelligence
Jia Wang, Chengyu Wang, Qiuzhen Lin, Chengwen Luo, Chao Wu, Jianqiang Li
Summary: This paper provides a comprehensive survey of recent advances in adversarial attack and defense methods. It analyzes and compares the pros and cons of various schemes, and discusses the main challenges and future research directions in this field.
Article
Chemistry, Analytical
Ren-Hung Hwang, Jia-You Lin, Sun-Ying Hsieh, Hsuan-Yu Lin, Chia-Liang Lin
Summary: Deep learning technology has developed rapidly and has been successfully applied in various fields, including face recognition. However, most previous studies on adversarial attacks assume the attacker knows the architecture and parameters of the attacked deep learning model, which is not representative of real-world scenarios. This study proposes a Generative Adversarial Network method for generating adversarial patches to carry out dodging and impersonation attacks on a black-box face recognition system, achieving a higher attack success rate than previous works.
Article
Computer Science, Artificial Intelligence
Sekitoshi Kanai, Masanori Yamada, Hiroshi Takahashi, Yuki Yamanaka, Yasutoshi Ida
Summary: Adversarial training is a method to improve against adversarial attacks, but it still lags behind standard training in practical performance. Our analysis reveals that the non-smoothness of the loss function in adversarial training is caused by the constraint of adversarial attacks, which is dependent on the type of constraint. Furthermore, we find that a flatter loss surface in the input space corresponds to a less smooth adversarial loss surface in the parameter space. We demonstrate that smooth adversarial loss achieved through EntropySGD improves the performance of adversarial training.
IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS
(2023)
Editorial Material
Dermatology
Alejandro Garcia-Vazquez, Santiago Guillen-Climent, Marti Pons Benavent, Saray Porcar Saura, Maria Dolores Ramon-Quiles
INDIAN JOURNAL OF DERMATOLOGY VENEREOLOGY & LEPROLOGY
(2023)
Article
Engineering, Electrical & Electronic
Vincenzo Sciancalepore, Lanfranco Zanzi, Xavier Costa-Perez, Antonio Capone
Summary: Virtualization and network slicing provide mobile network operators with the opportunity to deploy multiple logical networks, known as network slices, on their physical network infrastructure. This paper introduces ONETS, an online network slicing solution that incorporates a mathematical model and analytical bounds to maximize multiplexing gains. The feasibility of ONETS is demonstrated through a proof-of-concept implementation on commercial hardware, supporting three network slices and seamless integration with the 3GPP architecture.
IEEE TRANSACTIONS ON WIRELESS COMMUNICATIONS
(2022)
Article
Computer Science, Information Systems
Antonio Albanese, Vincenzo Sciancalepore, Xavier Costa-Perez
Summary: This paper presents SARDO, a drone-based search and rescue solution that leverages mobile phones to localize missing people. SARDO uses pseudo-trilateration and machine-learning techniques to rapidly determine the location of mobile phones with high accuracy and low battery consumption.
IEEE TRANSACTIONS ON MOBILE COMPUTING
(2022)
Article
Computer Science, Information Systems
Antonio Albanese, Vincenzo Sciancalepore, Albert Banchs, Xavier Costa-Perez
Summary: This paper proposes a new base station placement solution that maximizes throughput and localization accuracy by selecting the location of new-generation base stations. This solution enables the provision of location-based services in 5G networks and can be readily applied to current and future roll-out processes.
IEEE TRANSACTIONS ON MOBILE COMPUTING
(2023)
Proceedings Paper
Computer Science, Hardware & Architecture
Antonio Albanese, Francesco Devoti, Vincenzo Sciancalepore, Marco Di Renzo, Xavier Costa-Perez
Summary: Reconfigurable Intelligent Surfaces (RISs) are considered a key disruptive technology for future 6G networks that revolutionize wireless communication by controlling wave propagation properties. However, the need for fast and complex control channels to adapt to changing wireless conditions is a challenge. This paper proposes a self-configuring smart surface solution that can be easily installed in the environment.
IEEE CONFERENCE ON COMPUTER COMMUNICATIONS (IEEE INFOCOM 2022)
(2022)
Proceedings Paper
Telecommunications
Antonio Alhanese, Guillermo Encinas-Lago, Vincenzo Sciancalepore, Xavier Costa-Perez, Dinh-Thuy Phan-Huy, Stephane Ros
Summary: This paper discusses the application of reconfigurable intelligent surface (RIS) technology in wireless networks. By controlling the propagation environment, RIS can improve communication performance and solve dead-zone problems. The authors showcase the capabilities of RIS through theoretical analysis and practical validation in synthetic topologies and real indoor scenarios.
IEEE INTERNATIONAL CONFERENCE ON COMMUNICATIONS (ICC 2022)
(2022)
Proceedings Paper
Computer Science, Interdisciplinary Applications
Leonardo Lo Schiavo, Marco Fiore, Marco Gramaglia, Albert Banchs, Xavier Costa-Perez
Summary: Forecasting is becoming increasingly important for mobile network operations, enabling anticipatory decisions and supporting zerotouch service and network management models. This research presents a hybrid approach that combines statistical modeling and machine learning for predictor design in mobile networks. Experimental results demonstrate that the new model outperforms current state-of-the-art predictors in network resource allocation and mobile traffic anomaly prediction.
2022 IEEE 23RD INTERNATIONAL SYMPOSIUM ON A WORLD OF WIRELESS, MOBILE AND MULTIMEDIA NETWORKS (WOWMOM 2022)
(2022)
Proceedings Paper
Computer Science, Interdisciplinary Applications
Carmen Delgado, Lanfranco Zanzi, Xi Li, Xavier Costa-Perez
Summary: Battery life is a key challenge for collaborative robotics, especially in mission-critical tasks. This paper proposes a novel orchestration approach called OROS, which optimizes robotic navigation, sensing, and infrastructure resources to significantly reduce task completion time.
2022 IEEE 23RD INTERNATIONAL SYMPOSIUM ON A WORLD OF WIRELESS, MOBILE AND MULTIMEDIA NETWORKS (WOWMOM 2022)
(2022)
Article
Computer Science, Information Systems
Jose A. Ayala-Romero, Andres Garcia-Saavedra, Marco Gramaglia, Xavier Costa-Perez, Albert Banchs, Juan J. Alcaraz
Summary: This paper presents vrAIn, a resource orchestrator for vRANs based on deep reinforcement learning. By using an autoencoder to project high-dimensional context data and employing a deep deterministic policy gradient algorithm, vrAIn effectively maps contexts into resource control decisions. Experimental evaluation demonstrates the superior performance of vrAIn in terms of saving computing capacity, improving QoS targets, and increasing throughput.
IEEE TRANSACTIONS ON MOBILE COMPUTING
(2022)
Proceedings Paper
Computer Science, Information Systems
Francesco Linsalata, Antonio Albanese, Vincenzo Sciancalepore, Francesca Roveda, Maurizio Magarini, Xavier Costa-Perez
Summary: This paper explores a novel localization technique for UAVs equipped with cellular base stations in emergency scenarios using OTFS modulation for ToA measurements. The optimal UAV speed is determined as a trade-off between accuracy of ranging technique and power consumption. Results show that the proposed solution outperforms standard PRACH-based localization techniques in terms of RMSE.
2021 IEEE GLOBAL COMMUNICATIONS CONFERENCE (GLOBECOM)
(2021)
Proceedings Paper
Telecommunications
Jose A. Ayala-Romero, Ihtisham Khalid, Andres Garcia-Saavedra, Xavier Costa-Perez, George Iosifidis
Summary: This study evaluates and analyzes the power consumption of virtualized Base Stations (vBS) experimentally, identifying interesting tradeoffs between power savings and performance. Two linear mixed-effect models are proposed to approximate the experimental data, helping to understand the power behavior of vBS and select power-efficient configurations. The release of the experimental dataset aims to encourage further research efforts in this area.
IEEE INTERNATIONAL CONFERENCE ON COMMUNICATIONS (ICC 2021)
(2021)
Article
Computer Science, Information Systems
Xi Li, Carlos Guimaraes, Giada Landi, Juan Brenes, Josep Mangues-Bafalluy, Jorge Baranda, Daniel Corujo, Vitor Cunha, Joao Fonseca, Joao Alegria, Aitor Zabala Orive, Jose Ordonez-Lucena, Paola Iovanna, Carlos J. Bernardos, Alain Mourad, Xavier Costa-Perez
Summary: This article proposes multiple multi-domain solutions for deploying private 5G networks in vertical industries and interconnecting them with public networks. The solutions have been validated in real industry verticals, demonstrating feasibility and efficiency.
Article
Engineering, Electrical & Electronic
Placido Mursia, Francesco Devoti, Vincenzo Sciancalepore, Xavier Costa-Perez
Summary: This study focuses on air-to-ground networks where UAVs equipped with Reconfigurable Intelligent Surfaces (RIS) can provide connectivity over selected areas. By compensating for flight effects, the proposed RiFe algorithm and its practical implementation Fair-RiFe automatically configure RIS parameters to account for undesired UAV oscillations due to adverse atmospheric conditions. Results show that both algorithms provide robustness and reliability, outperforming state-of-the-art solutions in various conditions.
IEEE OPEN JOURNAL OF THE COMMUNICATIONS SOCIETY
(2021)
Proceedings Paper
Computer Science, Hardware & Architecture
Jose A. Ayala-Romero, Andres Garcia-Saavedra, Xavier Costa-Perez, George Iosifidis
Summary: Radio Access Network Virtualization (vRAN) technology will lead the development of flexible radio stacks that adapt to various infrastructure. Research shows that analyzing the energy consumption of virtualized Base Stations (vBSs) is complex and influenced by human behavior, network load, and user mobility, highlighting the potential of machine learning in improving control over virtual base stations.
IEEE CONFERENCE ON COMPUTER COMMUNICATIONS (IEEE INFOCOM 2021)
(2021)