Gregory Blanc's research while affiliated with Institut Mines-Télécom and other places

... The researchers in [20] provide a comprehensive review of the research on XAI and its applications in cybersecurity. The paper highlights the growing need for AI models that can be understood and interpreted by humans in the context of cybersecurity. ...

... Machinelearning-based encrypted TA [224] 2022 A survey consists of an analysis of IoT traffic data acquisition approaches, a classification of public datasets, a literature evaluation of IoT traffic processing, and a comparison of ML approaches for IoT device classification. ...

... They used Generative Adversarial Imitation Learning (GAIL) as an alternative GAN method with backward reinforcement learning. Shahid et al. [62] presented a method with GAN for IoT Network Traffic Generation. They used GAN for categorical data generation and combined it with a WGAN autoencoder. ...

... In more recent works, researchers attempt to promote solutions concerning different aspects. For instance, research and innovation [18], legal and police [19], historical to maintaining computer security [20], challenges to face the interoperable cyber-security of information [21], barriers that have been present in the implementation of CSIRTs [22]. Finally, at a local level, the ECU-CERT of the Ecuadorian government, the CSIRT of Ecuadorian Consortium for the Development of Research and Academia (CEDIA), and the CSIRT of the National Polytechnic School, which function with traditional services, are still in the implementation and improvement to reach their maturity levels. ...

... In more recent works, researchers attempt to promote solutions concerning different aspects. For instance, research and innovation [18], legal and police [19], historical to maintaining computer security [20], challenges to face the interoperable cyber-security of information [21], barriers that have been present in the implementation of CSIRTs [22]. Finally, at a local level, the ECU-CERT of the Ecuadorian government, the CSIRT of Ecuadorian Consortium for the Development of Research and Academia (CEDIA), and the CSIRT of the National Polytechnic School, which function with traditional services, are still in the implementation and improvement to reach their maturity levels. ...

... In this scenario, an Impersonation Attack could be suffered, where an attacker impersonates an NSM, having total control of the segments, impacting integrity, including confidentiality of the system, with the API being an attack point [11] [12] [13]. If the traffic between VNFS is routed through an insecure physical route, it could suffer replay attacks and MitM affecting confidentiality and can escalate further [11] [12] [15]. In addition, control plane functions such as the network segment selection function (NSSF) are common in several segments. ...

... Authors in [32] studied a security problem during migration problems. They propose monitoring techniques in order to discover security attacks on the migration process. ...

... IoTrelated Objective [19] Identify DDoS attacks as soon as they are launched from IoT devices that were recruited to a botnet [20] Create an IoT anomaly detection service [29] Create an anomaly detection IoT system for smart cities [30] Review machine learning (ML) and deep learning (DL) based intrusion detection systems (IDSs) used in IoT settings [31] Overcome certain limitations of the Density-Based Spatial Clustering of Applications with Noise (DBSCAN) algorithm [32] Compare several ML algorithms for attack and anomaly detection in IoT systems [33] Create an IDS for low-powered and resource-constrained IoT devices, using unsupervised learning [34] Create a real-time, self-training, easily deployed, anomaly-based DDoS detection system for IoT networks [35] Detect compromised IoT devices without sharing data [36] Improve the effectiveness of traditional ML IDSs on low-frequency attacks in high-dimensional networks [37] Propose a federated-based approach for the detection of botnet attacks, using on-device decentralized traffic data [38] Create an anomaly detector for IoT network communication [39] Create a DL-based botnet attack detector for compromised IoT devices [40] Create a software-defined network (SDN) based security mechanism to detect and mitigate DDoS attacks on IoT networks [41] Detect cyber-attacks based on user behavior [42] Create an anomaly-based DDoS attack detector, using autoencoders (AEs) ...

... Shawe and Abbas [11] introduced a standard for increasing accuracy using singular value decomposition (SVD) enhancement by reducing data, and the classification algorithms used is back propagation neural network (BPNN) with an accuracy of 94,34%. Studies in [12]- [17] have higher detection rates and false alarm rates, also integrated of classification and clustering reached better results. In [18] and [19], analysis of fuzzy cluster is used to classify data and achieved better results. ...

... Moreover, a quick and dynamic response should be offered to satisfy the constraints of service level agreement (SLA). Mainly, the present solutions for resource allocation and network slicing are expensive computationally while not supporting the mixed slice requests [5,6]. ...