Soft Computing: Theory Innovations and Problem-Solving Benefits
Permanent URI for this collectionhttps://hdl.handle.net/10125/107448
Browse
Recent Submissions
Item type: Item , Explainable Intrusion Detection System in IoT Scenarios: A Cross-Device Model Training and Evaluation for Traffic Classification(2024-01-03) Ducange, Pietro; Fazzolari, Michela; Marcelloni, FrancescoThe proliferation of IoT devices in our daily lives has raised concerns about the security of transmitted data. Due to their limited resources, IoT devices are vulnerable to malware attacks and cyber threats. Detecting and classifying these attacks is crucial to mitigate their impact. Various intrusion detection techniques have been proposed for IoT, including approaches based on ML and AI. Most of the ML/AI-based intrusion detection techniques, though effective, often lack transparency and trustworthiness. To address these aspects, XAI has emerged as a promising solution, providing insights on AI model decisions. In this work, we describe an explainable IDS in IoT networks which embeds a multi-way FDT as an XAI model for traffic classification. We propose a Cross-Device training and evaluation approach in which we evaluate the generalization capability of the IDS when new devices are connected to the IoT network without retraining the FDT.Item type: Item , EllipScape: A Genetic Algorithm Based Approach to Non-Photorealistic Colored Image Reconstruction for Evolutionary Art(2024-01-03) Datta, Soumil; Walter, CharlesThe image reconstruction problem has uses in several areas, including the medical field, resolution scaling, and art. The goal of image reconstruction is to recreate an image as close to the original image as possible. While in many cases this goal is to nearly match a target image (or, in the case of generative machine learning problems, to match an image to a target group), it can be useful to examine methods of recreating an image with imperfections, often for the purposes of artistic expression or an understanding of the underlying structure of the image being reconstructed. In this paper, we introduce EllipScape -- a colored, non-photorealistic image reconstruction algorithm utilizing a genetic algorithm. This algorithm produces a resulting image that is similar to the original image but is created from ellipses of different sizes and colors. We show that the performance of this algorithm scales well and executes in a reasonable amount of time for an arbitrarily sized image.Item type: Item , Methodology based on linguistic protoforms for activity detection in patients with type 2 diabetes mellitus(2024-01-03) Díaz Jiménez, David; López Ruiz, José Luis; González Lama, Jesús; Espinilla, MacarenaNowadays, activity recognition systems are a very hot topic with a high applicability in almost any field. These types of systems are capable of detecting human activities using Internet of Things devices that incorporate a set of sensors that allow us to collect events associated with such activities. This study presents a general methodology based on linguistic protoforms for human activity detection. This methodology approaches one of the main challenges of this type of systems, multi-occupancy, and for this purpose it incorporates an indoor localisation system. Furthermore, this methodology is applied in a real environment in patients affected by type 2 diabetes mellitus with the aim of enabling health care professionals to check the degree of compliance with the therapeutic contract. Finally, an analysis is conducted of the alignment of the Sustainable Development Goals with this research.Item type: Item , A Study of Different Protocols of Distribution of Information Granularity to Build Consensus in Fuzzy Group Decision-Making(2024-01-03) González-Quesada, Juan Carlos; Pérez, Ignacio Javier; Herrera-Viedma, Enrique; Cabrerizo, FranciscoInformation granularity has been regarded as a crucial design asset whose careful application becomes essential to create more realistic models. In processes of group decision-making, by admitting an average information granularity level, models can capture the diversity of knowledge sources, which allow them to be more reflective of reality. Concretely, a distribution of information granularity along with an optimization of the distribution process have been applied to build consensus while limiting the information loss. Given that there exist different protocols of distribution of information granularity, viz. a symmetric and uniform distribution, an asymmetric but uniform distribution, a symmetric but non-uniform distribution, and an asymmetric and non-uniform distribution, this study aims to discuss how we can take advantage of them to build consensus in group decision-making with fuzzy preference relations. Some numerical experiments are also conducted to analyze the performance and effectiveness of these protocols to build consensus.Item type: Item , Introduction to the Minitrack on Soft Computing: Theory Innovations and Problem-Solving Benefits(2024-01-03) Cabrerizo, Francisco; Pérez, Ignacio; Herrera-Viedma, Enrique