Search

Federated learning (FL) is a decentralized learning technique that enables participating devices to collaboratively build a shared Machine Leaning (ML) or Deep Learning (DL) model without revealing their raw data to a third party. Due to its privacy-preserving nature, FL has sparked widespread attention for building Intrusion Detection Systems (IDS) within the realm of cybersecurity. However, the data heterogeneity across participating domains and entities presents significant challenges for the reliable implementation of an FL-based IDS. In this paper, we propose an effective method called Statistical Averaging (StatAvg) to alleviate non-independently and identically (non-iid) distributed features across local clients' data in FL. In particular, StatAvg allows the FL clients to share their individual data statistics with the server, which then aggregates this information to produce global statistics. The latter are shared with the clients and used for universal data normalisation. It is worth mentioning that StatAvg can seamlessly integrate with any FL aggregation strategy, as it occurs before the actual FL training process. The proposed method is evaluated against baseline approaches using datasets for network and host Artificial Intelligence (AI)-powered IDS. The experimental results demonstrate the efficiency of StatAvg in mitigating non-iid feature distributions across the FL clients compared to the baseline methods., Comment: 10 pages, 8 figures

High-end components that conduct complicated tasks automatically are a part of modern industrial systems. However, in order for these parts to function at the desired level, they need to be maintained by qualified experts. Solutions based on Augmented Reality (AR) have been established with the goal of raising production rates and quality while lowering maintenance costs. With the introduction of two unique interaction interfaces based on wearable targets and human face orientation, we are proposing hands-free advanced interactive solutions in this study with the goal of reducing the bias towards certain users. Using traditional devices in real time, a comparison investigation using alternative interaction interfaces is conducted. The suggested solutions are supported by various AI powered methods such as novel gravity-map based motion adjustment that is made possible by predictive deep models that reduce the bias of traditional hand- or finger-based interaction interfaces

The objective of augmented reality (AR) is to add digital content to natural images and videos to create an interactive experience between the user and the environment. Scene analysis and object recognition play a crucial role in AR, as they must be performed quickly and accurately. In this study, a new approach is proposed that involves using oriented bounding boxes with a detection and recognition deep network to improve performance and processing time. The approach is evaluated using two datasets: a real image dataset (DOTA dataset) commonly used for computer vision tasks, and a synthetic dataset that simulates different environmental, lighting, and acquisition conditions. The focus of the evaluation is on small objects, which are difficult to detect and recognise. The results indicate that the proposed approach tends to produce better Average Precision and greater accuracy for small objects in most of the tested conditions., Comment: 11 pages, 4 figures, conference

Augmented Reality (AR) solutions are providing tools that could improve applications in the medical and industrial fields. Augmentation can provide additional information in training, visualization, and work scenarios, to increase efficiency, reliability, and safety, while improving communication with other devices and systems on the network. Unfortunately, tasks in these fields often require both hands to execute, reducing the variety of input methods suitable to control AR applications. People with certain physical disabilities, where they are not able to use their hands, are also negatively impacted when using these devices. The goal of this work is to provide novel hand-free interfacing methods, using AR technology, in association with AI support approaches to produce an improved Human-Computer interaction solution.

Argumentation mining is a rising subject in the computational linguistics domain focusing on extracting structured arguments from natural text, often from unstructured or noisy text. The initial approaches on modeling arguments was aiming to identify a flawless argument on specific fields (Law, Scientific Papers) serving specific needs (completeness, effectiveness). With the emerge of Web 2.0 and the explosion in the use of social media both the diffusion of the data and the argument structure have changed. In this survey article, we bridge the gap between theoretical approaches of argumentation mining and pragmatic schemes that satisfy the needs of social media generated data, recognizing the need for adapting more flexible and expandable schemes, capable to adjust to the argumentation conditions that exist in social media. We review, compare, and classify existing approaches, techniques and tools, identifying the positive outcome of combining tasks and features, and eventually propose a conceptual architecture framework. The proposed theoretical framework is an argumentation mining scheme able to identify the distinct sub-tasks and capture the needs of social media text, revealing the need for adopting more flexible and extensible frameworks., Comment: Journal of Information Processing & Management, Elsevier - Accepted Version

Smart video sensors for applications related to surveillance and security are IOT-based as they use Internet for various purposes. Such applications include crowd behaviour monitoring and advanced decision support systems operating and transmitting information over internet. The analysis of crowd and pedestrian behaviour is an important task for smart IoT cameras and in particular video processing. In order to provide related behavioural models, simulation and tracking approaches have been considered in the literature. In both cases ground truth is essential to train deep models and provide a meaningful quantitative evaluation. We propose a framework for crowd simulation and automatic data generation and annotation that supports multiple cameras and multiple targets. The proposed approach is based on synthetically generated human agents, augmented frames and compositing techniques combined with path finding and planning methods. A number of popular crowd and pedestrian data sets were used to validate the model, and scenarios related to annotation and simulation were considered., Comment: IoTI4 Workshop 2019. arXiv admin note: text overlap with arXiv:1707.02655

The increased maturity level of technological achievements towards the realization of the Internet of Things (IoT) vision allowed sophisticated solutions to emerge, offering reliable monitoring in highly dynamic environments that lack well-defined and well-designed infrastructures. In this paper, we use a bio-inspired IoT architecture, which allows flexible creation and discovery of sensor-based services offering self-organization and self-optimization properties to the dynamic network, in order to make the required monitoring information available. The main contribution of the paper is the introduction of a new algorithm for following mobile monitored targets/individuals in the context of an IoT system, especially a dynamic one as the aforementioned. The devised technique, called Hot-Cold, is able to ensure proximity maintenance by the tracking robotic device solely based on the strength of the RF signal broadcasted by the target to communicate its sensors' data. Complete geometrical, numerical, simulation, and convergence analyses of the proposed technique are thoroughly presented, along with a detailed simulation-based evaluation that reveals the higher following accuracy of Hot-Cold compared to the popular concept of trilateration-based tracking. Finally, a prototype of the full architecture was implemented not only to demonstrate the applicability of the presented approach for monitoring in dynamic environments, but also the operability of the introduced tracking technique., Comment: Accepted version that replaces the previous pre-print. Pervasive and Mobile Computing, Elsevier, 2020

The field of Argumentation Mining has arisen from the need of determining the underlying causes from an expressed opinion and the urgency to develop the established fields of Opinion Mining and Sentiment Analysis. The recent progress in the wider field of Artificial Intelligence in combination with the available data through Social Web has create great potential for every sub-field of Natural Language Process including Argumentation Mining., Comment: 12th Annual South-East European Doctoral Student Conference (DSC2018), ISBN: 978-960-9416-20-7, pp. 66-74, Thessaloniki, Greece, May 2018

Understanding the financial burden of chronic diseases in developing regions still remains an important economical factor which influences the successful implementation of sensor based applications for continuous monitoring of chronic conditions. Our research focused on a comparison of literature-based data with real costs of the management and treatment of chronic diseases in a developing country, and we are using Kosovo as an example here. The results reveal that the actual living costs exceed the minimum expenses that chronic diseases impose. Following the potential of a positive economic impact of sensor based platforms for monitoring chronic conditions, we further examined the users perception of digital technology. The purpose of this paper is to present the varying cost levels of treating chronic diseases, identify the users concerns and requirements towards digital technology and discuss issues and challenges that the application of sensor based platforms imply in low and middle income countries., Comment: 17th International Symposium on Health Information Management Research, ISHIMR 2015, York, UK, June 2015

The purpose of this paper is to provide a multilayer review of the two major standards in next generation Passive Optical Networks (PONs) and technologies, the ITU-T 10-Gigabit-capable PON (XG-PON) and the IEEE 10 Gigabit Ethernet PON (10G-EPON). A study and a discussion on the standards are performed. The main intention of this paper is to compare XG-PON and 10G-EPON, mainly in terms of physical and data link layers. The paper answers the question of what are the common elements and the basic execution differences of the two standards. Moreover, critical points are raised regarding the Dynamic Bandwidth Allocation (DBA) schemes of both standards. Special focus is also pointed in the coexistence of XG-PON and 10G-EPON. Finally, the paper includes a discussion of open issues and continuing research regarding the two standards., Comment: In Proceedings of the 9th South East European Doctoral Student Conference DSC2014, ISBN 978-960-9416-07-8, ISSN 1791-3578, pp. 286-298, Thessaloniki, Greece, September 2014