Federated Learning (FL) allows multiple clients to collaboratively train a machine learning model without directly sharing their raw data. While federated learning provides a degree of data privacy, model updates are still susceptible to various inference attacks. This paper presents FLASS, a lightweight and decentralized federated learning framework that leverages additive secret sharing within a multi-server architecture. Each client encodes its local model update into additive shares and shares them with several non-colluding servers, which execute secure aggregation without gaining knowledge of individual contributions. FLASS does not rely on heavy-weight cryptography or a trusted authority. The security analysis and experiments demonstrate the effectiveness and efficiency of the suggested scheme. The analysis indicates that, while achieving the same accuracy as conventional FL schemes, FLASS ensures robust privacy protection with acceptable computational and communication overhead.

Automated assessment of textual disclosures along the Environmental, Social, and Governance (ESG) pillars demands language models that are at once lightweight and transparent. We deliver the first systematic comparison of two compact transformer families—knowledge-distilled models (Tiny Bidirectional Encoder Representations from Transformers and Distilled BERT) and efficient-attention models (Fourier Network and Longformer)-on three balanced datasets of 2000 documents each, separately targeting the Environmental, Social, and Governance dimensions. All models are fine-tuned on a single 8 gigabyte Graphics Processing Unit and judged on predictive accuracy, inference latency, token-level interpretability obtained with SHapley Additive exPlanations, and concise resource indicators such as per-epoch training time and peak memory consumption. Distilled transformers achieve the best classification performance while offering significant speed-ups and memory savings relative to a BERT-base reference, whereas efficient-attention models become appealing only when inputs exceed four thousand tokens. Attribution analysis reveals that distillation focuses explanatory mass on domain-specific terms such as “diversity” and “emissions”, whereas efficient-attention diffuses importance across generic words. These findings position knowledge distillation as the most practical route for productionready ESG text classification under tight compute and audit constraints, and establish baselines to steer future adaptations of efficient-attention architectures to this domain.

Programmable matter in the context of microrobots is like mobile microcomputers with size of a millimeter, which can move around each other, communicate to form different shapes. Current challenges are as much theoretical as technological. The main research challenges in computer science focus on clustering, synchronization time, decision-making, analysis of detected data, security and the autoconfiguration process, which today remains the most fundamental. When the number of modules is very high, the self-reconfiguration challenge becomes more crucial for applications requiring rapid transition between two forms. Clustering can enable parallel transitions by allowing geographically close modules to make intra-cluster transitions, thus reducing reconfiguration complexity, the number of transitions and the time required. Clustering-based autoconfiguration solutions are available, but they do not include the security aspect. In this paper, we propose a security protocol for clustering-based self-reconfiguration solutions. This solution relies on resource-efficient cryptographic mechanisms to implement a robust authentication solution that underpins a flexible key management mechanism coupled with a confidential data exchange of the resulting structure

The Bill of Lading (BoL) is the most important commercial document in the entire in international transport, particularly maritime transport. It has three main functions: it contains proof of the contract of carriage, serves as a receipt for the goods, as well as the title to the goods, but is generally issued in the form of a physical paper document signed and stamped by the carrier or from PDF files. However, both methods have serious disadvantages such as the risk of fraud, high costs, production, administration and storage times, as well as significant environmental impacts. Our goal in this article is to make the BoL electronically smart using blockchain technology by replacing the traditional paper document by ensuring full traceability of all logistical operations and events. Our Smart BoL proposal using a private cloud blockchain can only be transferred between connected and registered companies that become users (shipper, carrier, consignee, as well as other stakeholders). The connection can only be accepted or initiated through the published address book of these users. An intelligent BoL will serve as an immutable audit trail, absolute transparency of business events and unbreakable confidentiality between business partners.

With the proliferation of IoT devices and the exponential growth of data generated at the edge, federated learning (FL) emerges as a powerful method for training machine learning models on decentralized data sources. In security-critical applications, such as anomaly and threat detection, ensuring the confidentiality and integrity of sensitive data is paramount. In this paper, we introduce VeSAFL, a novel scheme for verifiable secure aggregation for privacy-preserving FL designed for edge computing environments. VeSAFL decentralizes model updates across edge nodes, reducing dependency on centralized cloud servers and mitigating risks associated with single points of failure. By leveraging multi-server aggregators, our approach fortifies system resilience and reliability against potential cyber threats. To bolster trust in the learning process, we implement a robust verification mechanism that guarantees the integrity and authenticity of local and global updates. Our experimental results highlight the efficacy and efficiency of VeSAFL in safeguarding against active adversaries and accurately identifying anomalous activities. Furthermore, our comprehensive security analysis affirms the scheme's correctness, verifiability, and privacy preservation in adversarial scenarios.

Modern containerized applications deployed on Kubernetes demand efficient resource scaling to adapt to varying workloads. Auto-scaling reduces cloud infrastructure costs, increases application stability and improves the Quality of Service from user perspective. This paper introduces an innovative approach for Kubernetes vertical pod autoscaling. The proposed model DTR-Max takes in consideration maximum resources limit of pods allocation using predictive Decision Tree Regression (DTR) policy. The system leverages predictive resource management, enabling proactive adjustments to pod resource requests based on historical utilization patterns and pods resources limit range. In experimental simulations, the system effectively manages resource scaling decisions and demonstrates its potential to adapt Kubernetes pods to dynamic workloads efficiently.

In recent years, there has been substantial attention directed towards the development of proficient Transformers, which display considerable potential in effectively managing extensive sequences in various fields, including but not limited to natural language processing and computer vision. While researchers have conducted survey papers on this field, there remains a notable gap in papers that specifically present and discuss benchmark datasets tailored to evaluate the efficiency and performance of efficient transformers in specific challenges, notably hierarchical reasoning on long-range sequences, question answering on large documents, and key phrase extraction. Moreover, our paper aims to fill this void by offering a clear and comprehensive survey of efficient Transformers. We focus on three main approaches: sparse attention, linearized attention, and low rank attention, all geared towards mitigating the computational and memory challenges associated with traditional Transformers while ensuring consistently high-performance levels. By presenting and analyzing these critical aspects, our paper endeavors to contribute to the broader understanding and advancements in efficient Transformers, paving the way for further research and innovation in this dynamic field.

As the quantity of cars increases on the road, so does the focus on improving road safety and in-vehicle entertainment. In response to this need, the creation of new in-car applications and services is on the rise. Common applications include collision avoidance, safety and traffic monitoring, multimedia streaming, and smart city data collection using wireless sensor networks. The Internet of Things (IOT) is a recent innovation that is accelerating the universe towards automated procedures and intelligent transportation management solutions. This is an important contribution to automation and intelligent civilization. Another way to improve the transport system is through the machine learning (ML). an Existing traffic management solution face a number of problems, resulting in traffic congestion, delays and high fatality rates. In addition, vehicular ad hoc network (VANET) has emerged as a new technology capable of supporting such emerging vehicular applications. This paper presents a comprehensive review of the latest state of the traffic control, traffic management and parking and smart city strategy of the vehicle by using VANET, ML and Internet of Things technology.

Today Industry 4.0 offers innovative technologies for the growth of industrial companies. Technologies such as Internet of Things (IoT), Artificial Intelligence (AI), Blockchain, Big Data, Machine Learning (ML) and Deep learning all are used to implement Industry 4.0. Predictive maintenance as a service (PdMaaS) consists of anticipating future failures on equipment, an object in aaS mode. Concretely, it is a question of anticipating a breakdown or a malfunction thanks to the accumulation of a set of data collected in the IoT cloud. Blockchain technology has the ability to revolutionize predictive maintenance (PdM) processes in various industries including automotive, oil, gas, and aerospace, heavy-machinery industry, etc. Predictive maintenance involves identifying potential equipment failures before they occur. Our main objective in this article is to propose a new predictive maintenance architecture as a service to predict equipment failures based on certain parameters and factors using Blockchain technology between communication equipment (Connected Objects) and also trace all communications between Machine-to-machine (M2M) and storage in Big Data. Once predicted, manufacturers take the necessary steps to prevent this failure through smart preventative maintenance as a service.

The increasing prevalence of Visually-rich Documents (VRDs) in diverse domains has led to a growing interest in Visually-rich Document Understanding (VrDU). Researchers have been developing sophisticated systems to extract explicit and implicit information from such documents. However, a comprehensive literature review on this field still missing. In this paper, we present a comprehensive overview of VrDU systems, focusing on their applications in Key Information Extraction (KIE), Visual Question Answering (VQA) and document classification tasks. We delve into the fundamental concepts and techniques used in VrDU, providing a high-level abstraction of these approaches. To facilitate understanding and comparison, we propose a taxonomy that categorizes VrDU systems based on four key aspects: modality, geometric approach, OCR integration type and task. Additionally, we present the benchmarks commonly used in the field and discuss the current challenges faced by VrDU systems. We believe that this paper advances the research in the field of VrDU by providing a comprehensive resource for researchers and practitioners

Traceability is the ability to verify, follow, manage, and control medical data (or records) that are created during initial consultations in public and private healthcare and hospital centers. As far as Morocco is concerned, these medical records are still created and captured using manual processes and/or traditional electronic systems. These systems are not standardized and can differ among physicians in private centers and even among healthcare institutions. With the advancement of technology, it has become clear that the current trend regarding healthcare is towards common (remotely controlled and) regularly updated digital medical profiles. As a matter of fact, all medical data can be digitized and structured, by developing individual and unique IDs provided for each patient. In this paper, we highlight the importance and benefits of medical traceability and discuss how it can improve and resolve some of the main concerns related to privacy, technology, and systems in Moroccan healthcare. We also introduce a lasting traceability-based approach using RFID to facilitate lifelong health management for Moroccan patients, using a standard and unique ID since birth.

Recently, the mobile ad hoc network (MANET) has enjoyed a great reputation thanks to its advantages such as: high performance, no expensive infrastructure to install, use of unlicensed frequency spectrum, and fast distribution of information around the transmitter. But the topology of MANETs attracts the attention of several attacks. Although authentication and encryption techniques can provide some protection, especially by minimizing the number of intrusions, such cryptographic techniques do not work effectively in the case of unseen or unknown attacks. In this case, the machine learning approach is successful to detect unfamiliar intrusive behavior. Security methodologies in MANETs mainly focus on eliminating malicious attacks, misbehaving nodes, and providing secure routing. In this paper we present to most recent works that propose or apply the concept of Machine Learning (ML) to secure the MANET environment.

The traditional cloud Service-Level Agreement (C-SLA) ensures that cloud providers meet certain business-level requirements and provide customers with a set of clearly defined services as clauses in a simple IT contract. These agreed terms create a foundation of trust on which a customer agrees to use the services of a cloud provider. They also reflect the provider's commitments to Quality of Service (QoS) and cloud infrastructure. Suppliers must also have KPIs that they can measure the reality is essential to have a dashboard and monitoring for the governance of these agreements such contracts require dedicated administration teams with financial, operational and time-consuming compliance. Our objective in this paper is to propose a new service-oriented smart contract model of smart cloud SALs to cloud providers is a microservices approach that also allows to create several independent services that then interact in a secure and very useful way. During the execution of these microservices, all the validation steps are recorded in the blockchain, which makes it possible to secure and control the effectiveness of their actions using predefined performance indicators.

K-nearest neighbors is one of the most popular classification algorithms. It assumes that similar things or people are near to each other. One of the primordial steps in this algorithm is K value, which is given by the user. This value influences the algorithm result performance. In this paper, we suggest an enhanced approach that eliminates the use of K value with keeping the same performance and increasing it for a specific datasets type. We propose a combined approach named `Zone classifier' that provides an excellent performance which is estimated, in average, by more than 85%; for Iris, Wine, Digits, The breast cancer and Olivetti faces dataset.

Mobile ad hoc network (MANET) is a temporary network used in any application where the deployment of a wired network infrastructure is too restrictive. MANET is consists of wireless mobile nodes without any infrastructure and it has no centralized administration and no centralized equipment like access points. Each node in MANET need specifics routing protocols to adapt mobile environment that characterized by changing network topology dynamically and should be efficient in energy and bandwidth. These routing protocols are affected by the threat of malicious nodes attacks on MANET. In this paper, we analyzed two important routing protocols under Black Hole attack, On Demand Distance Vector (AODV) as a reactive protocol and optimized Link State Routing protocols (OLSR) as a proactive protocol. To evaluate the impact of the Black hole attack on routing protocol performance we used four metrics to compare, on the basis of Throughput, Packet Delivery Ratio (PDR), End-to-End Delay (EED) and Packet Lost to validate which protocol is more vulnerable to Black Hole attack all measurements were carried on Simulating in Network Simulation Tool (NS3).

Target tracking with the wireless sensors networks is to detect and locate a target on its entire path through a region of interest. This application arouses interest in the world of research for its many fields of use. Wireless sensor networks, thanks to their versatility, can be used in many hostile environments and inaccessible to humans. However, with a limited energy, they cannot remain permanently active which can significantly reduce their lifetime. The formation of a cluster network seems an effective mechanism to increase network lifetime ". We propose to build optimal dynamic clusters on the target trajectory. For increasing energy efficiency, our algorithm integrates for the first time, to our knowledge, strategies to avoid overlapping clusters and a model to wake up the sensors, adapting to the context of targets with large and variable speed.

In the healthcare context, decisions are crucial. Therefore Decision Support Systems must ensure a high level of dependability. In order to justify the required dependability level, its requirements must be addressed seriously during the system life cycle. To overcome this issue, our approach is based on Model Driven Architecture to model dependability requirements from the earlier stages of decision support system development. Then, it proposes to map them to functional one in order to obtain integral models. The purpose of this paper is to prove the applicability and the benefits of our approach using a case study of urgent clinical care dashboard which is a project recommended by the National Health Service (NHS) of United Kingdom (UK).

In today's distributed organizations, there is a huge need of establishing a collaborative environment where multiple users can cooperate to achieve common goals. This paper proposes a lightweight approach for automatic and transparent collaboration workflow environment, supporting generic collaboration scenario. The interesting aspect of our work is the fact that collaboration designer needs just to specify the workflow description according to a simple and structured model. The collaboration is built from a high-level specification form in XML schema, making it very easy for non-professionals to establish collaboration. As a proof of concept, we apply our approach to medical collaborative diagnosis.

Today, One of the most important performance measures in the new burst traffic gets to 3G and 4G systems is the End-to-End jitter. However, in order to study and analyze Jitter, many works and models have been devoted to estimate and evaluate this parameter. In this paper, analytical formulas for jitter of interrupted Poisson traffic (IPP) and MMPP-2 are presented. We provide Jitter approximation for The IPP Model. Moreover, we found that the jitter calculated directly using the IPP process is almost equal to that found by the EPP model. We have found that jitter can be a decreasing function of load. Simulation results show that the delay and jitter behave inversely which confirms the result already found in [8]. Finally, we found value of the average arrival rate and service rate required to provide flexible and efficient QoS guarantees for VoIP.