Information, Volume 14, Issue 12 (December 2023) – 42 articles

The increasing complexity and connectivity of automotive systems have raised concerns about their vulnerability to security breaches. As a result, the integration of formal methods and validation techniques has become crucial in ensuring the security of automotive systems. This survey research paper aims to provide a comprehensive overview of the current state-of-the-art formal methods and validation techniques employed in the automotive industry for system security. The paper begins by discussing the challenges associated with automotive system security and the potential consequences of security breaches. Then, it explores various formal methods, such as model checking, theorem proving, and abstract interpretation, which have been widely used to analyze and verify the security properties of automotive systems. Additionally, the survey highlights the validation techniques employed to ensure the effectiveness of security measures, including penetration testing, fault injection, and fuzz testing. Furthermore, the paper examines the integration of formal methods and validation techniques within the automotive development lifecycle, including requirements engineering, design, implementation, and testing phases. It discusses the benefits and limitations of these approaches, considering factors such as scalability, efficiency, and applicability to real-world automotive systems. Through an extensive review of relevant literature and case studies, this survey provides insights into the current research trends, challenges, and open research questions in the field of formal methods and validation techniques for automotive system security. The findings of this survey can serve as a valuable resource for researchers, practitioners, and policymakers involved in the design, development, and evaluation of secure automotive systems. Full article

The evolvement of COVID-19 vaccines is rapidly being revolutionized using artificial intelligence-based technologies. Small compounds, peptides, and epitopes are collected to develop new therapeutics. These substances can also guide artificial intelligence-based modeling, screening, or creation. Machine learning techniques are used to leverage pre-existing data for COVID-19 drug detection and vaccine advancement, while artificial intelligence-based models are used for these purposes. Models based on artificial intelligence are used to evaluate and recognize the best candidate targets for future therapeutic development. Artificial intelligence-based strategies can be used to address issues with the safety and efficacy of COVID-19 vaccine candidates, as well as issues with manufacturing, storage, and logistics. Because antigenic peptides are effective at eliciting immune responses, artificial intelligence algorithms can assist in identifying the most promising COVID-19 vaccine candidates. Following COVID-19 vaccination, the first phase of the vaccine-induced immune response occurs when major histocompatibility complex (MHC) class II molecules (typically bind peptides of 12–25 amino acids) recognize antigenic peptides. Therefore, AI-based models are used to identify the best COVID-19 vaccine candidates and ensure the efficacy and safety of vaccine-induced immune responses. This study explores the use of artificial intelligence-based approaches to address logistics, manufacturing, storage, safety, and effectiveness issues associated with several COVID-19 vaccine candidates. Additionally, we will evaluate potential targets for next-generation treatments and examine the role that artificial intelligence-based models can play in identifying the most promising COVID-19 vaccine candidates, while also considering the effectiveness of antigenic peptides in triggering immune responses. The aim of this project is to gain insights into how artificial intelligence-based approaches could revolutionize the development of COVID-19 vaccines and how they can be leveraged to address challenges associated with vaccine development. In this work, we highlight potential barriers and solutions and focus on recent improvements in using artificial intelligence to produce COVID-19 drugs and vaccines, as well as the prospects for intelligent training in COVID-19 treatment discovery. Full article

Artificial intelligence (AI) has rapidly emerged as a transformative force in multiple sectors, with digital marketing being a prominent beneficiary. As AI technologies continue to advance, their potential to reshape the digital marketing landscape becomes ever more apparent, leading to profound implications for businesses and their digital outreach strategies. This research seeks to answer the pivotal question: “How could AI applications be leveraged to optimize digital marketing strategies”? Drawing from a systematic literature review guided by the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) framework, this study has identified 211 pertinent articles. Through a comprehensive bibliometric analysis, the findings were categorized into distinct clusters, namely: AI/ML (Machine Learning) Algorithms, Social Media, Consumer Behavior, E-Commerce, Digital Advertising, Budget Optimization, and Competitive Strategies. Each cluster offers insights into how AI applications can be harnessed to augment digital marketing efforts. The conclusion synthesizes key findings and suggests avenues for future exploration in this dynamic intersection of AI and digital marketing. This research contributes to the field by providing a comprehensive bibliometric analysis of AI in digital marketing, identifying key trends, challenges, and future directions. Our systematic approach offers valuable insights for businesses and researchers alike, enhancing the understanding of AI’s evolving role in digital marketing strategies. Full article

Drug side effects (DSEs) or adverse drug reactions (ADRs) are a major concern in the healthcare industry, accounting for a significant number of annual deaths in Europe alone. Identifying and predicting DSEs early in the drug development process is crucial to mitigate their impact on public health and reduce the time and costs associated with drug development. Objective: In this study, our primary objective is to predict multiple drug side effects using 2D chemical structures, especially for COVID-19, departing from the conventional approach of relying on 1D chemical structures. We aim to develop a novel model for DSE prediction that leverages the CNN-based transfer learning architecture of ResNet152V2. Motivation: The motivation behind this research stems from the need to enhance the efficiency and accuracy of DSE prediction, enabling the pharmaceutical industry to identify potential drug candidates with fewer adverse effects. By utilizing 2D chemical structures and employing data augmentation techniques, we seek to revolutionize the field of drug side-effect prediction. Novelty: This study introduces several novel aspects. The proposed study is the first of its kind to use 2D chemical structures for predicting drug side effects, departing from the conventional 1D approaches. Secondly, we employ data augmentation with both conventional and diffusion-based models (Pix2Pix), a unique strategy in the field. These innovations set the stage for a more advanced and accurate approach to DSE prediction. Results: Our proposed model, named CHEM2SIDE, achieved an impressive average training accuracy of 0.78. Moreover, the average validation and test accuracy, precision, and recall were all at 0.73. When evaluated for COVID-19 drugs, our model exhibited an accuracy of 0.72, a precision of 0.79, a recall of 0.72, and an F1 score of 0.73. Comparative assessments against established transfer learning and machine learning models (VGG16, MobileNetV2, DenseNet121, and KNN) showcased the exceptional performance of CHEM2SIDE, marking a significant advancement in drug side-effect prediction. Conclusions: Our study introduces a groundbreaking approach to predicting drug side effects by using 2D chemical structures and incorporating data augmentation. The CHEM2SIDE model demonstrates remarkable accuracy and outperforms existing models, offering a promising solution to the challenges posed by DSEs in drug development. This research holds great potential for improving drug safety and reducing the associated time and costs. Full article

Cloud computing and relevant emerging technologies have presented ordinary methods for processing edge-produced data in a centralized manner. Presently, there is a tendency to offload processing tasks as close to the edge as possible to reduce the costs and network bandwidth used. In this direction, we find efforts that materialize this paradigm by introducing distributed deep learning methods and the so-called Federated Learning (FL). Such distributed architectures are valuable assets in terms of efficiently managing resources and eliciting predictions that can be used for proactive adaptation of distributed applications. In this work, we focus on deep learning local loss functions in multi-cloud environments. We introduce the MulticloudFL system that enhances the forecasting accuracy, in dynamic settings, by applying two new methods that enhance the prediction accuracy in applications and resources monitoring metrics. The proposed algorithm’s performance is evaluated via various experiments that confirm the quality and benefits of the MulticloudFL system, as it improves the prediction accuracy on time-series data while reducing the bandwidth requirements and privacy risks during the training process. Full article

Within organizational settings, communication dynamics are influenced by various factors, such as email content, historical interactions, and interpersonal relationships. We introduce the Email MultiModal Architecture (EMMA) to model these dynamics and predict future communication behavior. EMMA uses data related to an email sender’s social network, performance metrics, and peer endorsements to predict the probability of receiving an email response. Our primary analysis is based on a dataset of 0.6 million corporate emails from 4320 employees between 2012 and 2014. By integrating features that capture a sender’s organizational influence and likability within a multimodal structure, EMMA offers improved performance over models that rely solely on linguistic attributes. Our findings indicate that EMMA enhances email reply prediction accuracy by up to 12.5% compared to leading text-centric models. EMMA also demonstrates high accuracy on other email datasets, reinforcing its utility and generalizability in diverse contexts. Our findings recommend the need for multimodal approaches to better model communication patterns within organizations and teams and to better understand how relationships and histories shape communication trajectories. Full article

In the face of increasing flood risks intensified by climate change, accurate flood inundation mapping is pivotal for effective disaster management. This study introduces a novel explainable deep learning architecture designed to generate precise flood inundation maps from diverse satellite data sources. A comprehensive evaluation of the proposed model is conducted, comparing it with state-of-the-art models across various fusion configurations of Multispectral Optical and Synthetic Aperture Radar (SAR) images. The proposed model consistently outperforms other models across both Sentinel-1 and Sentinel-2 images, achieving an Intersection Over Union (IOU) of 0.5862 and 0.7031, respectively. Furthermore, analysis of the different fusion combinations reveals that the use of Sentinel-1 in combination with RGB, NIR, and SWIR achieves the highest IOU of 0.7053 and that the inclusion of the SWIR band has the greatest positive impact on the results. Gradient-weighted class activation mapping is employed to provide insights into its decision-making processes, enhancing transparency and interpretability. This research contributes significantly to the field of flood inundation mapping, offering an efficient model suitable for diverse applications. This study not only advances flood inundation mapping but also provides a valuable tool for improved understanding of deep learning decision-making in this area, ultimately contributing to improved disaster management strategies. Full article

The global impact of the COVID-19 pandemic has been profound, placing significant challenges upon healthcare systems and the world economy. The pervasive presence of illness, uncertainty, and fear has markedly diminished overall life satisfaction. Consequently, sentiment analysis has gained substantial traction among scholars seeking to unravel the emotional and attitudinal dimensions of this crisis. This research endeavors to provide a bibliometric perspective, shedding light on the principal contributors to this emerging field. It seeks to spotlight the academic institutions associated with this research domain, along with identifying the most influential publications in terms of both paper volume and h-index metrics. To this end, we have meticulously curated a dataset comprising 646 papers sourced from the ISI Web of Science database, all centering on the theme of sentiment analysis during the COVID-19 pandemic. Our findings underscore a burgeoning interest exhibited by the academic community in this particular domain, evident in an astonishing annual growth rate of 153.49%. Furthermore, our analysis elucidates key keywords and collaborative networks within the authorship, offering valuable insights into the global proliferation of this thematic pursuit. In addition to this, our analysis encompasses an n-gram investigation across keywords, abstracts, titles, and keyword plus, complemented by an examination of the most frequently cited works. The results gleaned from these endeavors offer crucial perspectives, contribute to the identification of pertinent issues, and provide guidance for informed decision-making. Full article

Apache Spark is a high-speed computing engine for processing massive data. With its widespread adoption, there is a growing need to analyze its correctness and temporal properties. However, there is scarce research focused on the verification of temporal properties in Spark programs. To address this gap, we employ the code-level runtime verification tool UMC4M based on the Modeling, Simulation, and Verification Language (MSVL). To this end, a Spark program S has to be translated into an MSVL program M, and the negation of the property P specified by a Propositional Projection Temporal Logic (PPTL) formula that needs to be verified is also translated to an MSVL program M1; then, a new MSVL program “M and M1” can be compiled and executed. Whether program S violates the property P is determined by the existence of an acceptable execution of “M and M1”. Thus, the key issue lies in how to formalize model Spark programs using MSVL programs. We previously proposed a solution to this problem—using the MSVL functions to perform Resilient Distributed Datasets (RDD) operations and converting the Spark program into an MSVL program based on the Directed Acyclic Graph (DAG) of the Spark program. However, we only proposed this idea. Building upon this foundation, we implement the conversion from RDD operations to MSVL functions and propose, as well as implement, the rules for translating Spark programs to MSVL programs based on DAG. We confirm the feasibility of this approach and provide a viable method for verifying the temporal properties of Spark programs. Additionally, an automatic translation tool, S2M, is developed. Finally, a case study is presented to demonstrate this conversion process. Full article

To identify objects in images, a complex set of skills is needed that includes understanding the context and being able to determine the borders of objects. In computer vision, this task is known as semantic segmentation and it involves categorizing each pixel in an image. It is crucial in many real-world situations: for autonomous vehicles, it enables the identification of objects in the surrounding area; in medical diagnosis, it enhances the ability to detect dangerous pathologies early, thereby reducing the risk of serious consequences. In this study, we compare the performance of various ensembles of convolutional and transformer neural networks. Ensembles can be created, e.g., by varying the loss function, the data augmentation method, or the learning rate strategy. Our proposed ensemble, which uses a simple averaging rule, demonstrates exceptional performance across multiple datasets. Notably, compared to prior state-of-the-art methods, our ensemble consistently shows improvements in the well-studied polyp segmentation problem. This problem involves the precise delineation and identification of polyps within medical images, and our approach showcases noteworthy advancements in this domain, obtaining an average Dice of 0.887, which outperforms the current SOTA with an average Dice of 0.885. Full article

Reverse engineering (RE) in Integrated Circuits (IC) is a process in which one will attempt to extract the internals of an IC, extract the circuit structure, and determine the gate-level information of an IC. In general, the RE process can be done for validation as well as Intellectual Property (IP) stealing intentions. In addition, RE also facilitates different illicit activities such as the insertion of hardware Trojan, pirating, or counterfeiting a design, or developing an attack. In this work, we propose an approach to introduce cognitive perturbations, with the aid of adversarial machine learning, to the IC layout that could prevent the RE process from succeeding. We first construct a layer-by-layer image dataset of 45 nm predictive technology. With this dataset, we propose a conventional neural network model called RecoG-Net to recognize the logic gates, which is the first step in RE. RecoG-Net is successful in recognizing the gates with more than 99.7% accuracy. Our thwarting approach utilizes the concept of adversarial attack generation algorithms to generate perturbation. Unlike traditional adversarial attacks in machine learning, the perturbation generation needs to be highly constrained to meet the fab rules such as Design Rule Checking (DRC) Layout vs. Schematic (LVS) checks. Hence, we propose CAPTIVE as a constrained perturbation generation satisfying the DRC. The experiments show that the accuracy of reverse engineering using machine learning techniques can decrease from 100% to approximately 30% based on the adversary generator. Full article

In this work, we propose a CAD (computer-aided diagnosis) system using advanced deep-learning models and computer vision techniques that can improve diagnostic accuracy and reduce transmission risks using the YOLOv7 (You Only Look Once, version 7) object detection architecture. The proposed system is capable of accurate object detection, which provides a bounding box denoting the area in the X-rays that shows some possibility of TB (tuberculosis). The system makes use of CNNs (Convolutional Neural Networks) and YOLO models for the detection of the consolidation of cavitary patterns of the lesions and their detection, respectively. For this study, we experimented on the TBX11K dataset, which is a publicly available dataset. In our experiment, we employed class weights and data augmentation techniques to address the data imbalance present in the dataset. This technique shows a promising improvement in the model’s performance and thus better generalization. In addition, it also shows that the developed model achieved promising results with a mAP (mean average precision) of 0.587, addressing class imbalance and yielding a robust performance for both obsolete pulmonary TB and active TB detection. Thus, our CAD system, rooted in state-of-the-art deep-learning and computer vision methodologies, not only advances diagnostic accuracy but also contributes to the mitigation of TB transmission risks. The substantial improvement in the model’s performance and the ability to handle class imbalance underscore the potential of our approach for real-world TB detection applications. Full article

Offline handwritten text recognition (HTR) is a long-standing research project for a wide range of applications, including assisting visually impaired users, humans and robot interactions, and the automatic entry of business documents. However, due to variations in writing styles, visual similarities between different characters, overlap between characters, and source document noise, designing an accurate and flexible HTR system is challenging. The problem becomes serious when the algorithm has a low learning capacity and when the text used is complex and has a lot of characters in the writing system, such as Ethiopic script. In this paper, we propose a new model that recognizes offline handwritten Ethiopic text using a gated convolution and stacked self-attention encoder–decoder network. The proposed model has a feature extraction layer, an encoder layer, and a decoder layer. The feature extraction layer extracts high-dimensional invariant feature maps from the input handwritten image. Using the extracted feature maps, the encoder and decoder layers transcribe the corresponding text. For the training and testing of the proposed model, we prepare an offline handwritten Ethiopic text-line dataset (HETD) with 2800 samples and a handwritten Ethiopic word dataset (HEWD) with 10,540 samples obtained from 250 volunteers. The experiment results of the proposed model on HETD show a 9.17 and 13.11 Character Error Rate (CER) and Word Error Rate (WER), respectively. However, the model on HEWD shows an 8.22 and 9.17 CER and WER, respectively. These results and the prepared datasets will be used as a baseline for future research. Full article

The European Installation Bus(EIB) protocol, also known as KNX/EIB, is widely used in building and home automation. An extension of the KNX/EIB protocol, EIBsec, is primarily designed to meet the requirements for data transmission security in distributed building automation systems. However, this protocol has some security issues in the request, key distribution, and identity authentication processes. This paper employs a formal analysis method that combines Colored Petri Net (CPN) theory with the Dolev-Yao attack model to evaluate and enhance the EIBsec protocol. It utilizes the CPN Tools to conduct CPN modeling analysis on the protocol and introduces a security assessment model to carry out intrusion detection and security assessment. Through this analysis, vulnerabilities in the protocol, such as tampering and replay attacks, are identified. To address these security concerns, we introduce hash verification and timestamp judgment methods into the original protocol to enhance its security. Subsequently, based on the improved protocol, we conduct CPN modeling and verify the security of the new scheme. Finally, through a comparison and analysis of the performance and security between the original protocol and the improved scheme, it is found that the improved scheme has higher security. Full article

In a world where the interconnection and interaction between human and artificial agents are continuously increasing, the dynamics of social bonds and dependence networks play a fundamental role. The core of our investigation revolves around the intricate interplay between dependence and trust within a hybrid society, populated by human and artificial agents. By means of a structural theory, this study offers valuable insights into the utilization of dependence networks and their impact on collaborative dynamics and resource management. Most notably, agents that leverage dependence, even at the cost of interacting with low-trustworthiness partners, achieve superior performance in resource-constrained environments. On the other hand, in contexts where the use of dependence is limited, the role of trust is emphasized. These findings underscore the significance of dependence networks and their practical implications in real-world contexts, offering useful practical implications in areas such as robotics, resource management, and collaboration among human and artificial agents. Full article

The inexorable march of technological advancement, particularly within the digital domain, continues to exert a profound influence on global economies, societies, and governance frameworks. This paper delves into the intricate coordination between digital business and digital governance against the backdrop of sustainable development. By introducing an index system to gauge the levels of digital business and governance, this study assesses their coupling coordination using a coupling coordination model. Through this level of coordination, this paper assesses their respective contributions to the sustainable development objectives of EU countries through panel-corrected standard error (PCSE) estimates. The paper’s findings underscore several key conclusions: (1) Notable upswings are evident in the composite indices for digital business and digital governance growth. Among these, the index of digital business has demonstrated the most pronounced surge. Furthermore, digital business has experienced a distinct upward trajectory in recent years. (2) Although observable, the rise of the coupling degree is restrained, with an overall coupling degree that remains relatively low. The coupling progression has transitioned from a stage of low-degree coupling to that of primary coupling, with EU countries demonstrating fluctuating rising trends in their coupling degrees, marked by conspicuous regional disparities. (3) Over the examined period, the extent of coordination between digital business and digital governance substantially impacts the Sustainable Development Goals (SDG) index. Focusing on the interplay and harmonization between digital business and governance offers a novel pathway toward attaining the objectives of the Sustainable Development Goals. Full article

Trade embargoes, often imposed for political, economic, or security reasons, have long been a tool of international diplomacy. Transshipments may be employed as a strategic mechanism by nations and organizations to circumvent trade embargoes. Transshipment involves rerouting goods through intermediary ports or countries to obscure their origin, destination, or the parties involved. This practice may be subject to investigation, which could lead to exposing the entities employing it. Strategic management of transshipments has to be devised by those entities (the attackers) battling against the transshipment detection mechanisms adopted by the embargo-setters (the defenders). In this paper, we consider an entity exploiting transshipments through several intermediaries. We derive an optimal strategy for that entity wishing to minimize the probability of being exposed. Our strategy provides the optimal number of intermediaries and the optimal distribution of goods among those intermediaries. Full article

With digital patient data increasing due to new diagnostic methods and technology, showing the right data in the context of decision support at the point of care becomes an even greater challenge. Standard operating procedures (SOPs) modeled in BPMN (Business Process Model and Notation) contain evidence-based treatment guidance for all phases of a certain diagnosis, while physicians need the parts relevant to a specific patient at a specific point in the clinical process. Therefore, integration of patient data from electronic health records (EHRs) providing context to clinicians is needed, which is stored and communicated in HL7 (Health Level Seven) FHIR (Fast Healthcare Interoperability Resources). To address this issue, we propose a method combining an integration of stored data into BPMN and a loss-free transformation from BPMN into FHIR, and vice versa. Based on that method, an identification of the next necessary decision point in a specific patient context is possible. We verified the method for treatment of malignant melanoma by using an extract of a formalized SOP document with predefined decision points and validated FHIR references with real EHR data. The patient data could be stored and integrated into the BPMN element ‘DataStoreReference’. Our loss-free transformation process therefore is the foundation for combining evidence-based knowledge from formalized clinical guidelines or SOPs and patient data from EHRs stored in FHIR. Processing the SOP with the available patient data can then lead to the next upcoming decision point, which will be displayed to the physician integrated with the corresponding data. Full article

Deep learning is used in various applications due to its advantages over traditional Machine Learning (ML) approaches in tasks encompassing complex pattern learning, automatic feature extraction, scalability, adaptability, and performance in general. This paper proposes an end-to-end (E2E) delay estimation method for 5G networks through deep learning (DL) techniques based on Gaussian Mixture Models (GMM). In the first step, the components of a GMM are estimated through the Expectation-Maximization (EM) algorithm and are subsequently used as labeled data in a supervised deep learning stage. A multi-layer neural network model is trained using the labeled data and assuming different numbers of E2E delay observations for each training sample. The accuracy and computation time of the proposed deep learning estimator based on the Gaussian Mixture Model (DLEGMM) are evaluated for different 5G network scenarios. The simulation results show that the DLEGMM outperforms the GMM method based on the EM algorithm, in terms of the accuracy of the E2E delay estimates, although requiring a higher computation time. The estimation method is characterized for different 5G scenarios, and when compared to GMM, DLEGMM reduces the mean squared error (MSE) obtained with GMM between 1.7 to 2.6 times. Full article

Anomaly detection plays a crucial role in data security and risk management across various domains, such as financial insurance security, medical image recognition, and Internet of Things (IoT) device management. Researchers rely on machine learning to address potential threats in order to enhance data security. In the financial insurance industry, enterprises tend to leverage the relation mining capabilities of knowledge graph embedding (KGE) for anomaly detection. However, auto insurance fraud labeling strongly relies on manual labeling by experts. The efficiency and cost issues of labeling make auto insurance fraud detection still a small-sample detection challenge. Existing schemes, such as migration learning and data augmentation methods, are susceptible to local characteristics, leading to their poor generalization performance. To improve its generalization, the recently emerging Decentralized Federated Learning (DFL) framework provides new ideas for mining more frauds through the joint cooperation of companies. Based on DFL, we propose a federated framework named DFLR for relation embedding aggregation. This framework trains the private KGE of auto insurance companies on the client locally and dynamically selects servers for relation aggregation with the aim of privacy protection. Finally, we validate the effectiveness of our proposed DFLR on a real auto insurance dataset. And the results show that the cooperative approach provided by DFLR improves the client’s ability to detect auto insurance fraud compared to single client training. Full article

Alzheimer’s disease (AD) is an illness affecting the neurological system in people commonly aged 65 years and older. It is one of the leading causes of dementia and, subsequently, the cause of death as it gradually affects and destroys brain cells. In recent years, the detection of AD has been examined in ways to mitigate its impacts while considering early detection through computer-aided diagnosis (CAD) tools. In this study, we developed deep learning models that focus on early detection and classifying each case, non-demented, moderate-demented, mild-demented, and very-mild-demented, accordingly through transfer learning (TL); an AlexNet, ResNet-50, GoogleNet (InceptionV3), and SqueezeNet by utilizing magnetic resonance images (MRI) and the use of image augmentation. The acquired images, a total of 12,800 images and four classifications, had to go through a pre-processing phase to be balanced and fit the criteria of each model. Each of these proposed models split the data into 80% training and 20% testing. AlexNet performed an average accuracy of 98.05%, GoogleNet (InceptionV3) performed an average accuracy of 97.80%, and ResNet-50 had an average performing accuracy of 91.11%. The transfer learning approach assists when there is not adequate data to train a network from the start, which aids in tackling one of the major challenges faced when working with deep learning. Full article

This paper proposes the concept of risk-aware actual value as a pivotal metric for evaluating the viability and desirability of AI projects and services in accordance with the AI Act. The framework establishes a direct correlation between the level of risk associated with a product or service and the resulting actual value generated. The AI Act reflects a concerted effort to harness the potential of AI while mitigating risks. The risk-based approach aligns regulatory measures with the specific attributes and potential hazards of distinct AI applications. As trilogue negotiations continue, the regulatory approach of the EU is evolving, highlighting its commitment to responsible and forward-thinking AI governance. Through a dedicated analysis of the AI Act, it becomes evident that products or services categorized as high-risk carry substantial compliance obligations, consequently diminishing their potential value. This underscores the imperative of exercising caution when engaging in projects with elevated risk profiles. Conversely, products or services characterized by lower risk levels are poised to accrue more substantial benefits from their AI and data potential, highlighting the incentive for a discerning approach to risk assessment. Methodologically, we propose an extension of an integrated AI risk management framework that is already existing in the literature, combining it with existing frameworks for measuring value creation from harnessing AI potential. Additionally, we contribute to the applied field of AI by implementing the proposed risk framework across nine industry-relevant use cases. In summation, this paper furnishes a comprehensive approach to achieving equilibrium between innovation and regulation in the realm of AI projects and services. By employing the risk-aware actual value metric, stakeholders are empowered to make informed decisions that prioritize safety and maximize the potential benefits of AI initiatives. This framework may stand as a reference point in this time when fostering responsible and sustainable AI development within the industry becomes of paramount importance. Full article

Eye-tracking systems have emerged as valuable tools in various research fields, including psychology, medicine, marketing, car safety, and advertising. However, the high costs of the necessary specialized hardware prevent the widespread adoption of these systems. Appearance-based gaze estimation techniques offer a cost-effective alternative that can rely solely on RGB cameras, albeit with reduced accuracy. Therefore, the aim of our work was to present a real-time eye-tracking system with low-end hardware that leverages appearance-based techniques while overcoming their drawbacks to make reliable gaze data accessible to more users. Our system employs fast and light machine learning algorithms from an external library called MediaPipe to identify 3D facial landmarks. Additionally, it uses a series of widely recognized computer vision techniques, like morphological transformations, to effectively track eye movements. The precision and accuracy of the developed system in recognizing saccades and fixations when the eye movements are mainly horizontal were tested through a quantitative comparison with the EyeLink 1000 Plus, a professional eye tracker. Based on the encouraging registered results, we think that it is possible to adopt the presented system as a tool to quickly retrieve reliable gaze information. Full article

In recent years, there has been rapid development in computer technology, leading to an increasing number of medical systems utilizing electronic medical records (EMRs) to store their clinical data. Because EMRs are very private, healthcare institutions usually encrypt these data before transferring them to cloud servers. A technique known as searchable encryption (SE) can be used by healthcare institutions to encrypt EMR data. This technique enables searching within the encrypted data without the need for decryption. However, most existing SE schemes only support keyword or range searches, which are highly inadequate for EMR data as they contain both textual and digital content. To address this issue, we have developed a novel searchable symmetric encryption scheme called SSE-RK, which is specifically designed to support both range and keyword searches, and it is easily applicable to EMR data. We accomplish this by creating a conversion technique that turns keywords and ranges into vectors. These vectors are then used to construct index tree building and search algorithms that enable simultaneous range and keyword searches. We encrypt the index tree using a secure K-Nearest Neighbor technique, which results in an effective SSE-RK approach with a search complexity that is quicker than a linear approach. Theoretical and experimental study further demonstrates that our proposed scheme surpasses previous similar schemes in terms of efficiency. Formal security analysis demonstrates that SSE-RK protects privacy for both data and queries during the search process. Consequently, it holds significant potential for a wide range of applications in EMR data. Overall, our SSE-RK scheme, which offers improved functionality and efficiency while protecting the privacy of EMR data, generally solves the shortcomings of the current SE schemes. Full article

Brain tumors (BT) present a considerable global health concern because of their high mortality rates across diverse age groups. A delay in diagnosing BT can lead to death. Therefore, a timely and accurate diagnosis through magnetic resonance imaging (MRI) is crucial. A radiologist makes the final decision to identify the tumor through MRI. However, manual assessments are flawed, time-consuming, and rely on experienced radiologists or neurologists to identify and diagnose a BT. Computer-aided classification models often lack performance and explainability for clinical translation, particularly in neuroscience research, resulting in physicians perceiving the model results as inadequate due to the black box model. Explainable deep learning (XDL) can advance neuroscientific research and healthcare tasks. To enhance the explainability of deep learning (DL) and provide diagnostic support, we propose a new classification and localization model, combining existing methods to enhance the explainability of DL and provide diagnostic support. We adopt a pre-trained visual geometry group (pre-trained-VGG-19), scratch-VGG-19, and EfficientNet model that runs a modified form of the class activation mapping (CAM), gradient-weighted class activation mapping (Grad-CAM) and Grad-CAM++ algorithms. These algorithms, introduced into a convolutional neural network (CNN), uncover a crucial part of the classification and can provide an explanatory interface for diagnosing BT. The experimental results demonstrate that the pre-trained-VGG-19 with Grad-CAM provides better classification and visualization results than the scratch-VGG-19, EfficientNet, and cutting-edge DL techniques regarding visual and quantitative evaluations with increased accuracy. The proposed approach may contribute to reducing the diagnostic uncertainty and validating BT classification. Full article

Usable security involves designing security measures that accommodate users’ needs and behaviors. Balancing usability and security poses challenges: the more secure the systems, the less usable they will be. On the contrary, more usable systems will be less secure. Numerous studies have addressed this balance. These studies, spanning psychology and computer science/engineering, contribute diverse perspectives, necessitating a systematic review to understand strategies and findings in this area. This systematic literature review examined articles on usable security from 2005 to 2022. A total of 55 research studies were selected after evaluation. The studies have been broadly categorized into four main clusters, each addressing different aspects: (1) usability of authentication methods, (2) helping security developers improve usability, (3) design strategies for influencing user security behavior, and (4) formal models for usable security evaluation. Based on this review, we report that the field’s current state reveals a certain immaturity, with studies tending toward system comparisons rather than establishing robust design guidelines based on a thorough analysis of user behavior. A common theoretical and methodological background is one of the main areas for improvement in this area of research. Moreover, the absence of requirements for Usable security in almost all development contexts greatly discourages implementing good practices since the earlier stages of development. Full article

Emerging digital technologies, such as telemedicine, artificial intelligence, the Internet of Medical Things, blockchain, and visual and augmented reality, have revolutionized the delivery of and access to healthcare services. Such technologies allow for real-time health monitoring, disease diagnosis, chronic disease management, outbreak surveillance, and rehabilitation. They help personalize treatment plans, identify trends, contribute to drug development, and enhance public health management. While emerging digital technologies have numerous benefits, they may also introduce new risks and vulnerabilities that can compromise the confidentiality, integrity, and availability of sensitive healthcare information. This review article discussed, in brief, the key emerging digital technologies in the health sector and the unique threats introduced by these technologies. We also highlighted the risks relevant to digital health cybersecurity, such as data breaches, medical device vulnerabilities, phishing, insider and third-party risks, and ransomware attacks. We suggest that the cybersecurity framework should include developing a comprehensive cybersecurity strategy, conducting regular risk assessments, implementing strong access control, encrypting data, educating staff, implementing secure network segmentation, backing up data regularly, monitoring and detecting anomalies, establishing an incident response plan, sharing threat intelligence, and auditing third-party vendors. Full article

Safety-critical cyber-physical systems (CPSs), such as high-tech cars having cyber capabilities, are highly interconnected. Automotive manufacturers are concerned about cyber attacks on vehicles that can lead to catastrophic consequences. There is a need for a new risk management approach to address and investigate cybersecurity risks. Risk management in the automotive domain is challenging due to technological improvements and advances every year. The current standard for automotive security is ISO/SAE 21434, which discusses a framework that includes threats, associated risks, and risk treatment options such as risk reduction by applying appropriate defences. This paper presents a residual cybersecurity risk management framework aligned with the framework presented in ISO/SAE 21434. A methodology is proposed to develop an integrated attack tree that considers multiple sub-systems within the CPS. Integrating attack trees in this way will help the analyst to take a broad perspective of system security. Our previous approach utilises a flow graph to calculate the residual risk to a system before and after applying defences. This paper is an extension of our initial work. It defines the steps for applying the proposed framework and using adaptive cruise control (ACC) and adaptive light control (ALC) to illustrate the applicability of our work. This work is evaluated by comparing it with the requirements of the risk management framework discussed in the literature. Currently, our methodology satisfies more than 75% of their requirements. Full article

The advent of Multilingual Language Models (MLLMs) and Large Language Models (LLMs) has spawned innovation in many areas of natural language processing. Despite the exciting potential of this technology, its impact on developing high-quality Machine Translation (MT) outputs for low-resource languages remains relatively under-explored. Furthermore, an open-source application, dedicated to both fine-tuning MLLMs and managing the complete MT workflow for low-resources languages, remains unavailable. We aim to address these imbalances through the development of adaptMLLM, which streamlines all processes involved in the fine-tuning of MLLMs for MT. This open-source application is tailored for developers, translators, and users who are engaged in MT. It is particularly useful for newcomers to the field, as it significantly streamlines the configuration of the development environment. An intuitive interface allows for easy customisation of hyperparameters, and the application offers a range of metrics for model evaluation and the capability to deploy models as a translation service directly within the application. As a multilingual tool, we used adaptMLLM to fine-tune models for two low-resource language pairs: English to Irish (EN$\leftrightarrow$ GA) and English to Marathi (EN$\leftrightarrow$MR). Compared with baselines from the LoResMT2021 Shared Task, the adaptMLLM system demonstrated significant improvements. In the EN$\to$GA direction, an improvement of 5.2 BLEU points was observed and an increase of 40.5 BLEU points was recorded in the GA$\to$EN direction representing relative improvements of 14% and 117%, respectively. Significant improvements in the translation performance of the EN$\leftrightarrow$MR pair were also observed notably in the MR$\to$EN direction with an increase of 21.3 BLEU points which corresponds to a relative improvement of 68%. Finally, a fine-grained human evaluation of the MLLM output on the EN$\to$GA pair was conducted using the Multidimensional Quality Metrics and Scalar Quality Metrics error taxonomies. The application and models are freely available. Full article

Scientometrics is a quantitative and statistical approach that analyzes research on certain themes. It originated from information/library science but has been applied in various disciplines, including information science, library science, natural science, technology, engineering, medical sciences, and social sciences and humanities. Numerous scientometric studies have been carried out, but no study has attempted to investigate the overall research status of scientometrics. The objective of this study was to investigate the research status of scientometrics based on 16,225 publications archived in the Web of Science Core Collection between 1992 and 2020. The results show that there has been a marked increase in publications on scientometric studies over the past decades, with “Information Science Library Science” being the predominant discipline publishing scientometric studies, but scientometrics has been widely adopted in a variety of other disciplines (240 of 254 Web of Science categories). It was found that Web of Science, Vosviewer, and Scientometrics are the most utilized database, software, and journal for scientometric studies, respectively. The most productive author (Lutz Bornmann from the Max Planck Society, Germany), organization (University of Granada, Spain), and country (USA) are also identified. In addition, high-impact scientometric studies and the research landscape are analyzed through citation networks and the co-occurrence of keywords method. Full article