Evidence plays a crucial role in judicial systems, and managing it securely and efficiently ensures justice. This paper introduces Decentralized Trust, a framework that combines blockchain technology, Non-Fungible Tokens (NFTs), and fog computing to address common issues like tampering, delays, and reliance on centralized systems. Traditional methods that depend on cloud computing often face high latency and slow processing, especially in remote areas. This research also builds upon the challenges identified in previous studies, such as tampering vulnerabilities, inefficiencies in evidence processing, and accessibility issues in underserved regions, providing a novel and comprehensive solution through Decentralized Trust. Fog computing handles tasks closer to where data is created, reducing delays and improving response times. Blockchain ensures that evidence records cannot be altered, while NFTs make each piece of evidence unique and tamper-proof. The framework is organized into layers: edge nodes at police stations capture evidence, fog nodes process the data and create NFTs, and cloud storage, supported by the Interplanetary File System (IPFS), provides secure long-term storage. Results demonstrate that the framework achieves average transaction delays of 24.5 seconds on low-performance devices (Node A) and 168.9 seconds on high-performance devices (Node B), with margins of error showing efficient scalability even under significant processing loads. The observed transaction delays are due to differences in system architecture and processing priorities. High-performance devices (Node B) have more complex validation processes, increased security checks, or resource contention, contributing to longer transaction times. By combining these technologies, Decentralized Trust offers a reliable, fast, and secure way to manage judicial evidence, building trust in the framework while addressing the needs of remote and underserved areas.

ABSTRACT Generative artificial intelligence (GAI) has emerged as a transformative force in STEM education, offering new possibilities for personalised instruction, content creation and interactive learning. This review examines the current landscape of GAI applications in science, technology, engineering and mathematics, highlighting key tools such as GPT‐4, DALLE, AlphaCode and CodeGen. The paper synthesises recent research and practices to identify the pedagogical benefits of GAI, including enhanced self‐paced learning, improved access to resources and support for interdisciplinary instruction. It also addresses critical challenges, such as the reliability of generated content, ethical concerns, data privacy and teacher preparedness. These challenges were identified through a synthesis of recent empirical studies, policy reports and expert commentaries in the field of GAI in education, which consistently highlight these issues as major barriers to effective implementation. Based on these findings, the review describes implications for curriculum integration, professional development and institutional policy. Furthermore, the study situates GAI within a broader historical and theoretical context, tracking its evolution from traditional machine learning and deep learning approaches and aligning its educational applications with constructivist, cognitive load and personalised learning theories. By categorising specific use cases across STEM disciplines, such as automated scientific explanations, AI‐generated visualisations, intelligent tutoring systems (ITS), virtual engineering labs and adaptive math assessments, the review illustrates the diverse and practical utility of GAI in classroom and remote learning environments. These cases were selected to represent a cross‐section of the core instructional needs in STEM education: explanation, visualisation, guidance, experimentation and assessment, where GAI offers distinct functional advantages. They were categorised according to the primary instructional role they fulfil, allowing a pedagogically meaningful organisation of GAI capabilities aligned with common learning processes in STEM. The analysis also emphasises the importance of teacher agency, student participation and equitable access in shaping effective GAI adoption. Finally, this review identifies key future research directions, including the need for longitudinal studies on learning outcomes, efforts to improve the transparency and explainability of GAI models in educational contexts, the development of domain‐specific generative tools tailored to STEM subfields, and the exploration of collaborative human–AI learning environments.

Generative AI has emerged as a transformative paradigm for diverse applications, yet the escalating scale of modern models exposes critical computational and memory bottlenecks in classical hardware. This paper investigates the intersection of quantum computing and generative artificial intelligence (QGAI) to address these limitations and scale modern generative models. As models grow to billions of parameters, classical systems face bottlenecks in memory, energy, and training efficiency, while quantum computing offers exponential representational benefits for high-dimensional data. The paper analyzes five core quantum generative architectures-Quantum Circuit Born Machines, Quantum Generative Adversarial Networks, Quantum Boltzmann Machines, Quantum Variational Autoencoders, and Quantum Diffusion models, highlighting their design principles, learning mechanisms, and applications. QGAI models have demonstrated significant promise in domains such as drug discovery, human-machine interaction, IoT security, and financial modelling. Despite these advances, QGAI remains constrained by qubit noise, barren plateaus, and integration challenges. We conclude by identifying ten open research challenges and propose directions for achieving scalable, interpretable, and energy-efficient quantum generative learning.

The integration of large-scale foundation models with physical embodiments has led to significant advancements in robotics termed Generative Physical Artificial Intelligence (GPAI). These agentic AI systems autonomously perceive, reason, and act in complex real-world situations. This survey comprehensively analyzes GPAI systems, focusing on their architectural foundations, current applications, and key limitations. We introduce a taxonomy of five distinct approaches: Robot Foundation Models (RFMs) for cross-platform skill transfer; Vision Language Action Models (VLAs) for end-to-end multi-modal perception and control; Large Behavior Models (LBMs) for human-like movement generation; Diffusion Policy Models (DPMs) for diffusion model-based temporally coherent action generation; and World Foundation Models (WFMs) for physics-compliant simulation and data generation. We examine how these approaches complement each other: WFMs generate training data for VLAs and DPMs, RFMs enable cross-platform deployment of learned policies, while LBMs provide motion priors for natural behavior. Through examples across autonomous vehicles, industrial automation, healthcare robotics, and humanoid systems, we identify significant performance improvements and summarize promising research directions in data-efficient learning, sim-to-real transfer, edge-compatible architectures, and safety frameworks. These insights advance embodied AI for IoT-connected environments where intelligent agents interact with networked sensors, actuators, and edge devices.

Wi-Fi Channel State Information (CSI) has become a widely studied modality for device-free sensing as it captures fine-grained wireless channel variations that can be mapped to human motion and presence while avoiding the explicit visual disclosure typical of vision-based systems. CSI-based pipelines have been explored for human activity and gesture recognition, fall detection, gait analysis, pose-related inference, and indoor localization. Despite strong results in controlled settings, practical deployment remains difficult due to measurement noise, sensitivity to environmental dynamics, multi-user interference, and system-level constraints in data acquisition and real-time processing. This article surveys machine learning methods forWi-Fi CSI sensing and analyzes more than 65 representative models, connecting algorithmic design choices with implementable end-to-end system design. We introduce a hierarchical taxonomy that organizes the literature into classical machine learning approaches, deep learning architectures, and hybrid strategies. Beyond modeling, we describe the full sensing pipeline- from hardware and network interface card (NIC) selection to software tools, antenna configuration, and signal conditioning- highlighting the design trade-offs that affect robustness and reproducibility. We further compare methods across major application domains and summarize open challenges in generalization to dynamic environments, multi-user separation, and resource-efficient inference. Finally, we outline research directions toward robust generalization, scalable deployment, and privacy-aware learning to support broader real-world adoption.

Abstract The internet has drastically changed trade and interpersonal interactions since the introduction of the pandemic. Food delivery services have expanded significantly in recent years. These services link customers with restaurants using the internet, enabling quick delivery of mobile meals. But as these businesses grow, dealing with problems like fraud and inefficiencies that may lower the quality of their services is more crucial than ever. This article lists the drawbacks of the existing food delivery services and recommends how to fix them. A blockchain‐based architecture is also used to increase security and reliability, and an auction process encourages fairness and transparency for both riders and restaurants. We have developed an auction model where the riders can quote the price they estimate for a particular delivery. Additionally, a machine‐predicted bid is also used to prevent the case where the riders would try to exploit the restaurants. This method leads to a better way of implementing online food delivery services, safeguarding the interests of customers, riders, and restaurants.

Abstract Online voting is gaining traction in contemporary society to reduce costs and boost voter turnout, allowing individuals to cast their ballots from anywhere with an internet connection. This innovation is cautiously met due to the inherent security risks, where a single vulnerability can lead to widespread vote manipulation. Blockchain technology has emerged as a promising solution to address these concerns and create a trustworthy electoral process. Blockchain offers a decentralized network of nodes that enhances transparency, security, and verifiability. Its distributed ledger and non‐repudiation features make it a compelling alternative to traditional electronic voting systems, ensuring the integrity of elections. To further bolster the security of online voting, we propose DemocracyGuard platform on the Ethereum blockchain, which incorporates facial recognition technology to authenticate voters. By leveraging these advancements, DemocracyGuard aims to provide a secure and resilient platform for online voting, paving the way for its broader adoption and revolutionizing the electoral landscape.

ABSTRACT Generative AI (GAI), which has become increasingly popular nowadays, can be considered a brilliant computational machine that can not only assist with simple searching and organising tasks but also possesses the capability to propose new ideas, make decisions on its own and derive better conclusions from complex inputs. Finance comprises various difficult and time‐consuming tasks that require significant human effort and are highly prone to errors, such as creating and managing financial documents and reports. Hence, incorporating GAI to simplify processes and make them hassle‐free will be consequential. Integrating GAI with finance can open new doors of possibility. With its capacity to enhance decision‐making and provide more effective personalised insights, it has the power to optimise financial procedures. In this paper, we address the research gap of the lack of a detailed study exploring the possibilities and advancements of the integration of GAI with finance. We discuss applications that include providing financial consultations to customers, making predictions about the stock market, identifying and addressing fraudulent activities, evaluating risks, and organising unstructured data. We explore real‐world examples of GAI, including Finance generative pre‐trained transformer (GPT), Bloomberg GPT, and so forth. We look closer at how finance professionals work with AI‐integrated systems and tools and how this affects the overall process. We address the challenges presented by comprehensibility, bias, resource demands, and security issues while at the same time emphasising solutions such as GPTs specialised in financial contexts. To the best of our knowledge, this is the first comprehensive paper dealing with GAI for finance.

ABSTRACT Precious metals are rare metals with high economic value. Forecasting the price volatility of precious metals is essential for investment purposes. In this work, we propose a novel hybrid model of random convolutional kernels‐based neural network model (RCK) and generalized autoregressive conditional heteroscedasticity (GARCH) model for forecasting the metal price volatilities of gold, silver, and platinum. Realized volatility calculated on logarithmic returns is used as an estimate for the volatility of prices, and data standardization is performed before feeding the price volatility data to the RCK model. RCK model applies multiple carefully designed random convolution kernels on the time series input to extract robust features for forecasting. The proportion of positive values (PPV) is extracted as features from the output of convolving convolutional kernels with time‐series inputs, which are then passed through a regressor to forecast volatility. Compared to the existing methods, the proposed method has the advantage that the weights of the random convolutional kernels need not be trained, unlike other neural network models. Further, no other work has made use of random convolutional kernels for precious metal forecasting, to the best of our knowledge. We incorporated novel learning and data augmentation strategies to achieve better performance. In particular, we used the cosine annealing learning rate strategy and Mixup data augmentation technique to improve the proposed model's performance. We have used MSE (Mean Squared Error), RMSE (Root Mean Squared Error), MAE (Mean Absolute Error), and MAPE (Mean Absolute Percentage Error) as metrics to compare the proposed models' performance. The proposed model decreases the MSE by 53% compared to the GARCH‐LSTM model, which is the current state‐of‐the‐art hybrid model for volatility forecasting.

Electric vehicles (EVs) are becoming more popular as people try to live more eco-friendly ways. A major challenge slowing down their widespread adoption is limited driving range and inadequate charging infrastructure, particularly in rural and highway areas where charging station deployment remains uneven. This paper proposed a blockchain-based framework for Vehicle-to-Vehicle (V2V) and Vehicle-to-Grid (V2G) energy trading to address these challenges, enabling efficient decentralized energy exchanges. The framework integrates three-game theory models, contract theory, Bayesian game theory, and evolutionary game theory to optimize trading strategies, ensure fairness, and enhance grid stability. By utilizing a lightweight blockchain architecture on Hyperledger Fabric, the proposed system ensures secure, transparent, and efficient transactions while reducing operational costs. The performance evaluations show that the proposed framework surpasses existing methods, reaching a transaction throughput of 49.8 transactions per second (TPS) for payment settlements and 44.7 TPS for energy trade proposals, with an average latency ranging from 0.09 to 0.18 seconds. Resource utilization analysis shows that peer nodes experience an average CPU usage of 23.76% and memory consumption of 169.5 MB during trade proposals. These results highlight the robustness and scalability of the framework in enabling decentralized, carbon-intelligent energy trading, offering a promising solution for advancing sustainable and intelligent EV energy ecosystems.

Data converters significantly contribute to efficient and accurate data processing in Internet of Things (IoT) systems. As IoT expands into agriculture, industrial automation, and healthcare (AIH), precise and low-power data conversion has become crucial to support longer battery life and reliable performance in IoT devices. Efficient data converters are key to reducing energy use, especially in components like comparator circuits, which consume significant energy in successive approximation register analog-to-digital converters (SAR ADCs). This survey provides an in-depth review of recent developments in low-power data converter design, examining techniques that help reduce power consumption at various stages. It emphasizes advancements, such as energy scaling, dynamic voltage references, and architectural optimizations that enhance efficiency without compromising performance. A specific analysis of emerging technology trends, such as the application of machine learning in data converter design, is explored to stimulate further innovation. Machine learning (ML)-based optimization, including adaptive calibration, noise reduction, and real-time performance optimization, presents new opportunities for enhancing efficiency and accuracy while addressing critical design constraints in IoT applications. While quantum encryption offers promising advancements in securing IoT data transmission, a broader security perspective beyond encryption is necessary, including concerns, such as attack detection and data integrity, ensuring the robustness of IoT systems. This review also examines latency, signal integrity, and accuracy issues, offering a roadmap for next-generation converter designs and reducing power consumption in data converters, which are fundamental to enhancing the performance and lifespan of IoT devices.

Nowadays, real-time object detection, which is a crucial task, is being performed through image processing and deep learning techniques. As there are several high-performance computing edge devices available, selecting the best-fit device for a particular problem is a tough task and keeping in mind the cost, performance, and weight of the device in mind. One faces several challenges while performing this task in real-time such as a lack of resources in terms of power and mobility. We have provided an insight into the computation power of devices in terms of Frames per Second (FPS) by deploying object detection models on them. This paper will provide insight into selecting the appropriate combination of device and object detection models for real-time applications. Raspberry Pi 3 (RPi3), Raspberry Pi 4 (RPi4), Intel Neural Compute Stick 2 (NCS2), and Nvidia Jetson NANO are popular devices with high computation power used for real-time applications. The memory constraints of devices along with the deployment of different You Only Look Once (YOLO) and Single-Shot Detector (SSD) are the two object detection models that have been explained in this paper. A deep learning inference optimiser, TensorRT, has been used in NANO to achieve high throughput in the performance of object detection. The precision, recall, and F1 score achieved on deploying each tested model have been presented. After observing the devices during experimentation, RPi4+NCS2 showed the best execution with the blend of factors i.e. speed, portability, and user-friendliness.

Abstract Face recognition is an essential feature required for a range of computer vision applications such as security, attendance systems, emotion detection, airport check‐in, and many others. The super‐resolution of subject images is an important and challenging element in numerous scenarios. At times the images are low resolution and need to be processed through super‐resolution techniques to gain more accurate results. For the problem of image super‐resolution, deep learning‐based face recognition systems have been explored in recent years; however, low‐resolution face recognition remains an arduous task. Generative adversarial network (GAN) based models are a promising approach to address this challenge. However, conventional GAN‐based models may generate images that differ significantly from an original high‐resolution image in the test set to the point that the identity of the target face may be changed. To address this shortcoming, we propose a novel U‐Net style generator architecture, where skip‐connections between the encoder and decoder layer can help in preserving the facial characteristics of the input image in the generated image, thus curbing the generator's ability to generate an entirely new image and training it to generate an image more similar in characteristics to the original image. In addition to statistical metrics like structural similarity index measure and Fréchet inception distance, we compute the pixel‐wise distance between the original and model‐generated images to ascertain that our model generates as close to the original images as possible. While we train the model for 4× super‐resolution (64 × 64 images to 256 × 256), our architecture can also be trained for an arbitrary resizing scale. Finally, the number of faces detected over high‐resolution images generated by our model is shown to be higher than state‐of‐the‐art high‐resolution image creation models for face recognition tasks.

Skin diseases are reported to contribute 1.79% of the global burden of disease. The accurate diagnosis of specific skin diseases is known to be a challenging task due, in part, to variations in skin tone, texture, body hair, etc. Classification of skin lesions using machine learning is a demanding task, due to the varying shapes, sizes, colors, and vague boundaries of some lesions. The use of deep learning for the classification of skin lesion images has been shown to help diagnose the disease at its early stages. Recent studies have demonstrated that these models perform well in skin detection tasks, with high accuracy and efficiency. Our paper proposes an end-to-end framework for skin lesion classification, and our contributions are two-fold. Firstly, two fundamentally different algorithms are proposed for segmenting and extracting features from images during image preprocessing. Secondly, we present a deep convolutional neural network model, S-MobileNet that aims to classify 7 different types of skin lesions. We used the HAM10000 dataset, which consists of 10000 dermatoscopic images from different populations and is publicly available through the International Skin Imaging Collaboration (ISIC) Archive. The image data was preprocessed to make it suitable for modeling. Exploratory data analysis (EDA) was performed to understand various attributes and their relationships within the dataset. A modified version of a Gaussian filtering algorithm and SFTA was applied for image segmentation and feature extraction. The processed dataset was then fed into the S-MobileNet model. This model was designed to be lightweight and was analysed in three dimensions: using the Relu Activation function, the Mish activation function, and applying compression at intermediary layers. In addition, an alternative approach for compressing layers in the S-MobileNet architecture was applied to ensure a lightweight model that does not compromise on performance. The model was trained using several experiments and assessed using various performance measures, including, loss, accuracy, precision, and the F1-score. Our results demonstrate an improvement in model performance when applying a preprocessing technique. The Mish activation function was shown to outperform Relu. Further, the classification accuracy of the compressed S-MobileNet was shown to outperform S-MobileNet. To conclude, our findings have shown that our proposed deep learning-based S-MobileNet model is the optimal approach for classifying skin lesion images in the HAM10000 dataset. In the future, our approach could be adapted and applied to other datasets, and validated to develop a skin lesion framework that can be utilised in real-time.

The field of autonomous driving research has made significant strides towards achieving full automation, endowing vehicles with self-awareness and independent decision-making. However, integrating automation into vehicular operations presents formidable challenges, especially as these vehicles must seamlessly navigate public roads alongside other cars and pedestrians. An intriguing yet relatively underexplored domain within autonomous driving is overtaking. Overtaking involves a dynamic interplay of complex tasks, including precise steering and speed control, rendering it one of the most intricate operations for implementing augmented intelligence driving technologies. Surprisingly, the overtaking of autonomous vehicles remains largely uncharted territory in the context of augmented intelligence for autonomous systems. This void in knowledge beckons researchers to embark on explorations and investigations in this nascent field. Our review paper systematically synthesises overtaking methodologies hinging on computer vision techniques tailored for augmented intelligence autonomous driving scenarios in response to this pressing need. Our analysis encompasses an array of domains central to overtaking in augmented intelligence autonomous vehicles, encompassing Object Detection, Lane/Line Detection, Depth Estimation, Obstacle Detection, Segmentation, and Pedestrian Detection. We meticulously analyze each domain using well-established Multimodal datasets. We assess different models’ performance across various parameters by employing graphical structures, enabling visual comparative analyses. In object detection, YOLOv4 achieves a top performance with 0.90 mAP on the BDD100K dataset. For lane detection, CLRNET excels with the highest F1 score of around 0.96 on the LLAMAS dataset. ViT-Adapter-L leads in segmentation tasks, boasting an impressive mIoU score of 83 on Cityscapes. The Hierarchical Model achieves a superior mAP of 0.90 in road sign detection on the Tsinghua-Tencent Dataset. Steering angle computation sees InterFuser as the standout, achieving the highest driving score of approximately 74.0. This paper’s primary contributions include a comprehensive assessment of diverse models for each Multimodal dataset, aiding future research in this evolving domain.

Historical patients’ medical data has an important impact on the healthcare industry for providing the best care to patients through intelligent health diagnosis and prediction of diseases. The existing intelligent health diagnosis systems collect data from medical institutions or laboratories and then use machine learning algorithms to predict diseases. But, in most cases, the medical institutions have incomplete medical data of the patients since a patient may consult different specialists (from various hospitals) during the treatment process. To overcome this problem, we build a smart and secure federated learning framework for intelligent health diagnosis with a blockchain-based incentive mechanism and nonfungible tokens (NFTs)-based marketplace. We make use of NFTs to develop clear demarkations on the ownership and accessibility of the data of patients. We create an NFT marketplace that manages access to the historical medical data of patients. A comprehensive incentive mechanism based on several factors, including the quality and relevance of the data, the frequency, regularity of data uploading, etc., is incorporated to encourage and penalize the patients based on their contributions to the global model. We used the Polyak-averaging technique for aggregating local models to form a global model. The extensive analysis shows that the proposed model achieves comparable performance with the centralized machine learning models while affording better security and access to better data. The results also show the efficacy of the proposed blockchain-based incentive mechanism.

Real-time video stream monitoring is gaining huge attention lately with an effort to fully automate this process. On the other hand, reporting can be a tedious task, requiring manual inspection of several hours of daily clippings. Errors are likely to occur because of the repetitive nature of the task causing mental strain on operators. There is a need for an automated system that is capable of real-time video stream monitoring in social systems and reporting them. In this article, we provide a tool aiming to automate the process of anomaly detection and reporting. We combine anomaly detection and video captioning models to create a pipeline for anomaly reporting in descriptive form. A new set of labels by creating descriptive captions for the videos collected from the UCF-Crime (University of Central Florida-Crime) dataset has been formulated. The anomaly detection model is trained on the UCF-Crime, and the captioning model is trained with the newly created labeled set UCF-Crime video description (UCFC-VD). The tool will be used for performing the combined task of anomaly detection and captioning. Automated anomaly captioning would be useful in the efficient reporting of video surveillance data in different social scenarios. Several testing and evaluation techniques were performed. Source code and dataset:https://github.com/Adit31/Captionomaly-Deep-Learning-Toolbox-for-Anomaly-Captioning.

Fake news has spread across social media platforms and with the ease of access, negative consequences have come with it on individuals and society. This issue has become a focus of interest among various research communities, including artificial intelligence (AI) researchers. Existing AI-based fake news detection techniques primarily make use of a 1-D convolutional neural network (1D-CNN) with unidirectional word embedding. We propose a multichannel deep convolutional neural network (CNN) with different kernel sizes and filters as an AI technique. Multiple embedding of the same dimension with different kernel sizes technically allows the news article to be processed at different resolutions of different n-grams at the same time. Different kernel sizes increase the learning ability of the proposed classification model. The proposed model determines how to integrate these interpretations (different n-grams) most suitably. Three real-world fake news datasets were used in experiments to validate the classification performance. The classification results showed that the proposed model has high accuracy in detecting fake news. Regardless of the dataset, the proposed model can be used for fake news detection in binary classification problems.

Designing an Emergency Vehicle Management (EVM) system that can provide competent services with the shortest possible delay is challenging. This is primarily due to the highly complex environments in which they are deployed and the diverse scenarios a vehicle could face after deployment. Compared to other types of vehicles, emergency vehicles operate under time-critical circumstances, move at much higher speeds, and require much greater alertness from the driver. Traditionally, emergency vehicles have not relied on dedicated smart/intelligent systems for increased safety and effectiveness. While much research has been carried out to improve the services of commercial and non-commercial vehicles in general, road emergency vehicles have not received the same research attention. The role of emergency vehicles in new environments (e.g., smart cities) will become increasingly important. To address this issue, we review state-of-the-art research results reported to date, focusing on the deployment and use of EVM. We discuss different dimensions of EVM which include route planning, patient information retrieval, accident detection, driver inattention detection, and several assisting tools and technologies. We identify key advancements in Intelligent Transportation Systems that can be leveraged for safe and effective EVM. We identify challenges inhibiting the usage and deployment of such advancements in EVM, and finally we identify future research directions for a more effective EVM that current developments in ITS have not yet addressed. We hope this work will serve as an accurate and updated source of information for the ever-evolving field of smart cities and ITS.

Traditional multimedia experiences are undergoing a transformation as generative AI integration fosters enhanced creative workflows, streamlines content creation processes, and unlocks the potential for entirely new forms of multimedia storytelling. It has potential to generate captivating visuals to accompany a documentary based solely on historical text descriptions, or creating personalized and interactive multimedia experiences tailored to individual user preferences. From the high-resolution cameras in our smartphones to the immersive experiences offered by the latest technologies, the impact of generative imaging undeniable. This study delves into the burgeoning field of generative AI, with a focus on its revolutionary impact on image generation. It explores the background of traditional imaging in consumer electronics and the motivations for integrating AI, leading to enhanced capabilities in various applications. The research critically examines current advancements in state-of-the-art technologies like DALL-E 2, Craiyon, Stable Diffusion, Imagen, Jasper, NightCafe, and Deep AI, assessing their performance on parameters such as image quality, diversity, and efficiency. It also addresses the limitations and ethical challenges posed by this integration, balancing creative autonomy with AI automation. The novelty of this work lies in its comprehensive analysis and comparison of these AI systems, providing insightful results that highlight both their strengths and areas for improvement. The conclusion underscores the transformative potential of generative AI in image generation, paving the way for future research and development to further enhance and refine these technologies. This article serves as a critical guide for understanding the current landscape and future prospects of AI-driven image creation, offering a glimpse into the evolving synergy between human creativity and artificial intelligence.

Abstract The prediction of news popularity is having substantial importance for the digital advertisement community in terms of selecting and engaging users. Traditional approaches are based on empirical data collected through surveys and applied statistical measures to prove a hypothesis. However, predicting news popularity based on statistical measures applied to past data is highly questionable. Therefore, in this article, we predict news popularity using machine learning classification models and deep residual neural network models. Articles are usually made up of textual content and in many cases, images are also used. Although it is evident that the appropriate amount of textual data is required to extract features and create models, image data is also helpful in gaining useful information. In this article, we present a novel multimodal online news popularity prediction model based on ensemble learning. This research work acts as a guide for extensive feature engineering, feature extraction, feature selection and effective modelling to create a robust news popularity Prediction Model. Three kinds of features—meta‐features, text features and image features are used to design an influential and robust model. The relative error performance measure Root Mean Squared logarithmic error (RMSLE) is used to quantify the popularity prediction error. Further, the RMSLE outcome shows 0.351 which is the lowest error value given by the proposed model. Further, the most important features are also sought out to show the dependence of the best‐fit model on text and image features.

Abstract Online social networks have grown exponentially in the recent years while finding applications in real life like marketing, recommendation systems, and social awareness campaigns. An important research area in this field is Influence Maximization, which pertains to finding methods for maximizing the spread of information (influence) across an OSN. Existing works in IM widely use a pre‐defined edge propagation probability for node activation. Hurst exponent (H), which depicts the self‐similarity in the time series depicting a user's past interaction behaviour, has also been used as activation criteria. In this work, we propose a Time Series Characteristic based Hurst‐based Diffusion Model (TSC‐HDM), which calculates H based on the stationary or non‐stationary characteristic of the time series. TSC‐HDM selects a handful of seed nodes and activates a seed node's inactive successor only if H > 0.5. The proposed model has been tested on four real‐world OSN datasets. The results have been compared against four other IM models – Independent Cascade, Weighted Cascade, Trivalency, and Hurst‐based Influence Maximization. TSC‐HDM is found to have achieved as much as 590% higher expected influence spread as compared to the other models. Moreover, TSC‐HDM has attained 344% better average influence spread than other state‐of‐the‐art models namely LIR, A‐Greedy, LPIMA, Genetic Algorithm with Dynamic Probabilities, NeighborsRemove, DegreeDecrease, IGIM, IRR, and PHG.

Blockchain (BC) and artificial intelligence (AI) technologies have independent applications in multiple industries, including banking, finance, healthcare, construction, transportation, hospitality, manufacturing, and insurance, to name a few. Moreover, these two technologies can be integrated seamlessly, thanks to their complementary and mutually supportive features. AI algorithms can make the medical BC storage efficient by their processing algorithms, also playing the role of knowledgeable gatekeepers. BC can support AI models by providing secure, sizeable, traceable, diverse, and immutable healthcare data for the training purpose. The integration of BC and AI has multiple use cases in the healthcare industry ranging from disease prediction to pandemic management. Previously, researchers have reviewed the applications of each of these technologies in healthcare independently. Although the integration of BC and AI has been fruitful, to the best of our knowledge, there has been no work in the past reviewing the confluence of these two technologies in the healthcare sector. We have classified the works based on two different classification schemes: 1) application-based and 2) AI-training paradigm-based classification. We have also provided a compilation of tools used in the integrated systems of BC and AI for healthcare. We identified that the integration of BC and AI technologies had been applied in quite different areas of healthcare ranging from biomedical research to pandemic management. It is also noted that the supervised learning algorithms and federated learning paradigm for secure decentralized AI model training are often used in the integration. Our findings reveal that majority of the reviewed works use BC as a secure database for AI models. Furthermore, we also have pointed out the potential applications of these two technologies in healthcare.

Health insurance is crucial for each person, bearing in mind the increasing medical costs. COVID-19 has been an eye-opener as to how important it is to have health insurance. Medical emergencies can have a severe emotional and financial impact. Thus, a health insurance policy can help mitigate financial risks in unpredictable circumstances. However, the current insurance system is very expensive, as thousands of people pay the premiums, and very few take the claims. Furthermore, the claim settlement process is excruciatingly long and tiresome. In this article, we focus on establishing a rapid and cost-effective framework for the health insurance market, based on machine learning and blockchain technology. By developing a smart contract, blockchain may eliminate any third-party organizations and make the complete process safer, easier, and more efficient. The contract pays the claim based on the claimant’s documentation. We optimized the premiums using a regression model based on the net amount claimed during the current policy tenure and various other criteria. For anticipating risk, a random forest classifier is used, which aids in the risk-rated premium rebate computation for policyholders for their next term of insurance.

Technological interventions are becoming commonplace in everyday vehicles. But utilization of biosignals that can enhance the overall driving experience is still limited. Microsleep is one such issue that needs intervention, owing to the difficulty in its detection and social acceptance of using wearable BCI devices during transit. Microsleep is a short duration of sleep that lasts from few to several seconds. It could occur unconsciously without the person in context realizing it. This, therefore, happens before the deep sleep and could also occur when performing critical tasks such as driving on a highway. By using modern-day advancements in Internet of Things (IoT) and Machine Learning, we can provide efficient solutions to prevent accidents due to microsleep during vehicular transit. However, it is noteworthy that distinguishing microsleep using a single channel system is a challenge. We have explored this using datasets provided by International BCI Competition Committee. Given the fact that the participants’ values might not match the exact scenario, approaches for exploiting transitory phases using ANN/CNN have been developed and discussed in this paper. Transitory phases could include Wakefulness$\leftrightarrow $Non-Rapid Eye Movement-1 phase (NREM-1). Results show ≈95% increase in mean statistical agreements, which are represented by kappa values (CNN NREM$1~\rightarrow $CNN Transition) and ≈77% increase in mean kappa (ANN NREM$1~\rightarrow $ANN Transition). Hence, this work gives an initial indication whether classifiers trained on night sleep data can be used for microsleep detection in more real-world scenarios.

There has been a notable increase in the research and development of Vehicular Ad-hoc Networks (VANETs) to efficiently and safely manage large amounts of traffic. Such networks are, however, also prone to various cyber threats to data integrity, privacy, authentication, and network availability, and given the potential risk to life under the event of a malfunction and misinformation, it is important to provide security measures against such threats. This paper presents the Multi-branch Reconstruction Error (MbRE) Intrusion Detection System (IDS) for edge-based anomaly detection in VANETs for data integrity, network availability and user authentication-based misbehaviors without the need to train on them. Vehicular data is first sequenced and separated into three data branches - frequency (F) derived from the message timestamps, pseudo-identities (I), and the motion data (M) i.e. position and velocity. The proposed model comprises of three Convolutional Neural Networks (CNN)-based reconstruction models trained to reconstruct normal F-I-M vehicular behavior. The IDS classifies each branch of a sequence as 0/1 based on the reconstruction error threshold for the respective branch and, therefore, has the ability to detect 8 possible binary encoded behaviors for each sequence of vehicular data. These results are then used to find the overall behavior of each vehicle using carefully selected detection thresholds. MbRE is able to classify frequency, identity and motion-based behavior samples with an accuracy of 100%, 98.5-100%, and 95.4-100%, respectively, without the need to train on such behaviors. The study also emulates the IDS on Google Colaboratory and Jetson Nano to show its practicality in cloud and edge environments.

Researchers nowadays are particularly focusing on the interpretation of EEG signals to understand and exploit the information they provide for brain activities. Deep learning architectures performing sleep staging have recently grown to their full potential with their ability to learn and interpret highly complex mathematical contexts. This has been catered to owing to the increasing availability of large EEG data sets. In this paper, we describe how sleep staging differs from microsleep prediction. We also provide a fresh methodology for the microsleep classification job that works with even less training data. Our proposed model exploits the attention-based mechanism that clubs the advantages available in Wavelet transform with Short Time Fourier Transform(STFT) Spectrogram. We also put forward a robust deep learning model that contains separate “time-dependent” and “time-independent” parts, which can record contexts from the sequence of features and simultaneously learn intra-epoch relations. A single-electrode EEG signal was employed for our analysis to accommodate such procedures’ social acceptance. For the task of microsleep detection on the MWT dataset, our model achieves fairly high accuracy rates (92% training and 89.9% testing accuracy), and an overall improvement in the kappa value by ≈ 42%, as compared to prior novel approaches.

The future needs of the telecommunication system lie in deploying a heterogeneous ultra-dense network with varied topographical use cases. However, this increase in ultra-denseness in 5g and beyond poses several challenges in resource allocation, requiring an accurate learning-based prediction. This paper proposes a novel framework using Federated Learning (FL) and Distributed Machine Learning (DML) for Mobile Edge based resource provisioning to User Equipment (UEs). This work formulates the correlation-based novel procedures between UEs in applying Federated and Distributed Machine Learning through Kolmogorov tests for predicting SNR. The correlations of the distribution obtained through the Kolmogorov test check the extent of Independent and Identically Distributed (IID) - ness between modelled data and evaluate the global model for resource provisioning accuracy. Further, correlation-based DML is also employed to balance the computational load of a mobile edge, which acts as a small cell base station and a computational node. In this approach, we account for correlation-based resource predictive model training to balance the uniform computational load by data distribution methods among the neighbouring mobile Edge SCBS nodes for computation. Together with both DML and FL, we create a novel Framework for resource prediction with minimal time for achieving high accuracy without over-fitting.

Currently, the world is witnessing a huge growth in additional computing proficiency and extensive network coverage capability, which resulted in a paradigm shift from VANETs to Internet of Vehicles (IoV). Moreover, enhanced network capabilities facilitate enabling of IoV technology for latency-critical applications in energy-constrained smart IoT devices. However, IoV networks demand energy efficiency due to its dynamic nature for which Roadside Units (RSUs) are critical. However, in cities, huge deployment of RSUs and their maintenance is expensive in IoV infrastructure, requiring a trade-off between the network coverage and installation-related expenses. Also, the latency issues in IoV are highly dependent on the positioning of accessible RSUs. Motivated by the above highlighted issues, we propose an upgraded RSU placement method to boost network efficiency through placement of RSUs in optimal locations in a given road map. The Memetic Framework-based Optimal RSU Deployment (MFRD) algorithm is proposed to maximize the coverage area among the vehicles in an IoV and minimize the overlap in the coverage of the different RSUs. We observed from simulation results based on real-world maps that MFRD yields a significantly higher fitness score as compared to the existing state-of-the-art in terms of optimal positioning of the RSUs.

Abstract The traditional target recognition and classification is mostly done manually, with low efficiency and high cost. Improving the level of target recognition automatically has become an important research topic. This paper proposes a target recognition method based on transfer learning to effectively complete the classification and recognition of targets using a brain–computer interface (BCI) model. Based on the construction of the faster‐RCNN deep learning model, the pre‐training of the model is achieved by VGG‐16 and Inception‐v2, and the transfer learning algorithm is used to optimize the faster‐RCNN deep learning model based on the kinematics model. Experiments are carried out with the aim to detect tableware by the persons whose brain signals recognition rate has been substantially improved using faster‐RCNN. Compared with the traditional recognition methods, the results at the lab‐scale level illustrated that the proposed algorithm can effectively improve the speed and accuracy of target recognition by using the BCI model to classify tableware of different colors and shapes in a complex background.

Edge computing (EC), is a technological game changer that has the ability to connect millions of sensors and provide services at the device end. The broad vision of EC integrates storage, processing, monitoring, and control of operations in the Edge of the network. Though EC provides end-to-end connectivity, speeds up operation, and reduces latency of data transfer, security is a major concern. The tremendous growth in the number of Edge Devices and the amount of sensitive information generated at the device and the cloud creates a broad surface of attack and therefore, the need to secure the static and mobile data is imperative. This article is a comprehensive survey that describes the security and privacy issues in various layers of the EC architecture that result from the networking of heterogeneous devices. Second, it discusses the wide range of machine learning and deep learning algorithms that are applied in EC use cases. Following this, this article broadly details the different types of attacks that the Edge network confronts, and the intrusion detection systems and the corresponding machine learning algorithms that overcome these security and privacy concerns. The details of machine learning and deep learning techniques for EC security are tabulated. Finally, the open issues in securing Edge networks and future research directions are provided.

Object Detection in low-light aerial images is a challenging problem due to considerable variation in brightness and varying contrast. Deep Learning-based approaches have recently demonstrated great promise in image enhancement. Many existing neural networks used for image quality enhancement first encode the input into low-resolution representations and then decode these representations back to a higher resolution for the contextual information. However, this method leads to the loss of semantic content. Recent research has demonstrated the advantage of maintaining high-resolution information along with lower resolution representations, which maintains image features throughout the network. In this paper, we propose a novel architecture named RNet for low-light image enhancement of aerial images. The proposed network contains multi-resolution branches for better understanding of different levels of local and global context through different streams. The performance of RNet is evaluated on a recent synthetic dataset. We also present a comprehensive evaluation with a representative set of state-of-the-art enhancement techniques and neural net architectures.

International audience

For the above article[1], the affiliation information is presented here.

Due to the exponential increase in the number of users, the next-generation cellular networks are resource-constrained in power and bandwidth. Power consumption is one of the critical consideration for the next-generation wireless networks, therefore, management of available resources is essential to achieve power efficiency. With the growing incentive to ‘go green’ and to reduce the carbon footprint, the fifth generation (5G) and beyond wireless networks will derive power from renewable sources to solve the energy efficiency problems. This work focuses on integrated regulation of the traditional, i.e., the grid-based and the renewable, i.e., the solar-based power supplies for the 5G and beyond 5G green base stations (BSs) in a smart city scenario. We propose a pricing model for suppliers to charge the BSs for electricity consumption when the renewable power supply cannot meet their total energy requirements. We propose a game-theoretic analysis for cost optimization by proposing two games, i.e., the power control game and the best supplier game. Each BS acts as a game player and has some actions like power reduction and supplier selection to reduce the total energy costs. We also provide the game transition profiles for the BSs. Furthermore, the Nash Equilibrium’s existence is verified for each of these games and an optimal cost solution is proposed for the green BSs.

Vehicular ad-hoc network (VANET) security has been an active area of research over the past decade. However, with the increasing adoption of the Internet of Things (IoT) in VANETs, the number of connected vehicles is set to grow exponentially over the next few years, which translates to a higher number of communication interfaces and a greater possibility of cybersecurity attacks. Along with these cybersecurity attacks, the instances of compromised vehicles sending faulty information about their positions and speeds also increase exponentially. Thus, there is a need to augment the existing security schemes with anomaly detection schemes which can differentiate normal vehicle data from malicious and faulty data. Since, the number of anomaly types can be many, deep neural networks would work best in this scenario. In this paper, we propose a deep neural network-based vehicle anomaly detection scheme. We use a sequence reconstruction approach to differentiate normal vehicle data from anomalous data. Numerical results show that we can correctly detect data corresponding to several anomaly types.

Having a health insurance is important for everybody, bearing in mind the increasing medical costs. Medical emergencies can have a severe financial and emotional impact. However, the current insurance system is very expensive and the claim settlement process is excessively lengthy, making it tedious. This results in policyholders not being able to successfully make a claim with their insurance company. In this paper, we focus on developing a fast and cost-effective framework based on blockchain technology and machine learning for the health insurance industry. Blockchain is capable of removing all third-party organisations by forming a smart contract, making the entire process more smooth, secure, and efficient. The contract settles the claim on documents submitted by the claimant. A ridge regression model is used for computing the premiums optimally, based on the total amount claimed under the current policy tenure, along with several other factors. A random forest classifier is applied for predicting the risk that helps in the computation of risk-rated premium rebate.

In recent times, there is a paradigm shift to cloud services that offer on-demand computer system resources, especially data storage and computing power. The main reason for the shift is that it removes the user's active participation to perform computationally intensive tasks. However, current cloud-based services incur high user latency as being deployed very far from the user. One alternative solution to the traditional cloud-based paradigm is drone-based edge computing. In drone edge computing, drones are located near the user and deployed to provide data offload services. There have been many works that have addressed the issue of efficient task assignment in edge devices. This paper presents a concrete analytical performance model for drone cloudlet networks and factors that influence the service response time to the user. The results can be helpful for network administrators to make the current edge computing paradigm faster, more robust and, cost-effective.

This article provides a literature review of state-of-the-art machine learning (ML) algorithms for disaster and pandemic management. Most nations are concerned about disasters and pandemics, which, in general, are highly unlikely events. To date, various technologies, such as IoT, object sensing, UAV, 5G, and cellular networks, smartphone-based system, and satellite-based systems have been used for disaster and pandemic management. ML algorithms can handle multidimensional, large volumes of data that occur naturally in environments related to disaster and pandemic management and are particularly well suited for important related tasks, such as recognition and classification. ML algorithms are useful for predicting disasters and assisting in disaster management tasks, such as determining crowd evacuation routes, analyzing social media posts, and handling the post-disaster situation. ML algorithms also find great application in pandemic management scenarios, such as predicting pandemics, monitoring pandemic spread, disease diagnosis, etc. This article first presents a tutorial on ML algorithms. It then presents a detailed review of several ML algorithms and how we can combine these algorithms with other technologies to address disaster and pandemic management. It also discusses various challenges, open issues and, directions for future research.

The applications of unmanned aerial vehicles (UAVs) in military, intelligent transportation, agriculture, rescue operations, natural environment mapping, and many other allied domains has increased exponentially during the past few years. Some of the use cases of their applications range from aerial surveillance, data retrieval to their use in real-time communicative networks. Though UAVs were traditionally used only outdoors, many of its indoor applications like for rescue operations, inventory tracking in warehouses, etc., have recently emerged and these use cases are being actively explored. One of the major challenges for indoor drone applications is navigation and obstacle avoidance. Due to indoor operations, the global positioning system fails in accurate localization and navigation. To address this issue, we introduce a scheme that facilitates the autonomous navigation of UAVs (which have an onboard camera) in the indoor corridors of a building using deep-neural-networks-based processing of images. For a deep neural network, the selection of a good combination of hyperparameters for a better prediction is a complicated task. In this article, the hyperparameters tuning of a convolutional neural network is achieved by using genetic algorithms. The proposed architecture (DCNN-GA) is compared with state-of-the-art ImageNet models. The experimental results show the minimum loss and high performance of the proposed algorithm.

Internet of Vehicles (IoVs) are fast becoming the norm in our society, but such a trend also comes with its own set of challenges (e.g., new security and privacy risks due to the expanded attack vectors). In this work, we propose an edge-computing-based secure, efficient, and intelligent multitier heterogeneous IoVs network. We first discuss the functionality and objectives of such an architecture. Then, we demonstrate how unsupervised deep learning techniques can facilitate the identification of suspicious vehicle behavior and ensure the security of such an architecture. The findings from our evaluations demonstrate the learning spatiotemporal information and parameter efficiency of the proposed stacked long short-term memory (LSTM) model over single LSTMs.

The rapid improvement in the global connectivity standards has escalated the level of trade taking place among different parties. Advanced communication standards are allowing the trade of all types of commodities and services. Furthermore, the goods and services developed in a particular region are transcending boundaries to enter into foreign markets. Supply chains play an essential role in the trade of these goods. To be able to realize a connected world with no boundary restrictions in terms of goods and services, it is imperative to keep the associated supply chains transparent, secure, and trustworthy. Therefore, some fundamental changes in the current supply chain architecture are essential to achieve a secure trade environment. This article discusses the supply chain's security-critical application areas and presents a detailed survey of the security issues in the existing supply chain architecture. Various emerging technologies, such as blockchain, machine learning (ML), and physically unclonable functions (PUFs) as solutions to the vulnerabilities in the existing infrastructure of the supply chain have also been discussed. Recent studies reviewed in this work reveal a growing sentiment in the industry toward new and emerging technologies, such as Internet of Things (IoT), blockchain, and ML. While many organizations have already adopted IoT applications and artificial intelligence systems in their businesses, widespread adoption of blockchain remains distant. It has also been found that over the past decade, PUF-based authentication systems have gained much ground. However, a proper reference model for their implementation in complex supply chains is still missing.

Governments and public sector entities around the world are actively exploring new ways to keep up with technological advancements to achieve smart governance, work efficiency, and cost optimization. Blockchain technology is an example of such technology that has been attracting the attention of Governments across the globe in recent years. Enhanced security, improved traceability, and lowest cost infrastructure empower the blockchain to penetrate various domains. Generally, governments release tenders to some third-party organizations for different projects. During this process, different competitors try to eavesdrop the tender values of others to win the tender. The corrupt government officials also charge high bribe to pass the tender in favor of some particular third party. In this article, we presented a secure and transparent framework for government tenders using blockchain. Blockchain is used as a secure and immutable data structure to store the government records that are highly susceptible to tampering. This work aims to create a transparent and secure edge computing infrastructure for the workflow in government tenders to implement government schemes and policies by limiting human supervision to the minimal.

With the recent use of IoT in the field of healthcare, a lot of patient data is being transmitted and made available online. This necessitates sufficient security measures to be put in place to prevent the possibilities of cyberattacks. In this regard, several authentication techniques have been designed in recent times to mitigate these challenges, but the physical security of the healthcare IoT devices against node tampering and node replacement attacks, in particular, is not addressed sufficiently in the literature. To address these challenges, a two-way two-stage authentication protocol using hardware security primitives called Physical Unclonable Functions (PUFs) is presented in this paper. Considering the memory and energy constraints of healthcare IoT devices, this protocol is made very lightweight. A formal security evaluation of this protocol is done to prove its validity. We also compare it with relevant protocols in the healthcare IoT scenario in terms of computation time and security to show its suitability and robustness.

Summary The emergence of mobile cloud computing enables mobile users to offload computation tasks to other resource‐rich mobile devices to reduce energy consumption and enhance performance. A direct peer‐to‐peer connection among mobile devices to offload computation tasks can be a highly promising solution to provide a fast mechanism, especially for deadline‐sensitive offloading tasks. The generic blockchain‐based system might fail in such a scenario due to it being a heavyweight mechanism requiring high power consumption in the mining process. To address these issues, in this article, we propose a directed acyclic graph‐enabled mobile offloading (DAGMO) algorithm. DAGMO model is empowered by traditional blockchain features and provides additional advantages to overcome the fundamental limitations of generic blockchain. A game‐theoretic approach is used to model the interactions between mobile devices. The numerical analysis proves the proposed model to enhance the overall welfare of the participating nodes in terms of computation cost and time.

The exponential surge in the number of vehicles on the road has aggravated the traffic congestion problem across the globe. Several attempts have been made over the years to predict the traffic scenario accurately and consequently avoiding further congestion. Crowdsourcing has come forward as one of the most adopted methods for predicting traffic intensity using live data. However, the privacy concerns and the lack of motivation for the live users to help in the traffic prediction process have rendered existing crowdsourcing models inefficient. Towards this end, we present an advanced blockchain-based secure crowdsourcing model. Not only does our model ensure privacy preservation of the users, but by incorporating a revenue model, it also provides them with an incentive to participate in the traffic prediction process willingly. For accurate and efficient traffic jam probability estimation, our work proposes a neural network-based smart contract to be deployed onto the blockchain network. The results reveal that the proposed model is highly efficient in terms of attaining high participation and consequently obtaining highly accurate predictions.

Adverse weather conditions such as fog, haze, snow, mist and glare create visibility problems for applications of autonomous vehicles. To ensure safe and smooth operations in frequent bad weather scenarios, image dehazing is crucial to any vehicular motion and navigation task on road or air. Moreover, the commonly deployed mobile systems are resource constrained in nature. Therefore, it is important to ensure memory, compute and run-time efficiency of dehazing algorithms. In this manuscript we propose ReViewNet, a fast, lightweight and robust dehazing system suitable for autonomous vehicles. The network uses components like spatial feature pooling, quadruple color-cue, multi-look architecture and multi-weighted loss to effectively dehaze images captured by cameras of autonomous vehicles. The effectiveness of the proposed model is analyzed by exhaustive quantitative evaluation on five benchmark datasets demonstrating its supremacy over other existing state-of-the-art methods. Further, a component-wise ablation and loss weight ratio analysis demonstrates the contribution of each and every component of the network. We also show the qualitative analysis with special use cases and visual responses on distinctive vehicular vision instances, establishing the effectiveness of the proposed method in numerous hazy weather conditions for autonomous vehicular applications.

Aerial scenes captured by UAVs have immense potential in IoT applications related to urban surveillance, road and building segmentation, land cover classification, and so on, which are necessary for the evolution of smart cities. The advancements in deep learning have greatly enhanced visual understanding, but the domain of aerial vision remains largely unexplored. Aerial images pose many unique challenges for performing proper scene parsing such as high-resolution data, small-scaled objects, a large number of objects in the camera view, dense clustering of objects, background clutter, and so on, which greatly hinder the performance of the existing deep learning methods. In this work, we propose ISDNet (Instance Segmentation and Detection Network), a novel network to perform instance segmentation and object detection on visual data captured by UAVs. This work enables aerial image analytics for various needs in a smart city. In particular, we use dilated convolutions to generate improved spatial context, leading to better discrimination between foreground and background features. The proposed network efficiently reuses the segment-mask features by propagating them from early stages using residual connections. Furthermore, ISDNet makes use of effective anchors to accommodate varying object scales and sizes. The proposed method obtains state-of-the-art results in the aerial context.

Vehicular Ad-hoc Network (VANET) poses to be a promising technology for the future since it increases the comfort level of the drivers while also enhancing the safety measures for them. The main aim of VANETs is to enable communication among vehicles and roadside units (RSUs) using vehicle-to-vehicle (V2V) and vehicle-to-RSU (V2R) networks. VANET applications have a vast potential for growth owing to the increasing number of smart cities around the globe and advancement taking place in the technology sector. However, with all their benefits, VANETs also face several security challenges. The sensitive nature of data being transferred turns VANETs prone to malicious attacks. To overcome the security challenges, this paper proposes a distributed Directed Acyclic Graph (DAG) enabled vehicular network comprising several requesting vehicles and RSUs. The proposed model is based on advanced blockchain and therefore provides a strong level of security and data immutability. Furthermore, the interactions between the requesting vehicles and the RSUs have been modeled using an auction-based game-theoretic smart contract deployed on the blockchain.

The rapid increase in the number of mobile devices across the globe has brought a new challenge to the forefront, one of mobile traffic management. The ever-increasing number of mobile devices leads to the generation of a large amount of data and computationally intensive applications, which contributes heavily to cellular network congestion. To solve this issue, we propose a mobile data offloading scheme based on a distributed ledger technology (DLT). Existing mobile data offloading schemes based on DLT employ conventional blockchain to set up a peer-to-peer (P2P) network of mobile users. Although these schemes have gained ground in improving the Quality of Experience (QoE) for end-users, they lack efficiency and scalability. Furthermore, generic blockchain does not provide timestamp ordering of events, which is necessary to ensure the computation of delay-sensitive tasks. To overcome these challenges, we propose the use of a directed acyclic graph (DAG) data structure for mobile data offloading. Finally, to ensure time and cost optimality, a game-theoretic approach has been proposed in this paper.

The rapidly increasing mobile traffic across the globe has proliferated the deployment of cellular base stations, which has, in turn, led to an increase in the power consumption and carbon footprint of the telecommunications industry. In recent times, solar-powered base stations (SPBSs) have gained much popularity in the telecom sector due to their ability to make operations more sustainable. However, some potential energy benefits rendered by the SPBSs have not yet been realized. In areas with dense base station deployment or low mobile traffic, SPBSs store surplus energy, which, in most instances, gets lost due to limited charge storage capacity of the batteries. To limit the wastage of energy, an appropriate mechanism enabling the utilization of excess energy produced by these base stations can be adopted. To this end, we model a Base Station-to-Grid (BS2G) network in which the grid can utilize surplus energy spared by the SPBSs. To overcome challenges in regards to scalability, robustness, and cost-optimization, we propose using the blockchain technology to create the BS2G network. Blockchain is a distributed ledger designed to record transactions in a transparent, lightweight, and tamper-proof manner. To make energy trade between base stations and the grid cost-effective, a game-theoretical approach has also been adopted in this paper. The proposed model simplifies the process of energy trading while also making it cost-optimal.

With the increasing incidence rate of lung cancer patients, early diagnosis could help in reducing the mortality rate. However, accurate recognition of cancerous lesions is immensely challenging owing to factors such as low contrast variation, heterogeneity and visual similarity between benign and malignant nodules. Deep learning techniques have been very effective in performing natural image segmentation with robustness to previously unseen situations, reasonable scale invariance and the ability to detect even minute differences. However, they usually fail to learn domain‐specific features due to the limited amount of available data and domain agnostic nature of these techniques. This work presents an ensemble framework Deep3DSCan for lung cancer segmentation and classification. The deep 3D segmentation network generates the 3D volume of interest from computed tomography scans of patients. The deep features and handcrafted descriptors are extracted using a fine‐tuned residual network and morphological techniques, respectively. Finally, the fused features are used for cancer classification. The experiments were conducted on the publicly available LUNA16 dataset. For the segmentation, the authors achieved an accuracy of 0.927, significant improvement over the template matching technique, which had achieved an accuracy of 0.927. For the detection, previous state‐of‐the‐art is 0.866, while ours is 0.883.

The notion of aggregation of data in Industrial Internet of Things (IIoT) environment is a common practice. It shortens the data and associated signatures to reduce the bandwidth requirement. The compact aggregate signature (CAS) scheme creates a constant length aggregate signature (AS). Thus, the length of the CAS is independent of the number of messages or signatures to be aggregated. This article presents the first pairing-free CAS scheme in certificate-based settings. Due to the certificate-based approach, the proposed scheme is free from key escrow and key distribution problems inherited in identity-based cryptography (IDC) and certificate-less cryptography (CLC), respectively. Being compact and pairing free, it is the least bandwidth-consuming and the most efficient provably secure aggregation method. The length and computational cost analysis show that the scheme is the most appealing to use in the IIoT environment.

Stock exchanges around the world are exploring the best possible solution that can improve trading efficiency, lower the risks and tighten secu- rity levels. The working and functioning of a stock exchange involves very hectic and cumbersome pro- cedures which are time consuming, cost inefficient and can be prone to numerous risks. Machine learning and Blockchain are most popular upcoming technologies. In this paper we present a novel secure and de- centralized intelligent stock market prediction model. We present a blockchain based solution for stock exchange model that uses machine learning accessible smart contracts. The machine learning model makes a prediction on the future of the stock market providing an intelligent solution for secure stock market.

Base stations equipped with resources to harvest renewable energy are not only environment-friendly but can also reduce the grid energy consumed, thus bringing cost savings for the cellular network operators. Intelligent management of the harvested energy can further increase the cost savings. Such management of energy savings has to be carefully coupled with managing the quality of service so as to ensure customer satisfaction. In such a process, there is a trade-off between the energy drawn from grid and the quality of service. Unlike prior studies which mainly focus on network energy minimization, this paper proposes a framework for jointly managing the grid energy savings and the quality of service (in terms of the network latency), which is achieved by downlink power control and user association reconfiguration. We use a real BS deployment scenario from London, U.K., to show the performance of our proposed framework and compare it against existing benchmarks. We show that the proposed framework can lead to around 60% grid energy savings as well as better network latency performance than the traditionally used scheme.

One of the major issues in the deployment of solar powered base stations (BSs) is to dimension the photovoltaic (PV) panel and battery size resources, while satisfying outage constraints with least cost. The fundamental step in this dimensioning is to evaluate the power outage probability associated with a particular configuration of PV panel and battery size. This paper addresses this issue by first proposing an analytic model to evaluate the power outage probability of a solar powered BS. The proposed model accounts for hourly as well as daily variation in the harvested solar energy as well as the load dependent BS power consumption. The model evaluates the steady state probability of the battery level, which is then used to estimate the BS power outage probability. Next, given a tolerable power outage probability, we address the problem of obtaining the cost-optimal PV panel and battery dimensions for the BS. The proposed model and the framework have been evaluated using empirical solar energy data for geographically diverse locations.

Solar powered cellular base stations are emerging as a key solution in green cellular networks. A major challenge in the design of such a base station (BS) is finding the optimal cost configuration of the photo-voltaic (PV) panel size and number of batteries which meets a tolerable outage probability with the least cost. One of the fundamental steps in this process is to calculate the outage probability associated with a particular PV panel size and battery size configuration. To address this issue, this paper proposes an analytic model to evaluate the outage probability of a solar powered BS. The proposed model factors in the daily and hourly variations in the harvested solar energy and the traffic dependent BS load, and develops a discrete-time Markov process to model the battery level and thus the outage probability of the BS. Simulation results with empirical solar irradiance data for three different locations are used to validate the proposed model and demonstrate its accuracy.

The deployment of cellular network infrastructure powered by renewable energy sources is gaining popularity as an avenue to provide coverage in areas without reliable grid power and also as a means to reduce the environmental impact of the telecommunications industry. To facilitate the deployment of such networks, this paper addresses the problem of resource provisioning and dimensioning solar powered base stations in terms of the required battery capacity and photo-voltaic (PV) panel sizing. The paper first develops a framework for evaluating the outage probability associated with a base station at a given location as a function of the battery and panel size, by using the solar energy and traffic profiles as inputs. A model is then proposed to evaluate the optimal battery and PV panel sizing, subject to the desired limit on the worst month outage probability. The proposed framework for dimensioning the base station's energy resource requirements has been evaluated using real solar irradiation data for multiple locations.