Search

Abstract Along with the computer application technology progress, machine learning, and block-chain techniques have been applied comprehensively in various fields. The application of machine learning, and block-chain techniques into medical image retrieval, classification and auxiliary diagnosis has become one of the research hotspots at present. Brain tumor is one of the major diseases threatening human life. The number of deaths caused by these diseases is increasing dramatically every year in the world. Aiming at the classification problem of brain CT images in healthcare. We propose a Mask RCNN with attention mechanism method in this research. First, the ResNet-10 is utilized as the backbone model to extract local features of the input brain CT images. In the partial residual module, the standard convolution is substituted by deformable convolution. Then, the spatial attention mechanism and the channel attention mechanism are connected in parallel. The deformable convolution is embedded to the two modules to extract global features. Finally, the loss function is improved to further optimize the precision of target edge segmentation in the Mask RCNN branch. Finally, we make experiments on public brain CT data set, the results show that the proposed image classification fragrance can effectively refine the edge features, increase the degree of separation between target and background, and improve the classification effect.

Skin cancer is one of the most common, deadly, and widespread cancers worldwide. Early detection of skin cancer can lead to reduced death rates. A dermatologist or primary care physician can use a dermatoscope to inspect a patient to diagnose skin disorders visually. Early detection of skin cancer is essential, and in order to confirm the diagnosis and determine the most appropriate course of therapy, patients should undergo a biopsy and a histological evaluation. Significant advancements have been made recently as the accuracy of skin cancer categorization by automated deep learning systems matches that of dermatologists. Though progress has been made, there is still a lack of a widely accepted, clinically reliable method for diagnosing skin cancer. This article presented four variants of the Convolutional Neural Network (CNN) model (i.e., original CNN, no batch normalization CNN, few filters CNN, and strided CNN) for the classification and prediction of skin cancer in lesion images with the aim of helping physicians in their diagnosis. Further, it presents the hybrid models CNN-Support Vector Machine (CNNSVM), CNN-Random Forest (CNNRF), and CNN-Logistic Regression (CNNLR), using a grid search for the best parameters. Exploratory Data Analysis (EDA) and random oversampling are performed to normalize and balance the data. The CNN models (original CNN, strided, and CNNSVM) obtained an accuracy rate of 98%. In contrast, CNNRF and CNNLR obtained an accuracy rate of 99% for skin cancer prediction on a HAM10000 dataset of 10,015 dermoscopic images. The encouraging outcomes demonstrate the effectiveness of the proposed method and show that improving the performance of skin cancer diagnosis requires including the patient's metadata with the lesion image.

Collective intelligence systems like Chat Generative Pre-Trained Transformer (ChatGPT) have emerged. They have brought both promise and peril to cybersecurity and privacy protection. This study introduces novel approaches to harness the power of artificial intelligence (AI) and big data analytics to enhance security and privacy in this new era. Contributions could explore topics such as: leveraging natural language processing (NLP) in ChatGPT-like systems to strengthen information security; evaluating privacy-enhancing technologies to maximize data utility while minimizing personal data exposure; modeling human behavior and agency to build secure and ethical human-centric systems; applying machine learning to detect threats and vulnerabilities in a data-driven manner; using analytics to preserve privacy in large datasets while enabling value creation; crafting AI techniques that operate in a trustworthy and explainable manner. This article advances the state-of-the-art at the intersection of cybersecurity, privacy, human factors, ethics, and cutting-edge AI, providing impactful solutions to emerging challenges. Our research presents a revolutionary approach to malware detection that leverages deep learning (DL) based methodologies to automatically learn features from raw data. Our approach involves constructing a grayscale image from a malware file and extracting features to minimize its size. This process affords us the ability to discern patterns that might remain hidden from other techniques, enabling us to utilize convolutional neural networks (CNNs) to learn from these grayscale images and a stacking ensemble to classify malware. The goal is to model a highly complex nonlinear function with parameters that can be optimized to achieve superior performance. To test our approach, we ran it on over 6,414 malware variants and 2,050 benign files from the MalImg collection, resulting in an impressive 99.86 percent validation accuracy for malware detection. Furthermore, we conducted a classification experiment on 15 malware families and 13 tests with varying parameters to compare our model to other comparable research. Our model outperformed most of the similar research with detection accuracy ranging from 47.07% to 99.81% and a significant increase in detection performance. Our results demonstrate the efficacy of our approach, which unlocks the hidden patterns that underlie complex systems, advancing the frontiers of computational security.

IntroductionGlobal Cardiovascular disease (CVD) is still one of the leading causes of death and requires the enhancement of diagnostic methods for the effective detection of early signs and prediction of the disease outcomes. The current diagnostic tools are cumbersome and imprecise especially with complex diseases, thus emphasizing the incorporation of new machine learning applications in differential diagnosis.MethodsThis paper presents a new machine learning approach that uses MICE for mitigating missing data, the IQR for handling outliers and SMOTE to address first imbalance distance. Additionally, to select optimal features, we introduce the Hybrid 2-Tier Grasshopper Optimization with L2 regularization methodology which we call GOL2-2T. One of the promising methods to improve the predictive modelling is an Adaboost decision fusion (ABDF) ensemble learning algorithm with babysitting technique implemented for the hyperparameters tuning. The accuracy, recall, and AUC score will be considered as the measures for assessing the model.ResultsOn the results, our heart disease prediction model yielded an accuracy of 83.0%, and a balanced F1 score of 84.0%. The integration of SMOTE, IQR outlier detection, MICE, and GOL2-2T feature selection enhances robustness while improving the predictive performance. ABDF removed the impurities in the model and elaborated its effectiveness, which proved to be high on predicting the heart disease.DiscussionThese findings demonstrate the effectiveness of additional machine learning methodologies in medical diagnostics, including early recognition improvements and trustworthy tools for clinicians. But yes, the model’s use and extent of work depends on the dataset used for it really. Further work is needed to replicate the model across different datasets and samples: as for most models, it will be important to see if the results are generalizable to populations that are not representative of the patient population that was used for the current study.

Parkinson's disease (PD) is a globally significant health challenge, necessitating accurate and timely diagnostic methods to facilitate effective treatment and intervention. In recent years, self-supervised deep representation pattern learning (SS-DRPL) has emerged as a promising approach for extracting valuable representations from data, offering the potential to enhance the efficiency of voice-based PD detection. This research study focuses on investigating the utilization of SS-DRPL in conjunction with deep learning algorithms for voice-based PD classification. This study encompasses a comprehensive evaluation aimed at assessing the accuracy of various predictive models, particularly deep learning methods when combined with SS-DRPL. Two deep learning architectures, namely hybrid Long Short-Term Memory and Recurrent Neural Networks (LSTM-RNN) and Deep Neural Networks (DNN), are employed and compared in terms of their ability to detect voice-based PD cases accurately. Additionally, several traditional machine learning models are also included to establish a baseline for comparison. The findings of the study reveal that the incorporation of SS-DRPL leads to improved model performance across all experimental setups. Notably, the LSTM-RNN architecture augmented with SS-DRPL achieves the highest F1-score of 0.94, indicating its superior ability to detect PD cases using voice-based data effectively. This outcome underscores the efficacy of SS-DRPL in enabling deep learning models to learn intricate patterns and correlations within the data, thereby facilitating more accurate PD classification.

As more aerial imagery becomes readily available, massive volumes of data are being gathered constantly. Several groups can benefit from the data provided by this geographical imagery. However, it is time-consuming to manually analyze each image to gain information on land cover. This research suggests using deep learning methods for precise and rapid pixel-by-pixel classification of aerial imagery for land cover analysis, which would be a significant step forward in resolving this issue. The suggested method has several steps, such as the augmentation and transformation of data, the selection of deep learning models, and the final prediction. The study uses the three most popular deep learning models (Vanilla-UNet, ResNet50 UNet, and DeepLabV3 ResNet50) for the experiments. According to the experimental results, the ResNet50 UNet model achieved an accuracy of 94.37%, the DeepLabV3 ResNet50 model achieved an accuracy of 94.77%, and the Vanilla-UNet model achieved an accuracy of 91.31%. The accuracy, precision, recall, and F1-score of DeepLabV3 and ResNet50 are higher than those of the other two models. The proposed approach is also compared to the existing UNet approach, and the proposed approaches have produced greater probability prediction scores than the conventional UNet model for all classes. Our approach outperforms model DeepLabV3 ResNet50 on aerial image datasets based on the performance.

Abstract As cardiovascular disorders are prevalent, there is a growing demand for reliable and precise diagnostic methods within this domain. Audio signal-based heart disease detection is a promising area of research that leverages sound signals generated by the heart to identify and diagnose cardiovascular disorders. Machine learning (ML) and deep learning (DL) techniques are pivotal in classifying and identifying heart disease from audio signals. This study investigates ML and DL techniques to detect heart disease by analyzing noisy sound signals. This study employed two subsets of datasets from the PASCAL CHALLENGE having real heart audios. The research process and visually depict signals using spectrograms and Mel-Frequency Cepstral Coefficients (MFCCs). We employ data augmentation to improve the model’s performance by introducing synthetic noise to the heart sound signals. In addition, a feature ensembler is developed to integrate various audio feature extraction techniques. Several machine learning and deep learning classifiers are utilized for heart disease detection. Among the numerous models studied and previous study findings, the multilayer perceptron model performed best, with an accuracy rate of 95.65%. This study demonstrates the potential of this methodology in accurately detecting heart disease from sound signals. These findings present promising opportunities for enhancing medical diagnosis and patient care.

Energy providers and the power grid are severely harmed by electricity theft, which also causes economic and non-technical losses. Energy theft causes a decline in power quality and overall profitability. Smart grids may address the problem of power theft by merging data and energy flow. The analysis of smart grid data helps to find power theft. The prior methods, however, could have done a better job of identifying energy theft. In this research, we presented an active learning-based machine learning model for energy theft identification and classification of a smart grid. The suggested approach is based on the following steps. We use a dataset from the Open Energy Data Initiative (OEDI), an energy research database that gets information from numerous OEDI offices and labs. Next, we pre-process the data and employ machine learning methods like Active Learning (AL) based Random Forests (RFAL), eXtreme Gradient Boosting (XGboostAL), Decision Tree (DTAL), Gradient Boosting (GBAL), K-Nearest Neighbors (KNNAL), Categorical Boosting (CatboostAL) and Light Gradient Boosting Machine (LGBMAL) classifier. Using the smart grid-based energy theft detection dataset, the proposed RFAL model outperforms competing models and obtains an accuracy of 70.61%. The principles of smart grid tasks streamline decisions and enhance interaction between humans and machines by combining AL with machine learning. The application of this technology in this area has the potential to enhance the accuracy of energy theft detection and electricity-related problems and consequences.

This research explores the potential of technologies in human activity recognition among the elderly population. More precisely, using sensor data and implementing Active Learning (AL), Machine Learning (ML), and Deep learning (DL) techniques for elderly activity recognition. Moreover, the study leverages the HAR70+ dataset, providing insight into the daily activities of older individuals and AL-based ML and DL techniques to construct predictive models for these activities. The findings have implications for proactive and personalized elderly care, representing an approach to improving prediction performance in this domain. The research experiments are presented systematically, summarizing the outcomes of various machine-learning models across three iterative experiments. This research explored a diverse array of ML algorithms, including Random Forest (RF), Extreme Gradient Boosting (XGBoost), Logistic Regression (LR), K-Nearest Neighbors (KNN), Stochastic Gradient Descent (SGB) and DL methods such as Deep Neural Networks (DNN) and Long Short-Term Memory networks (LSTM) for experimentation. This research trained models on 7 activities: walking, shuffling, climbing stairs (up and down), standing, sitting, and lying down, and 4 activities separately: standing, sitting, walking, and lying down, using the same classifiers. Results reveal that LSTM achieved the best accuracy of 0.95% for 7 activities and 0.96% using RF on 4 actives, showing the potential of DL and ML techniques, particularly when integrated with AL, to enhance activity recognition rate, patient care, optimize medication strategies and improve the well-being of elderly individuals. Hence, the findings presented in this study have showcased the potential to enhance the quality of life for seniors using the blend of ML, DL and AL.

This research delves into applying active and machine learning techniques to predict student anxiety. This research explores how these technologies can be explored to understand and predict student anxiety levels. This study utilizes active learning strategies to increase the effectiveness of machine learning models in predicting anxiety levels among students. Additionally, this adaptation emphasizes the usefulness of active learning methodologies in enhancing the precision of machine learning models for student anxiety prediction. This study uses two datasets containing information on student behavior and leverages machine learning methods to construct predictive models for student anxiety. This study uses various machine learning models: K-Nearest Neighbors (KNN), Logistic Regression (LR), XGBoost (XGB), Naive Bayes (NB), and Random Forest (RF). Experiments revealed that active learning-based LR yielded a score of 0.61, and RF performed well with an average accuracy of 0.60 on the first dataset. Similarly, for the second dataset, RF is the most effective model, achieving an accuracy of 0.83. These results provide valuable insights into the models’ performance across key metrics. Further, this research highlights the potential of employing machine learning techniques and active learning methodologies to predict and manage student anxiety.

This paper introduces PureFed, an innovative Federated Learning (FL) framework designed for efficiency, collaboration, and trustworthiness. In the background of FL research, it was observed that previous frameworks often neglected participant privacy, a critical aspect not aligned with the core FL concept. Additionally, there was room for improving the efficiency of existing frameworks. PureFed addresses these shortcomings by offering participants the flexibility to initiate FL tasks or join existing ones without sharing any private data and removing unnecessary actions that led to an inefficient system. Leveraging blockchain technology, it employs smart contracts to ensure traceability and immutability, enhancing the security of the framework. Additionally, PureFed employs symmetric key encryption and dual digital signature mechanisms using ECDSA to guarantee the confidentiality and integrity of shared models. To expedite model convergence, PureFed incorporates a dynamic aggregation scheme, selecting the most suitable model from three distinct techniques: FedAvg, accuracy-based, and loss-based. Furthermore, the framework introduces a dynamic incentive and punishment mechanism to incentivize collaboration and maintain trust. Extensive performance evaluations reveal PureFed’s significant advantages. It outperforms its counterparts by 63.39% and 67.72% in terms of smart contract deployment and interaction gas costs, respectively. Lastly, scalability analyses indicate PureFed’s ability to adapt efficiently, achieving target accuracy in fewer rounds.

Suspicious Human activity recognition (SHAR) is crucial for improving surveillance and security systems by recognizing and reducing possible hazards in different situations. This study focuses on the task of precisely identifying potentially suspicious human behaviour by utilizing an innovative approach that harnesses advanced deep learning methods. Despite the abundance of research on the subject of SHAR, current methods frequently need to be revised with restricted levels of precision and efficiency. This research aims to address these constraints by presenting a thorough methodology for detecting and recognizing suspicious human activities. By rigorously collecting and preparing data, as well as training models, we aim to solve the issue of inaccurate and inefficient activity recognition in surveillance systems. By utilizing Convolutional Neural Networks (CNNs) and deep learning structures, such as the proposed time-distributed CNN model and Conv3D model, we attain notably enhanced accuracy rates of 90.14% and 88.23%, respectively, surpassing current research approaches. Moreover, the efficacy of our approach is illustrated by conducting prediction experiments on previously unreported test data and YouTube videos. Through the process of evaluating the trained models on unseen test data, we ascertain their accuracy and ability to apply learned knowledge to new situations. Moreover, the algorithms are utilized to predict dubious human conduct in a YouTube video, demonstrating their practical usefulness in real-life surveillance situations. The results of this study have important consequences for improving surveillance and security systems, allowing for better identification and reduction of possible dangers in various settings. Our methodology enhances the precision and effectiveness of SHAR, advancing the construction of more resilient and dependable surveillance systems and ultimately strengthening public safety and security.

Smartwatches with cutting-edge sensors are becoming commonplace in our daily lives. Despite their widespread use, it can be challenging to interpret accelerometer and gyroscope data efficiently for Human Activity Recognition (HAR). This study explores active learning integrated with machine learning, intending to maximize the use of smartwatch technology across a range of applications. The previous research on the HAR lacks promising performance, which could make it difficult to make highly accurate recognition. This paper proposes a novel approach to predict human activity from the Heterogeneity Human Activity Recognition (HHAR) dataset that integrates active learning with machine learning models: Random Forest (RF), Extreme Gradient Boosting (XGBoost), K-nearest Neighbors (KNN), Decision Tree (DT), Gradient Boosting (GB) and Light Gradient Boosting Machine (LGBM) classifier to predict heterogeneous activities accurately. We evaluated our approach to these models on the HHAR dataset that was generated using an accelerometer and gyroscope of smartwatches. The experiments are evaluated on 3 iterations where the results demonstrated that the proposed approach predicts human activities with the highest F1-Score of 99.99%. The results indicate that this approach is the most accurate and effective compared to the conventional approaches and baseline.

Artificial Intelligence (AI) has made tremendous progress in anomaly detection. However, AI models work as a black-box, making it challenging to provide reasoning behind their judgments in a Log Anomaly Detection (LAD). To the rescue, Explainable Artificial Intelligence (XAI) improves system log analysis. It follows a white-box model for transparency, understandability, trustworthiness, and dependability of Machine Learning (ML) and Deep Learning (DL) Models. In addition, Shapely Additive Explanation (SHAP), added to system dynamics, makes informed judgments and adoptable proactive methods to optimize system functionality and reliability. Therefore, this paper proposed the Shapely eXplainable Anomaly Detection (SXAD) framework to identify different events (features) that impact the models’ interpretability, trustworthiness, and explainability. The framework utilizes the Kernel SHAP approach, which is based on Shapley values principle, providing an innovative approach to event selection and identifying specific events causing abnormal behavior. This study addresses the LAD by transforming it from a black-box model into a white-box one, leveraging XAI to make it transparent, interpretable, explainable, and dependable. It utilizes benchmark data from the Hadoop Distributed File System (HDFS), organized using a Drain parser, and employs several ML models, such as Decision Tree (DT), Random Forest (RF), and Gradient Boosting (GB). These models achieve impressive accuracy rates of 99.99%, 99.85%, and 99.99%, respectively. Our contribution are novel because no earlier work has been done in the area of Log Anomaly Detection (LAD) with integration of XAI-SHAP.

The increasing occurrence of Hypertension highlights the need for advanced predictive tools in healthcare. This research proposes a novel approach that combines machine and deep learning for new feature generation and hypertension prediction. We explore machine learning-based models: Random Forest (RF), Logistic Regression (LR), Decision Tree (DT), XGBoost (XGB), and Gradient Boosting (GB) for new Feature Prediction (FP) and integrated these predictions with the original dataset for training deep Long short-term memory (LSTM) model. To evaluate the efficiency of the proposed approach, we compare all models, predicting new features, with those in the existing study. The results demonstrate that the GB-based FP + LSTM are standout performers. The GB-based FP + LSTM combination demonstrates the highest accuracy at 98.48%. On the contrary, the LR-based FP + LSTM combination exhibits a lower accuracy of 89.39%. The remaining combinations, including RF-based FP + LSTM, XGB-based FP + LSTM, and DT-based FP + LSTM, showcase accuracies ranging from 95.45% to 97.97%. In practical terms, the high F1-score of 98.48% is achieved by the combination of GB-based FP + LSTM, which implies a reliable tool for clinicians to aid in early hypertension detection. These findings hold deep practical implications, offering healthcare practitioners and policymakers a pathway to deploy accurate and timely hypertension identification tools.

In a smart home environment, multiple users can access a single smart device simultaneously. Moreover, these multiple users may have conflicting demands at a time; that is, one user’s demands differ from another for the same device based on the role of users and environmental factors. Therefore, existing single-user access control systems cannot handle such conflicting and dynamically changing demands, considering both roles and environmental factors in the multi-user smart home environment. Considering this issue, we proposed a Smart Access Control and Authorization framework (SAAC). It is a multi-user access control solution that has four modules, namely, a user interaction module, a backend server module, a policy manager module, and a policy execution module. The user interaction module collects user data and resource policies, which are processed by the backend server and forwarded to the policy manager. The policy manager resolves conflicts and generates final policies, which are stored in the backend server for enforcement by the policy execution module. The finalized policies are shared with the backend server module and saved there till needed for execution by the policy execution module to enforce the access control decision. We have implemented a proof of concept of the proposed framework on VS Code using the Casbin library. The performance evaluation results show our framework’s effectiveness and efficiency with lower computational complexity requirements than existing methods. Finally, we performed a security analysis of the proposed model based on the STRIDE model that confirms its robustness against access control attacks.

The widespread use of smartphones has brought convenience and connectivity to the fingertips of the masses. As a result, this has paved the way for potential security vulnerabilities concerning sensitive data, particularly by exploiting side-channel attacks. When typing on a smartphone’s keyboard, its vibrations can be misused to discern the entered characters, thus facilitating side-channel attacks. These smartphone hardware sensors can capture such information while users input sensitive data like personal details, names, email addresses, age, bank details and passwords. This study presents a novel Graph Neural Network (GNN) approach to predict side-channel attacks on smartphone keyboards; different GNN architectures were used, including GNN, DeepGraphNet, Gradient Boosting (GB)+DeepGraphNet, Extreme Gradient Boosting (XGB)+DeepGraphNet and K-Nearest Neighbor (KNN)+DeepGraphNet. The proposed approach detects the side channel attack using vibrations produced while typing on the smartphone soft keyboard. The data was collected from three smartphone sensors, an accelerometer, gyroscope, and magnetometer, and evaluated this data using common evaluation measures such as accuracy, precision, recall, F1-score, ROC curves, confusion matrix and accuracy and loss curves. This study demonstrated that GNN architectures can effectively capture complex relationships in data, making them well-suited for analyzing patterns in smartphone sensor data. Likewise, this research aims to fill a crucial gap by enhancing data privacy in the information entered through smartphone keyboards, shielding it from side-channel attacks by providing an accuracy of 98.26%. Subsequently, the primary objective of this study is to assess the effectiveness of GNN architectures in this precise context. Similarly, the GNN model exhibits compelling performance, achieving accuracy, precision, recall, and f1 score metrics that showcase the model’s effectiveness, with the highest values of 0.98, 0.98, 0.98, and 0.98, respectively. Significantly, the metrics mentioned in the study outperform those documented in the previous literature. Overall, the study contributes to the detection of side-channel smartphone attacks, which advances secure data practices.INDEX TERMS Graph neural networks (GNN), keystroke inference, motion sensors, machine learning, smartphone security, side-channel attacks.

Darknet refers to a significant portion of the internet that is hidden and not indexed by traditional search engines. It is often associated with illicit activities such as the trafficking of illicit goods, such as drugs, weapons, and stolen data. To keep our online cyber spaces safe in this era of rapid technological advancement and global connectivity, we should analyse and recognise darknet traffic. Beyond cybersecurity, this attention to detail includes safeguarding intellectual property, stopping illegal activity, and following the law. In order to improve accuracy and precision in identifying illicit activities, this study presents a novel approach named Active-Darknet that uses an active learning-based machine learning model for detecting darknet traffic. In order to guarantee high-quality analysis, our methodology includes extensive data preprocessing, such as numerically encoding categorical labels and improving the representation of minority classes using data balancing. In addition to machine learning models, we also use Deep Neural Networks (DNN), Bidirectional Long Short-Term Memory (BI-LSTM) and Flattened-DNN for experimentation. The majority of models exhibited encouraging outcomes; however, the models that utilised active learning, specifically the Random Forest (RF) and Decision Tree (DT) models, attained promising accuracy levels of 87%, rendering them the most efficient in detecting darknet traffic. Large traffic analysis is greatly enhanced by this method, which also increases the detection process’s robustness and effectiveness.

Adversarial attacks involve manipulating data to trick Artificial Intelligence (AI) models, making false predictions or classifications or even disrupting the normal functions of the smart grid. This can be done by providing the wrong information to the models, hence producing wrong predictions and therefore leading to instabilities, power imbalances and overall operational failure. Thus, employing an AI model is vital for controlling energy usage, identifying and even predicting equipment failures, and accurately determining power availability. Thus, their input dependency does not enable them to withstand cyber-attacks, which compromises the stability of the grid. These attacks affect energy control, cause losses and intrude into the critical infrastructure part of the smart grid, hence the call for enhanced smart grid protection. This study first generates a novel adversarial attack dataset on smart grids with three attacks: adversarial perturbation, backdoor injection, DOS attacks, and one benign class. Next, it provides a fine-tuned Deep Neural Network (DNN) model to significantly improve resistance against adversarial attacks on smart grids. Results from the various Machine Learning (ML) and DNN algorithms showed accuracy varying between 29.10% to 73.9%, with DNN recording the highest levels of accuracy. This shows how the approach can be utilised in the prediction of such attacks and how the grid can be protected and secured.

The statewide consumer transportation demand model analyzes consumers’ transportation needs and preferences within a particular state. It involves collecting and analyzing data on travel behavior, such as trip purpose, mode choice, and travel patterns, and using this information to create models that predict future travel demand. Naturalistic research, crash databases, and driving simulations have all contributed to our knowledge of how modifications to vehicle design affect road safety. This study proposes an approach named PODE that utilizes federated learning (FL) to train the deep neural network to predict the truck destination state, and in the context of origin-destination (OD) estimation, sensitive individual location information is preserved as the model is trained locally on each device. FL allows the training of our DL model across decentralized devices or servers without exchanging raw data. The primary components of this study are a customized deep neural network based on federated learning, with two clients and a server, and the key preprocessing procedures. We reduce the number of target labels from 51 to 11 for efficient learning. The proposed methodology employs two clients and one-server architecture, where the two clients train their local models using their respective data and send the model updates to the server. The server aggregates the updates and returns the global model to the clients. This architecture helps reduce the server’s computational burden and allows for distributed training. Results reveal that the PODE achieves an accuracy of 93.20% on the server side.

The robust development of the blockchain distributed ledger, the Internet of Things (IoT), and fog computing-enabled connected devices and nodes has changed our lifestyle nowadays. Due to this, the increased rate of device sales and utilization increases the demand for edge computing technology with collaborative procedures. However, there is a well-established paradigm designed to optimize various distinct quality-of-service requirements, including bandwidth, latency, transmission power, delay, duty cycle, throughput, response, and edge sense, and bring computation and data storage closer to the devices and edges, along with ledger security and privacy during transmission. In this article, we present a systematic review of blockchain Hyperledger enabling fog and edge computing, which integrates as an outsourcing computation over the serverless consortium network environment. The main objective of this article is to classify recently published articles and survey reports on the current status in the domain of edge distributed computing and outsourcing computation, such as fog and edge. In addition, we proposed a blockchain-Hyperledger Sawtooth-enabled serverless edge-based distributed outsourcing computation architecture. This theoretical architecture-based solution delivers robust data security in terms of integrity, transparency, provenance, and privacy-protected preservation in the immutable storage to store the outsourcing computational ledgers. This article also highlights the changes between the proposed taxonomy and the current system based on distinct parameters, such as system security and privacy. Finally, a few open research issues and limitations with promising future directions are listed for future research work.

Tuberculosis (TB) is an infectious disease caused by Mycobacterium. It primarily impacts the lungs but can also endanger other organs, such as the renal system, spine, and brain. When an infected individual sneezes, coughs, or speaks, the virus can spread through the air, which contributes to its high contagiousness. The goal is to enhance detection recognition with an X-ray image dataset. This paper proposed a novel approach, named the Tuberculosis Segmentation-Guided Diagnosis Model (TSSG-CNN) for Detecting Tuberculosis, using a combined semantic segmentation and adaptive convolutional neural network (CNN) architecture. The proposed approach is distinguished from most of the previously proposed approaches in that it uses the combination of a deep learning segmentation model with a follow-up classification model based on CNN layers to segment chest X-ray images more precisely as well as to improve the diagnosis of TB. It contrasts with other approaches like ILCM, which is optimized for sequential learning, and explainable AI approaches, which focus on explanations. Moreover, our model is beneficial for the simplified procedure of feature optimization from the perspectives of approach using the Mayfly Algorithm (MA). Other models, including simple CNN, Batch Normalized CNN (BN-CNN), and Dense CNN (DCNN), are also evaluated on this dataset to evaluate the effectiveness of the proposed approach. The performance of the TSSG-CNN model outperformed all the models with an impressive accuracy of 98.75% and an F1 score of 98.70%. The evaluation findings demonstrate how well the deep learning segmentation model works and the potential for further research. The results suggest that this is the most accurate strategy and highlight the potential of the TSSG-CNN Model as a useful technique for precise and early diagnosis of TB.

Thermal comfort is a crucial element of smart buildings that assists in improving, analyzing, and realizing intelligent structures. Energy consumption forecasts for such smart buildings are crucial owing to the intricate decision-making processes surrounding resource efficiency. Machine learning (ML) techniques are employed to estimate energy consumption. ML algorithms, however, require a large amount of data to be adequate. There may be privacy violations due to collecting this data. To tackle this problem, this study proposes a federated deep learning (FDL) architecture developed around a deep neural network (DNN) paradigm. The study employs the ASHRAE RP-884 standard dataset for experimentation and analysis, which is available to the general public. The data is normalized using the min-max normalization approach, and the Synthetic Minority Over-sampling Technique (SMOTE) is used to enhance the minority class’s interpretation. The DNN model is trained separately on the dataset after obtaining modifications from two clients. Each client assesses the data greatly to reduce the over-fitting impact. The test result demonstrates the efficiency of the proposed FDL by reaching 82.40% accuracy while securing the data.

The Internet of Things (IoT), considered an intriguing technology with substantial potential for tackling many societal concerns, has been developing into a significant component of the future. The foundation of IoT is the capacity to manipulate and track material objects over the Internet. The IoT network infrastructure is more vulnerable to attackers/hackers as additional features are accessible online. The complexity of cyberattacks has grown to pose a bigger threat to public and private sector organizations. They undermine Internet businesses, tarnish company branding, and restrict access to data and amenities. Enterprises and academics are contemplating using machine learning (ML) and deep learning (DL) for cyberattack avoidance because ML and DL show immense potential in several domains. Several DL teachings are implemented to extract various patterns from many annotated datasets. DL can be a helpful tool for detecting cyberattacks. Early network data segregation and detection thus become more essential than ever for mitigating cyberattacks. Numerous deep-learning model variants, including deep neural networks (DNNs), convolutional neural networks (CNNs), and recurrent neural networks (RNNs), are implemented in the study to detect cyberattacks on an assortment of network traffic streams. The Canadian Institute for Cybersecurity’s CICDIoT2023 dataset is utilized to test the efficacy of the proposed approach. The proposed method includes data preprocessing, robust scalar and label encoding techniques for categorical variables, and model prediction using deep learning models. The experimental results demonstrate that the RNN model achieved the highest accuracy of 96.56%. The test results indicate that the proposed approach is efficient compared to other methods for identifying cyberattacks in a realistic IoT environment.

Background Feature selection is a vital process in data mining and machine learning approaches by determining which characteristics, out of the available features, are most appropriate for categorization or knowledge representation. However, the challenging task is finding a chosen subset of elements from a given set of features to represent or extract knowledge from raw data. The number of features selected should be appropriately limited and substantial to prevent results from deviating from accuracy. When it comes to the computational time cost, feature selection is crucial. A feature selection model is put out in this study to address the feature selection issue concerning multimodal. Methods In this work, a novel optimization algorithm inspired by cuckoo birds’ behavior is the Binary Reinforced Cuckoo Search Algorithm (BRCSA). In addition, we applied the proposed BRCSA-based classification approach for multimodal feature selection. The proposed method aims to select the most relevant features from multiple modalities to improve the model’s classification performance. The BRCSA algorithm is used to optimize the feature selection process, and a binary encoding scheme is employed to represent the selected features. Results The experiments are conducted on several benchmark datasets, and the results are compared with other state-of-the-art feature selection methods to evaluate the effectiveness of the proposed method. The experimental results demonstrate that the proposed BRCSA-based approach outperforms other methods in terms of classification accuracy, indicating its potential applicability in real-world applications. In specific on accuracy of classification (average), the proposed algorithm outperforms the existing methods such as DGUFS with 32%, MBOICO with 24%, MBOLF with 29%, WOASAT 22%, BGSA with 28%, HGSA 39%, FS-BGSK 37%, FS-pBGSK 42%, and BSSA 40%.

The technology advancement is supported by additive manufacturing industries, especially 3D printing companies, that enable fast object prototyping and development in Industry 4.0. As 3D printed products are highly adopted in various fields, the final printed product must fulfill precise requirements without any defects. Therefore, an efficient framework that simultaneously learns and detects faults during the printing process is required. Unfortunately, most state-of-the-art studies utilize a centralized approach, which is inefficient for continuous model updates. This article presents a hierarchical federated transfer learning (HFTL) framework that employs edge, fog, and cloud concepts to enable an efficient model updating mechanism. The proposed HFTL uses a well-known DL model with a new classifier to significantly reduce the distributed training process while providing high detection and classification performance. Additionally, the fog server comprehensively exploits the data collection from several edge servers and performs local training. The extensive simulation results indicate that the performance of the proposed HFTL is more efficient, with 24% faster training time, effectively detects flaws in 3D printing products with 45% accuracy, and has a 59% F1-score improvement in non-IID data distribution compared to traditional FL architecture.

This research article presents an enhanced YOLOv8 model with an additional feature extraction layer integrated into the traditional YOLOv8 architecture to improve fault detection performance in smart additive manufacturing, specifically for FDM 3D printers. Hyperparameter optimization techniques are employed to ensure the model is trained with optimal input and batch size configurations. The findings demonstrate that the additional module successfully enhances the model’s performance in detecting faults during the FDM 3D printing process. The best results are achieved using the YOLOv8s model with an image input size of 640 and a batch size of 16, achieving a ${\mathrm {mAP}}_{val}$ (50-95) of 89.7%. Despite the increased complexity from additional layers, there is a favorable trade-off between performance and complexity. Furthermore, a testbed implementation is conducted to validate the model’s performance in a real-world setting, showing that the fault detection latency remains insignificant even with multiple Raspberry Pi clients. Overall, this research provides insights into improving fault detection in smart additive manufacturing and highlights the effectiveness of the proposed YOLOv8 model with additional extraction layers.

Over time, the amount of textual data has increased drastically, especially due to the publication of articles. As a consequence, there has been a rise in anonymous content. Research is being conducted to determine alternative methods for identifying unknown text authors. To this end, a system has to be developed to accurately determine the author of unknown texts, given a group of writing samples. Active Learning is utilized in this study because it iteratively selects the most informative samples to include in the training set, which enables a more precise and accurate authorship identification approach with fewer examples. Makes it useful for analyzing the rising amount of anonymous content and identifying unknown text authors. This study proposes a novel approach that utilizes active learning (AL) based machine models, namely Logistic Regression (AL-LR), Random Forest (AL-RF), XGboost (AL-XGB), and Multilayer Perceptron (AL-MLP) for authorship identification. The proposed approach extracts valuable characteristics of the writer using the Term Frequency-Inverse Document Frequency (TF-IDF). This study’s selected comprehensive dataset, “All the news,” is divided into three subsets: Article 1, Article 2, and Article 3. We have restricted the dataset’s scope and selected the top 50 authors for our experimentation. The experimental outcomes reveal that the proposed AL-XGB model achieves superior performance on Article 1 of the “All the news” dataset. Further, the AL-LR model performed well on Article 2, and the AL-MLP performed well on Article 3. The results suggest using the proposed approach for authorship identification.

The advancement of Industry 4.0 has necessitated the development of reliable predictive maintenance systems in production and manufacturing processes. This need is particularly pronounced in facilities that employ multiple 3D printers concurrently. To address this, the present study proposes the establishment of an Industrial Internet of Things (IIoT)-based industrial setup, incorporating a robust machine-to-machine network. Within this framework, the paper introduces an additive manufacturing plant equipped with IIoT capabilities, focusing on a predictive maintenance technique tailored for Fused Deposition Modeling (FDM) 3D printers. The authors have devised a failure prediction algorithm that leverages a data-driven methodology and a Deep Learning (DL) algorithm called Multi-Flow BiLSTM. The proposed Multi-Flow BiLSTM employs a multi-learning flow process augmented by a residual connection, facilitating a more comprehensive understanding of historical data patterns. Notably, the findings demonstrate that the proposed Multi-Flow BiLSTM achieves an impressive mean absolute error of 2.95, the lowest error rate compared to alternative methods while exhibiting a respectable $R^{2}$ value of 0.9121.By employing the proposed predictive maintenance technique, efficient and uninterrupted operation of 3D printers can be ensured, thereby enhancing the overall productivity of the manufacturing plant. Moreover, the study’s results offer valuable insights that can inform further research in developing predictive maintenance systems for industrial applications.

Drug-drug interaction (DDI) is a significant public health issue that accounts for 30% of unanticipated clinically hazardous medication events. The past decade has seen an evolution in informatics-based research for DDI signal identification. This paper aims to create an ensemble stacking machine learning (ML) approach capable of accurately predicting novel DDI hazard indicators. The stacking ensemble machine learning architecture for predicting the signals of drug-drug interactions is supported by one of the most reliable sources of pharmacological data, DrugBank. We scrap a large dataset that contains drug-related information, including drug types, drug names, and other aspects of drug indicators, and make it publicly available to the research community. The proposed approach includes data preprocessing, balancing through a random oversampling, label encoding and one hot encoding technique used for categorical variables, and model prediction using an ensemble stacking technique. The proposed ensemble approach uses the Gradient Boosting (GB) classifier, Adaboost, and Gaussian Naive Bayes (GNB) to classify the drug indication types. The experimental results demonstrate that the suggested approach outperforms traditional machine learning approaches regarding accuracy and efficiency. The stacking model obtained the highest accuracy of 99.0%. The test indicates that the proposed model is more effective than traditional methods for detecting drug-drug interaction signals.

The advancement of the communications system has resulted in the rise of the Internet of Things (IoT), which has increased the importance of cybersecurity research. IoT, which incorporates a range of devices into networks to offer complex and intelligent services, must maintain user privacy and deal with attacks such as spoofing, denial of service (DoS), jamming, and eavesdropping. Attacks change with time, and new ones develop every day. Numerous researchers look into IoT system attack models and evaluate machine, deep, and federated learning-based IoT security approaches. However, existing methods do not produce reliable and encouraging performance. Therefore, this study proposes a novel approach for leveraging federated learning to identify large attacks on IoT devices using the novel CIC_IoT 2023 dataset. The approach uses a federated deep neural network to achieve precise categorization. Before model training, the data was preprocessed using various data preparation techniques to guarantee the creation of a trustworthy dataset for categorization. The suggested approach involves feature normalization, data balancing, and model prediction utilizing federated learning. The experimental findings show that the proposed approach attained an exceptional accuracy of 99.00%, endorsing it for attack detection.

The increasing demand for mobile phones has resulted in abundant online reviews, making it challenging for consumers to make informed purchasing decisions. In this study, we propose Graph Neural Network (GNN) models to classify mobile phone ratings using Term Frequency-Inverse Document Frequency (TF-IDF) features. We collected a dataset of over 13,000 mobile phone evaluations from the Flipkart website. The proposed method includes data purification, balancing, feature extraction from the TF-IDF, and model prediction using deep learning models. The proposed approach utilized other models such as Deep Neural Network (DNN), Long Short-Term Memory (LSTM), and Bidirectional LSTM to compare other classifiers. The experiments’ outcomes demonstrate that the suggested model performs better than conventional deep learning methods regarding accuracy and efficiency. The GNN model achieved the best 99.0% accuracy rate. The proposed approach can help consumers make informed purchasing decisions and can be extended to other e-commerce platforms with large datasets of online reviews.

Pulmonary Fibrosis (PF) is an immedicable respiratory condition distinguished by permanent fibrotic alterations in the pulmonary tissue for which there is no cure. Hence, it is crucial to diagnose PF swiftly and precisely. The existing research on deep learning-based pulmonary fibrosis detection methods has limitations, including dataset sample sizes and a lack of standardization in data preprocessing and evaluation metrics. This study presents a comparative analysis of four vision transformers regarding their efficacy in accurately detecting and classifying patients with Pulmonary Fibrosis and their ability to localize abnormalities within Images obtained from Computerized Tomography (CT) scans. The dataset consisted of 13,486 samples selected out of 24647 from the Pulmonary Fibrosis dataset, which included both PF-positive CT and normal images that underwent preprocessing. The preprocessed images were divided into three sets: the training set, which accounted for 80% of the total pictures; the validation set, which comprised 10%; and the test set, which also consisted of 10%. The vision transformer models, including ViT, MobileViT2, ViTMSN, and BEiT were subjected to training and validation procedures, during which hyperparameters like the learning rate and batch size were fine-tuned. The overall performance of the optimized architectures has been assessed using various performance metrics to showcase the consistent performance of the fine-tuned model. Regarding performance, ViT has shown superior performance in validation and testing accuracy and loss minimization, specifically for CT images when trained at a single epoch with a tuned learning rate of 0.0001. The results were as follows: validation accuracy of 99.85%, testing accuracy of 100%, training loss of 0.0075, and validation loss of 0.0047. The experimental evaluation of the independently collected data gives empirical evidence that the optimized Vision Transformer (ViT) architecture exhibited superior performance compared to all other optimized architectures. It achieved a flawless score of 1.0 in various standard performance metrics, including Sensitivity, Specificity, Accuracy, F1-score, Precision, Recall, Mathew Correlation Coefficient (MCC), Precision-Recall Area under the Curve (AUC PR), Receiver Operating Characteristic and Area Under the Curve (ROC-AUC). Therefore, the optimized Vision Transformer (ViT) functions as a reliable diagnostic tool for the automated categorization of individuals with pulmonary fibrosis (PF) using chest computed tomography (CT) scans.

Given the increasing frequency of network attacks, there is an urgent need for more effective network security measures. While traditional approaches such as firewalls and data encryption have been implemented, there is still room for improvement in their effectiveness. To effectively address this concern, it is essential to integrate Artificial Intelligence (AI)-based solutions into historical methods. However, AI-driven approaches often encounter challenges, including lower detection rates and the complexity of feature engineering requirements. Finding solutions to overcome these hurdles is critical for enhancing the effectiveness of intrusion detection systems. This research paper introduces a deep learning-based approach for network intrusion detection to overcome these challenges. The proposed approach utilizes various classification algorithms, including the AutoEncoder (AE), Long-short-term-memory (LSTM), Multi-Layer Perceptron (MLP), Linear Support Vector Machine (L-SVM), Quantum Support Vector Machine (Q-SVM), Linear Discriminant Analysis (LDA), and Quadratic Discriminant Analysis (QDA). To validate the effectiveness of the proposed approach, three datasets, namely IOT23, CICIDS2017, and NSL KDD, are used for experimentation. The results demonstrate impressive accuracy, particularly with the LSTM algorithm, achieving a 97.7% accuracy rate on the NSL KDD dataset, 99% accuracy rate on the CICIDS2017 dataset, and 98.7% accuracy on the IOT23 dataset. These findings highlight the potential of deep learning algorithms in enhancing network intrusion detection. By providing network administrators with robust security measures for accurate and timely intrusion detection, the proposed approach contributes to network safety and helps mitigate the impact of network attacks.

High efficiency and safety are critical factors in ensuring the optimal performance and reliability of systems and equipment across various industries. Fault monitoring (FM) techniques play a pivotal role in this regard by continuously monitoring system performance and identifying the presence of faults or abnormalities. However, traditional FM methods face limitations in fully capturing the complex interactions within a system and providing real-time monitoring capabilities. To overcome these challenges, Digital Twin (DT) technology has emerged as a promising solution to enhance existing FM practices. By creating a virtual replica or digital copy of a physical equipment or system, DT offers the potential to revolutionize fault monitoring approaches. This paper aims to explore and discuss the diverse range of predictive methods utilized in DT and their implementations in FM across industries. Furthermore, it will showcase successful implementations of DT in FM across a wide array of industries, including manufacturing, energy, transportation, and healthcare. The utilization of DT in FM enables a comprehensive understanding of system behavior and performance by leveraging real-time data, advanced analytics, and machine learning algorithms. By integrating physical and virtual components, DT facilitates the monitoring and prediction of faults, providing valuable insights into the system’s health and enabling proactive maintenance and decision making.

Contemporary advancements in wearable equipment have generated interest in continuously observing stress utilizing various physiological indicators. Early stress detection can improve healthcare by lessening the negative effects of chronic stress. Machine learning (ML) methodologies have been modified for healthcare equipment to monitor user health situations utilizing sufficient user information. Nevertheless, more data are needed to make applying Artificial Intelligence (AI) methodologies in the medical field easier. This research aimed to detect stress using a stacking model based on machine learning algorithms using chest-based features from the Wearable Stress and Affect Detection (WESAD) dataset. We converted this natural dataset into a convenient format for the suggested model by performing data visualization and preprocessing using the RESP feature and feature analysis using the Z-score, SelectKBest feature, the Synthetic Minority Over-Sampling Technique (SMOTE), and normalization. The efficiency of the proposed model was estimated regarding accuracy, precision, recall, and F1-score. The experimental outcome illustrated the efficacy of the proposed stacking technique, achieving 0.99% accuracy. The results revealed that the proposed stacking methodology performed better than traditional methodologies and previous studies.

Additive manufacturing (AM) has emerged as a transformative technology for various industries, enabling the production of complex and customized parts. However, ensuring the quality and reliability of AM parts remains a critical challenge. Thus, image-based fault monitoring has gained significant attention as an efficient approach for detecting and classifying faults in AM processes. This paper presents a comprehensive survey of image-based fault monitoring in AM, focusing on recent developments and future directions. Specifically, the proponents garnered relevant papers from 2019 to 2023, gathering a total of 53 papers. This paper discusses the essential techniques, methodologies, and algorithms employed in image-based fault monitoring. Furthermore, recent developments are explored such as the use of novel image acquisition techniques, algorithms, and methods. In this paper, insights into future directions are provided, such as the need for more robust image processing algorithms, efficient data acquisition and analysis methods, standardized benchmarks and datasets, and more research in fault monitoring. By addressing these challenges and pursuing future directions, image-based fault monitoring in AM can be enhanced, improving quality control, process optimization, and overall manufacturing reliability.

With the most recent developments in wearable technology, the possibility of continually monitoring stress using various physiological factors has attracted much attention. By reducing the detrimental effects of chronic stress, early diagnosis of stress can enhance healthcare. Machine Learning (ML) models are trained for healthcare systems to track health status using adequate user data. Insufficient data is accessible, however, due to privacy concerns, making it challenging to use Artificial Intelligence (AI) models in the medical industry. This research aims to preserve the privacy of patient data while classifying wearable-based electrodermal activities. We propose a Federated Learning (FL) based approach using a Deep Neural Network (DNN) model. For experimentation, we use the Wearable Stress and Affect Detection (WESAD) dataset, which includes five data states: transient, baseline, stress, amusement, and meditation. We transform this raw dataset into a suitable form for the proposed methodology using the Synthetic Minority Oversampling Technique (SMOTE) and min-max normalization pre-processing methods. In the FL-based technique, the DNN algorithm is trained on the dataset individually after receiving model updates from two clients. To decrease the over-fitting effect, every client analyses the results three times. Accuracies, Precision, Recall, F1-scores, and Area Under the Receiver Operating Curve (AUROC) values are evaluated for each client. The experimental result shows the effectiveness of the federated learning-based technique on a DNN, reaching 86.82% accuracy while also providing privacy to the patient’s data. Using the FL-based DNN model over a WESAD dataset improves the detection accuracy compared to the previous studies while also providing the privacy of patient data.