Search

Computer scientists often publish their source code accompanying their publications, prominently using code repositories across various domains. Despite the concurrent existence of scholarly articles and their associated official code repositories, explicit references linking the two are often missing. Traditionally, identifying whether scholarly content and code repositories pertain to the same research project requires manual inspection, a time-consuming task. This paper proposes a deep learning-based algorithm for automatically matching scholarly articles with their corresponding official code repositories. Our findings indicate that the most common linking information includes the paper title and BibTeX entries, typically found in the repository’s readme document. In this study, we employed SPECTER for vector embedding of paper and repository metadata. Utilizing these embedding representations with the Light Gradient Boosting Machine (LGBM), our method achieved an F1 score of 0.94. Moreover, combining our best model with a rule-based approach improved performance by 5.31%. This study successfully delineates a connection between academic papers and associated official code repositories, minimizing reliance on explicit bibliographic information in repositories.

Recently, evolutionary computation (EC) has experienced significant advancements due to the integration of machine learning, distributed computing, and big data technologies. These developments have led to new research avenues in EC, such as distributed EC and surrogate-assisted EC. While these advancements have greatly enhanced the performance and applicability of EC, they have also raised concerns regarding privacy leakages, specifically the disclosure of optimal results and surrogate models. Consequently, the combination of evolutionary computation and privacy protection becomes an increasing necessity. However, a comprehensive exploration of privacy concerns in evolutionary computation is currently lacking, particularly in terms of identifying the object, motivation, position, and method of privacy protection. To address this gap, this paper aims to discuss three typical optimization paradigms, namely, centralized optimization, distributed optimization, and data-driven optimization, to characterize optimization modes of evolutionary computation and proposes BOOM (i.e., oBject, mOtivation, pOsition, and Method) to sort out privacy concerns related to evolutionary computation. In particular, the centralized optimization paradigm allows clients to outsource optimization problems to a centralized server and obtain optimization solutions from the server. The distributed optimization paradigm exploits the storage and computational power of distributed devices to solve optimization problems. On the other hand, the data-driven optimization paradigm utilizes historical data to address optimization problems without explicit objective functions. Within each of these paradigms, BOOM is used to characterize the object and motivation of privacy protection. Furthermore, this paper discuss the potential privacy-preserving technologies that strike a balance between optimization performance and privacy guarantees. Finally, this paper outlines several new research directions for privacy-preserving evolutionary computation. [ABSTRACT FROM AUTHOR]

The document is the table of contents for the August 2024 issue of the IEEE Transactions on Semiconductor Manufacturing. It includes a special section on the 2023 Advanced Semiconductor Manufacturing Conference, as well as regular issue papers on topics such as wafer-scale 3D integration, chemical uniformity impacts on electrodeposition, and thermal study of polymers for microelectronics. The document also includes letters and announcements, such as calls for papers and nominations for awards. [Extracted from the article]

Workers with demanding jobs are at risk of experiencing mental stress, leading to decreased performance, mental illness, and disrupted sleep. To detect elevated stress levels in the workplace, studies have explored stress measurement from physiological, psychological, and behavioral perspectives. This paper reviews the assessment methods and strategies for mitigating mental stress in the workplace and provides recommendations for early detection and mitigation of mental stress. Among the modalities, Electroencephalography (EEG), Electrocardiography (ECG) and Galvanic Skin Response (GSR) were found to be the most used in assessing mental stress in the workplace. Nevertheless, these modalities are sensitive to motion artifacts and are difficult to be integrated into real work environments. To further improve stress level assessment in the workplace, multimodality integration with a reduced number of sensors such as EEG, GSR and Functional near infrared spectroscopy (fNIRS) can be utilized. This would lead to developing strategies for stress management in real-time. Furthermore, combining EEG with fNIRS would improve source localization of mental stress. To mitigate stress, we recommend developing a closed loop system that incorporates brain data acquisition systems and machine learning with physical stimulations such as audio Binaural Beats Stimulation and/or Transcranial Electric Stimulation. [ABSTRACT FROM AUTHOR]

Recent years have witnessed significant advances brought by microfluidic biochips in automating biochemical protocols. Accurate preparation of fluid samples is an essential component of these protocols, where concentration prediction and generation are critical. Equipped with the advantages of convenient fabrication and control, microfluidic mixers demonstrate huge potential in sample preparation. Although finite element analysis (FEA) is the most commonly used simulation method for accurate concentration prediction of a given microfluidic mixer, it is time-consuming with poor scalability for large biochip sizes. Recently, machine learning models have been adopted in concentration prediction, with great potential in enhancing the efficiency over traditional FEA methods. However, the state-of-the-art machine learning-based method can only predict the concentration of mixers with fixed input flow rates and fixed sizes. In this paper, we propose a new concentration prediction method based on graph neural networks (GNNs), which can predict output concentrations for microfluidic mixters with variable input flow rates. Moreover, a transfer learning method is proposed to transfer the trained model to mixers of different sizes with reduced training data. Experimental results show that, for microfluidic mixers with fixed input flow rates, the proposed method obtains an average reduction of 88% in terms of prediction errors compared with the state-of-the-art method. For microfluidic mixers with variable input flow rates, the proposed method reduces the prediction error by 85% on average. Besides, the proposed transfer learning method reduces the training data by 84% for extending the pre-trained model for microfluidic mixers of different sizes with acceptable prediction error.

Three dimensional Not-And (3D NAND) flash memory devices are scaling in the vertical direction to more than 200 oxide/sacrificial wordline nitride layers to further increase storage capacity and enhance energy efficiency. The accurate measurement of the thicknesses of these layers is critical to controlling stress-induced wafer warping and pattern distortion. While traditional optical metrology in the UV-vis-NIR range offers a non-destructive inline solution for high volume manufacturing, we demonstrate in this paper, that mid-IR metrology has advantages in de-correlating oxide and nitride thicknesses owing to their unique absorption signatures. Furthermore, because of the depths sensitivity of oxide and nitride absorptions, the simulated measurement results show the ability to differentiate thickness variations in the vertical zones. Good blind test results were obtained with a machine learning model trained on pseudo-references and pseudo spectra with added skew. [ABSTRACT FROM AUTHOR]

Molecular retrosynthesis is a significant and complex problem in the field of chemistry, however, traditional manual synthesis methods not only need well-trained experts but also are time-consuming. With the development of Big Data and machine learning, artificial intelligence (AI) based retrosynthesis is attracting more attention and has become a valuable tool for molecular retrosynthesis. At present, Monte Carlo tree search is a mainstream search framework employed to address this problem. Nevertheless, its search efficiency is compromised by its large search space. Therefore, this paper proposes a novel approach for retrosynthetic route planning based on evolutionary optimization, marking the first use of Evolutionary Algorithm (EA) in the field of multi-step retrosynthesis. The proposed method involves modeling the retrosynthetic problem into an optimization problem, defining the search space and operators. Additionally, to improve the search efficiency, a parallel strategy is implemented. The new approach is applied to four case products and compared with Monte Carlo tree search. The experimental results show that, in comparison to the Monte Carlo tree search algorithm, EA significantly reduces the number of calling single-step model by an average of 53.9%. The time required to search three solutions decreases by an average of 83.9%, and the number of feasible search routes increases by 1.38 times. The source code is available at https://github.com/ilog-ecnu/EvoRRP. [ABSTRACT FROM AUTHOR]

Heart sound auscultation has been applied in clinical usage for early screening of cardiovascular diseases. Due to the high demand for auscultation expertise, automatic auscultation can help with auxiliary diagnosis and reduce the burden of training professional clinicians. Nevertheless, there is a limit to classic machine learning's performance improvement in the era of Big Data. Deep learning has outperformed classic machine learning in many research fields, as it employs more complex model architectures with a stronger capability of extracting effective representations. Moreover, it has been successfully applied to heart sound analysis in the past years. As most review works about heart sound analysis were carried out before 2017, the present survey is the first to work on a comprehensive overview to summarise papers on heart sound analysis with deep learning published in 2017–2022. This work introduces both classic machine learning and deep learning for comparison, and further offer insights about the advances and future research directions in deep learning for heart sound analysis. [ABSTRACT FROM AUTHOR]

Innovations in digital health and machine learning are changing the path of clinical health and care. People from different geographical locations and cultural backgrounds can benefit from the mobility of wearable devices and smartphones to monitor their health ubiquitously. This paper focuses on reviewing the digital health and machine learning technologies used in gestational diabetes – a subtype of diabetes that occurs during pregnancy. This paper reviews sensor technologies used in blood glucose monitoring devices, digital health innovations and machine learning models for gestational diabetes monitoring and management, in clinical and commercial settings, and discusses future directions. Despite one in six mothers having gestational diabetes, digital health applications were underdeveloped, especially the techniques that can be deployed in clinical practice. There is an urgent need to (1) develop clinically interpretable machine learning methods for patients with gestational diabetes, assisting health professionals with treatment, monitoring, and risk stratification before, during and after their pregnancies; (2) adapt and develop clinically-proven devices for patient self-management of health and well-being at home settings (“virtual ward” and virtual consultation), thereby improving clinical outcomes by facilitating timely intervention; and (3) ensure innovations are affordable and sustainable for all women with different socioeconomic backgrounds and clinical resources. [ABSTRACT FROM AUTHOR]

Training an accurate classifier for EEG-based brain-computer interface (BCI) requires EEG data from a large number of users, whereas protecting their data privacy is a critical consideration. Federated learning (FL) is a promising solution to this challenge. This paper proposes Federated classification with local Batch-specific batch normalization and Sharpness-aware minimization (FedBS) for privacy protection in EEG-based motor imagery (MI) classification. FedBS utilizes local batch-specific batch normalization to reduce data discrepancies among different clients, and sharpness-aware minimization optimizer in local training to improve model generalization. Experiments on three public MI datasets using three popular deep learning models demonstrated that FedBS outperformed six state-of-the-art FL approaches. Remarkably, it also outperformed centralized training, which does not consider privacy protection at all. In summary, FedBS protects user EEG data privacy, enabling multiple BCI users to participate in large-scale machine learning model training, which in turn improves the BCI decoding accuracy.

Advancements in machine learning offer promising avenues for the identification of ADHD symptoms in adults, an endeavour traditionally encumbered by the intricacies of human behavioural patterns. In this paper, we introduce three innovative dual-stream models. The proposed approach utilises a novel multi-modal dataset recorded for ADHD symptoms detection, leveraging RGB video alongside facial, body posture and hand landmark data. The fusion of these different sub-modalities within video enhances the discriminative capability of the ADHD symptoms detection system. A primary objective was to maintain minimal model depth while achieving competitive performance. Through randomised search cross-validation and a rigorous leave-one-out validation scheme, the proposed model achieves high generalisability and robust symptom identification, suggesting strong potential for application in clinical environments. Evaluation boasts the state-of-the-art performance of the proposed model, demonstrating an accuracy of 98.67%, a precision of 98.01%, and a recall of 98.88%. These metrics attest to the model's ability to consistently identify ADHD symptoms while maintaining a minimal parameter footprint. This delicate balance provides a significant step forward in behavioural health analytics.

Robot-assisted motor training is applied for neurorehabilitation in stroke patients, using motor imagery (MI) as a representative paradigm of brain-computer interfaces to offer real-life assistance to individuals facing movement challenges. However, the effectiveness of training with MI may vary depending on the location of the stroke lesion, which should be considered. This paper introduces a multi-task electroencephalogram-based heterogeneous ensemble learning (MEEG-HEL) specifically designed for cross-subject training. In the proposed framework, common spatial patterns were used for feature extraction, and the features according to stroke lesions are shared and selected through sequential forward floating selection. The heterogeneous ensembles were used as classifiers. Nine patients with chronic ischemic stroke participated, engaging in MI and motor execution (ME) paradigms involving finger tapping. The classification criteria for the multi-task were established in two ways, taking into account the characteristics of stroke patients. In the cross-subject session, the first involved a direction recognition task for two-handed classification, achieving a performance of 0.7419 (±0.0811) in MI and 0.7061 (±0.1270) in ME. The second task focused on motor assessment for lesion location, resulting in a performance of 0.7457 (±0.1317) in MI and 0.6791 (±0.1253) in ME. Comparing the specific-subject session, except for ME on the motor assessment task, performance on both tasks was significantly higher than the cross-subject session. Furthermore, classification performance was similar to or statistically higher in cross-subject sessions compared to baseline models. The proposed MEEG-HEL holds promise in improving the practicality of neurorehabilitation in clinical settings and facilitating the detection of lesions.

Various measures have been proposed to quantify upper-limb use through wrist-worn inertial measurement units. The two most popular traditional measures of upper-limb use - thresholded activity counts (TAC) and the gross movement (GM) score suffer from high sensitivity and low specificity, and vice versa. We previously proposed a hybrid version of these two measures - the GMAC - that showed better overall detection performance than TAC and GM. In this paper, we answer two critical questions to improve the GMAC measure's usefulness: (a) can it be implemented using only the accelerometer data? (b) what are its optimal parameter values? Here, we propose a modified GMAC using only the accelerometer data and optimize its parameters to develop: (a) a generic measure that is both limb- and subject-independent, and (b) limb-specific measures that were only subject-independent. The optimized GMAC showed better detection performance than the previous GMAC and surprisingly had comparable performance to the best-performing machine learning-based measure (random forest inter-subject model). In hemiparetic data, its performance was similar to the previous GMAC and the random forest inter-subject model; the limb-specific GMAC measure, however, had a better performance than the generic measure. The optimized limb-specific GMAC is a simple, interpretable alternative to a machine learning-based inter-subject model. The optimized GMAC can be a valuable measure for offline or real-time detection and feedback of upper limb use. The preliminary results of this study, based on a small dataset, need to be validated on a larger dataset to evaluate its generalizability.

The steady-state visual evoked potential (SSVEP) has become one of the most prominent BCI paradigms with high information transfer rate, and has been widely applied in rehabilitation and assistive applications. This paper proposes a least-square (LS) unified framework to summarize the correlation analysis (CA)-based SSVEP spatial filtering methods from a machine learning perspective. Within this framework, the commonalities and differences between various spatial filtering methods appear apparent, the interpretation of computational factors becomes intuitive, and spatial filters can be determined by solving a generalized optimization problem with non-linear and regularization items. Moreover, the proposed LS framework provides the foundation of utilizing the knowledge behind these spatial filtering methods in further classification/regression model designs. Through a comparative analysis of existing representative spatial filtering methods, recommendations are made for the superior and robust design strategies. These recommended strategies are further integrated to fill the research gaps and demonstrate the ability of the proposed LS framework to promote algorithmic improvements, resulting in five new spatial filtering methods. This study could offer significant insights in understanding the relationships between various design strategies in the spatial filtering methods from the machine learning perspective, and would also contribute to the development of the SSVEP recognition methods with high performance.

This study presents a novel method to assess the learning effectiveness using Electroencephalography (EEG)-based deep learning model. It is difficult to assess the learning effectiveness of professional courses in cultivating students' ability objectively by questionnaire or other assessment methods. Research in the field of brain has shown that innovation ability can be reflected from cognitive ability which can be embodied by EEG signal features. Three navigation tasks of increasing cognitive difficulty were designed and a total of 41 subjects participated in the experiment. For the classification and tracking of the subjects' EEG signals, a convolutional neural network (CNN)-based Multi-Time Scale Spatiotemporal Compound Model (MTSC) is proposed in this paper to extract and classify the features of the subjects' EEG signals. Furthermore, Spiking neural networks (SNN) -based NeuCube is used to assess the learning effectiveness and demonstrate cognitive processes, acknowledging that NeuCube is an excellent method to display the spatiotemporal differences between spikes emitted by neurons. The results of the classification experiment show that the cognitive training traces of different students in solving three navigational problems can be effectively distinguished. More importantly, new information about navigation is revealed through the analysis of feature vector visualization and model dynamics. This work provides a foundation for future research on cognitive navigation and the training of students' navigational skills.

Thanks to the many submissions we have received, we can present this second part of our special issue on "AI-eXplained." In here, we continue our mission to demystify the intricate world of artificial intelligence and make it accessible to a broader audience. As AI continues to evolve and integrate into various aspects of our lives, it becomes increasingly important to bridge the gap between experts and those eager to understand the inner workings of AI systems. Our goal remains unchanged—to present AI concepts in a comprehensible, engaging, and interactive manner, empowering our readers to explore and grasp the enchanting world of AI. [ABSTRACT FROM AUTHOR]

As we enter a New Year, we can look back on another year of solid accomplishment at IEEE Transactions on Semiconductor Manufacturing. I am happy to report that our impact factor remains steady at 2.70, and our mean time to first decision remains competitive at 8.3 weeks. Our Editorial Board remains as strong as ever, with the addition of Dr. Jun-Haeng Lee in the area of machine learning and data science applications in 2023, and we are actively seeking new board members. Our submissions remain strong, as do the special sections from conferences (ASMC, ISSM and CS-MANTECH). The Special Issue on Production-Level Artificial Intelligence Applications in Semiconductor Manufacturing appeared in the November issue, and two additional special issues are in preparation. Prof. Duane Boning of MIT and Dr. Bill Nehrer of Technology Consultancy are co-editing a special issue on “Semiconductor Design for Manufacturing,” which will be a collaborative effort with the IEEE Transactions on Electron Devices. Drs. Oliver Patterson of Intel and Tomasz Brozek of PDF Solutions are also co-editing a special issue on sustainable semiconductor manufacturing. We are also happy to announce the Best paper Award for 2023, in the companion editorial appearing in this issue. Congratulations to all the honorees, and we hope we will continue to see their submissions in the future. Our thanks go to Drs. Jeanne Bickford, Dragan Djurdjanovic and Mahadeva Iyer Natarajan for their work on this committee. [ABSTRACT FROM AUTHOR]

This paper presents an adaptive pre-silicon integrated circuit verification algorithm that incorporates a machine learning algorithm. This approach can overcome the traditional corner-based process-voltage-temperature verification limitations considering variables coverage. The proposed solution employs an adaptive algorithm whose parameters have been analyzed in detail and optimized to achieve high performance. The optimal configurations for these parameters were demonstrated through the accuracy and efficiency of the algorithm. Additionally, this paper presents three methods that expand the use of the univariate algorithm to multivariate adaptive scenarios. These techniques aim to achieve the most accurate identification of worst-case circuit behavior through simultaneously modeling multiple electrical parameters (EP) outputs. The effectiveness of the proposed methods was validated through extensive testing on a large and diverse set of synthetic test functions that intend to replicate the behavior of real circuits. The algorithms consistency and accuracy are also validated on real Low Dropout Voltage Regulator (LDO) circuits.

Natural languages are written and spoken languages, and NLP (Natural Language Processing) is the ability of a computer program to recognize both written and spoken languages. Word Sense Disambiguation (WSD) is identified as a challenging area of research in Artificial Intelligence (AI), and Machine Translation (MT). WSD is the procedure for selecting the exact meaning of a word that has more than one meaning. This is an essential application for all-natural language processing applications. There are various knowledge-based, supervised, and unsupervised approaches to WSD process. The Naïve Bayes classifier as an example of approach supervised and unsupervised approaches is the most important method. In this paper, we emphasize on the use of the Naïve Bayes approach for text classification in WSD techniques. Bayes’ hypothesis is a probabilistic model and a reliable approach for text classification. Bayes’ hypothesis acknowledges that the occurrence of some other features is not dependent on the presence of a particular element in a class. This calculation can be used to solve multi-class prediction problems. This classifier performs better compared to the other methods of different approaches. This paper gives an itemized investigation of Naïve Bayes algorithms, which depicts its ideas, covering up Naïve Bayes,’ text characterization, traditional innocent Bayes,’ and machine learning. We have used the collocation method of feature extraction for the WSD of English sentences. Using this model we have disambiguate ambiguous English words by predicting part-of-speech inclusive of “noun,” “verb,” “adverb,” and “adjective.” This disambiguation module is an enhancement in machine translation. The system reported the performance measure of seventy-eight (78%) percent of the scale on F1-measure.

The increasing complexity of software development demands efficient automated bug report priority classification, and recent advancements in deep learning hold promise. This paper presents a comparative study of contemporary learning paradigms, including BERT, vector databases, large language models (LLMs), and a simple novel learning paradigm, contrastive learning for BERT. Utilizing datasets from bug reports, movie reviews, and app reviews, we evaluate and compare the performance of each approach. We find that transformer encoder-only models outperform in classification tasks measured by the precision, recall, and F1 score transformer decoder-only models despite an order of magnitude gap between the number of parameters. The novel use of contrastive learning for BERT demonstrates promising results in capturing subtle nuances in text data. This work highlights the potential of advanced NLP techniques for automated bug report priority classification and underscores the importance of considering multiple factors when developing models for this task. The paper’s main contributions are a comprehensive evaluation of various learning paradigms, such as vector databases and LLMs, an introduction of contrastive learning for BERT, an exploration of applicability to other text classification tasks, and a contrastive learning procedure that exploits ordinal information between classes.

Health index is an essential tool to evaluate the condition of power transformers. Generally, there are two methodologies to evaluate the health index of power transformers: conventional methods and machine learning-based approaches. The main challenge in employing the health index to assess transformer’s condition is the lack of supporting data, which hinders its ability to accurately reflect the actual health condition of the transformers. While several studies on estimating the transformers health index using machine learning techniques can be found in the literatures, not much attention was given to the cumulative impact of multiple data unavailability on the calculation’s accuracy. This paper presents a comprehensive analysis of health index predictions using two machine learning methods: specifically, regression and classification against conventional methods of weighting and scoring. Furthermore, the paper presents some strategies to overcome the unavailability of multiple data. Six methods are presented and evaluated across 15 missing data scenarios to assess their effectiveness compared with scenarios featuring complete data. Results reveal that Gradient Boosting Regression has the most accurate accuracy in predicting the health index with a 95%, assuming complete data, and 97.87% considering data unavailability. The paper also presents an economic analysis to highlight the profitability associated with predicting missing data compared with the certainty level and evaluation using asset owner interpretation to assess the feasibility of the proposed method. The main contribution of the paper is the presentation of a comprehensive view for asset owners to make informed decisions in case of data unavailability.

The transportation sector is one among the key sources of greenhouse gas emissions (GHGs) leading to climate change and global warming. Energy transition through electrified transportation is one of the solutions to tackle the issues. Electric vehicles (EVs) offer significant environmental and economic advantages against the conventional Internal Combustion Engine (ICE) vehicles. EVs are called mobility loads and their connectivity to the utility grid for charging is unpredictable. The large penetration of such unpredictable loads into the utility grid will lead to undesirable impacts on the utility service. This paper highlights the importance of managing and optimizing the charging schedules. The optimization of EV charging has diverse aspects, and the perspectives of EV charging differ among consumers, aggregators, and utility services. Proper planning and management of EV charging is essential to achieve harmony amongst these stakeholders. A comprehensive review on the objectives of electric vehicle charging optimization from various perspectives is presented and discussed in this paper. EV charging optimization techniques including mathematical programming, meta heuristics algorithms and machine learning techniques are explored. The main objectives, constraints, strength, and limitations of different charging optimization techniques are analyzed in detail. A brief discussion on the communication strategies for data exchange in EV charging framework is presented and the need for a communication security constrained EV charging scheduling is also emphasized.

Smart farming, popularly called precision agriculture, refers to technologies like the Internet of Things (IoT), Artificial Intelligence (AI), and drones that are rapidly transforming age-old farming traditions. This paper investigates how smart farming technologies are revolutionizing the practice of agriculture, specifically focusing on monitoring and control solutions of IoT devices and LoRaWAN networked. Therefore, the implementation of IoT devices would give the farmer access to current readings when needed for making decisions and optimizing their resources on the most important parameters for agricultural activity. And this is where IoT devices come in, making possible precision agriculture techniques with lowered costs by customizing farming practice for each crop or section of land. It means that even the most remote farms will benefit from monitoring and control solutions within LoRaWAN networks. Moreover, low power with long-range wireless connectivity technology devised by LoRaWAN ensures communication and data analysis to be a reliable one within the system. It introduces an intelligent LoRaWAN-based IoT device for the monitoring and controlling solutions with the parameters of performance evaluation, experimental setup, and the dataset used for its analysis. This paper elaborates on anomaly detection through Isolation Forest and proves it for the identification of anomalies in data related to temperature and humidity. Predominantly, the study also indicates precision in the temperature variation prediction model through the use of the predictive model based on the linear regression and random forest algorithms. This improves smart farming practices developed for precision agriculture in terms of efficiency, productivity, and sustainability.

Origin-Destination (OD) traffic flow estimations from traffic sensor data play an important role for transportation planning and management. This paper proposes a novel method to compare OD traffic estimated matrices (using data from traffic sensors). The proposed method uses the estimated OD traffic flow values together with COVID-19 incidence data in order to build a sequence of temporal graphs that are fed into a machine learning (ML) model. The ML model uses the input information to estimate/predict one-week ahead COVID-19 incidence data. A tailored Graph Neural Network (GNN) and Long Short-Term Memory (LSTM) model is designed adapted to the input information. The paper evaluates the proposed method with 3 different OD estimation alternatives and compares the accuracy achieved by different configurations of the ML model with a traffic agnostic baseline model. Data from 44 provinces in Spain during 2021 providing daily COVID-19 incidence data and 635 geo-located traffic sensors providing monthly traffic counts are used to evaluate the results. The 3 traffic-aware OD estimation methods were able to outperform the baseline model, achieving model gains up to 136%. The major application of the results of this paper is a novel mechanism to validate prior OD traffic matrices.

This paper addresses the hypothesis that the preference of a worker toward a work spot depends on attribute information such as her or his age and position and the type of work task on which she or he is working on that day. The system proposed in this paper is designed to estimate work-spot preferences using machine learning (ML) for activity-based workplaces. The system collects a variety of items related to attribute information and work-spot preferences for each work task from individual workers in advance. It constructs ML models using the collected data and estimates the work-spot preference for a worker. We extend the system to improve the accuracy for the case where the estimation is performed with a limited number of samples for training data. To examine the system, this paper presents two types of experiments: with computer graphics (CG) and with a real environment. Through these experiments, we collect data from people and construct ML models for the estimation of work-spot preferences based on attribute information and types of work tasks. The results demonstrate that the proposed system works for estimating the work-spot preference for each individual worker in accordance with the attribute information and the type of work tasks.

Soaring energy demand in power sector demands an intelligent electricity grid which is able to meet the energy requirements of consumers with higher efficiency and reliability. This paper discusses about how Smart Grid (SG) is the solution for such requirements in the energy sector. SG constitutes of three main components: Information Technology (IT), Operational Technology (OT), and Advanced Metering Infrastructure (AMI). It is important to synchronize these three components in order to seamlessly deliver electricity to the consumers connected to the electric network. Cyberattack is increasing sharply in the SG infrastructure and triggers disrupting or malicious manipulation of entire electricity network. It is crucial to maintain the data traffic, power transmission and distribution with high reliability. This paper aims at arriving at the taxonomy of attacks that are subjected to IT, OT and AMI components of SG. This exhaustive taxonomy further opens avenues to evolve mitigation strategies for these attacks which will eventually aid in SG protection. Elaborate discussion about various mitigation strategies for various types of attacks is performed. The systematic approach of methods and techniques along with the type of solution it provides to SG applications is discussed. It is evident from literature that Machine Learning (ML), Deep Learning (DL) and signal processing techniques are very efficient and robust in attack detection and retaliating in SG. ML model pro. Hence the importance of framework and tools in SG environment is analyzed. A case study is also presented that emphasizes on SG Attacks on the specific system and its mitigation strategies.

Alzheimer’s disease is a progressive disease for which researchers have yet to discover the main cause, but believe it probably involves a combination of age-related changes in the brain, genetic, environmental and lifestyle factors. Alzheimer’s is an irreversible disease that still has no cure. Therefore, its early diagnosis is very important to prevent its progression. Developing Machine Learning algorithms in healthcare, especially in brain disorders such as Alzheimer’s disease, provides new opportunities for early diagnosis and recognition of important biomarkers. This paper presents an overview of advanced studies based on Machine Learning techniques for diagnosing Alzheimer’s disease and different stages of mild cognitive impairment based on magnetic resonance imaging (MRI) images in the last 10 years. Also, this paper comprehensively describes the commonly efficient Machine Learning algorithms in each stage of magnetic resonance imaging processing used in the papers, which can facilitate the comparison of algorithms with each other and provide insight into the impact of each technique on classification performance. This review can be a valuable resource to gain a new perspective on the various research methods used in recent studies on Alzheimer’s disease.

The overall network volume in backbone optical networks is constantly growing, composed of many smaller network services with increasing trends and various seasonality. Due to the recent advances in machine learning algorithms, short- and long-term traffic fluctuations can be forecasted. Consequently, the backbone optical network can be adapted to traffic changes aiming to improve its performance. However, an important challenge lies in developing effective optimization methods capable of adapting to traffic changes to leverage the knowledge about the traffic. To this end, this paper addresses the time-varying traffic in spectrally-spatially flexible optical networks (SS-FONs), which are a promising technology to mitigate backbone network requirements of vast traffic volume transmission. The main contribution of this paper is twofold. Firstly, we introduce a new traffic prediction method using multioutput regression and temporal features to forecast traffic between all node pairs and integrate this prediction method into an optimization framework developed for dynamic resource allocation in translucent SS-FONs with time-varying traffic. Secondly, we evaluate potential network performance improvements from periodic lightpath reallocation through extensive numerical experiments. According to the results of experiments run on two representative optical network topologies, the proposed approach with periodic resource allocation allows achieving up to 7.8 percentage points reduction of bandwidth blocking compared to the reference scenario without reallocation. Consequently, the network requires up to 23.4% fewer transceivers to deliver the same traffic in considered scenarios and thus the power consumption savings are provided.

Precise forecasting of solar power output is crucial for integrating renewable energy into power networks, improving efficiency and dependability. This study assesses the efficacy of several Machine Learning (ML) algorithms in predicting solar power generation through a detailed performance comparison. This paper analyzes six algorithms: CatBoost, Gradient Boosting Machines (GBMs), Multilayer Perceptron (MLP) regressor, Support Vector Machines (SVMs), XGBoost, and Random Forest (RF). Using a dataset of 4213 sets of solar power generation data, each model was trained and tested, with performance evaluated based on R-squared (R2) scores for the whole dataset, training set, and test set. Also, this study examined the mean and standard deviation of test set predictions to gauge how consistent each model was. The results showed that RF had the highest overall R2 score of 0.940 and a training set score of 0.971. XGBoost demonstrated exceptional performance on the test set, attaining a high R2 score of 0.822. CatBoost and GBMs exhibited strong performance, albeit with slightly lower R2 values of 0.786 and 0.829, respectively. Although the MLP regressor and SVMs exhibited high training scores, they encountered difficulties in generalizing to unfamiliar data. This paper highlights the effectiveness of combining XGBoost and RF techniques in improving the accuracy of solar power forecasts. The investigation focuses on enhancing the precision and reliability of renewable energy projections through a comprehensive comparison of various contemporary ML techniques.

The security of Industrial Supply Chain (ISC) has emerged through the integration of Industrial Internet of Things (IIoT) and Blockchain (BC) technology. This new era involves effectively protecting IIoT systems from various threats and ensuring their smooth operation and resilience against potential cyber-attacks. Within the ISC ecosystem, combining machine learning (ML)-based security models for cyber-attack detection can play a crucial role in enhancing the ISC security and proactively identifying potential threats. This paper presents a BC-enabled ISC that embed ML security model integrated within a multi-layered approach. We conducted a comparative study and performance analysis of several ML classification techniques, with a focus on supervised methods to identify the lightweight model for cyber-attack detection suitable for deployment in resource-constrained IIoT environment. We investigate the performance of Gaussian Naive Bayes (NB), K-Nearest Neighbors (KNN), Random Forest (RF), Decision Tree (DT), and three ensemble techniques, namely Bagging, Stacking, and Boosting. The study employs the WUSTL-IIOT-2021 imbalanced dataset, which contains samples representing four types of attacks, including denial of service (DoS), SQL injection, reconnaissance, and backdoor. The paper addresses the imbalance in class representation by customizing the dataset for training and testing the ML models. Both Mutual Information (MI) and Extra-trees (ET) are applied as a one-stage ensemble feature selection. The performance of the ML models are investigated using classification accuracy (Acc), precision, recall, F1 score, Matthews correlation coefficient (MCC), model size (Mem), training time (TT) and prediction time (PT).

With increasing realised traffic on transport networks, greenhouse gas emissions show a similar trend. Reducing them is a modern aspiration, creating a better place to live and moving towards sustainability. Expanding the infrastructure is often not an appropriate solution, as the system would only be fully utilised at peak times, while at less frequent times it would not even approach capacity and would require huge investment costs. An alternative to further construction work is the implementation of intelligent traffic systems, where smoother flows can achieve higher capacity by reducing the variability in the system. In a motorway environment, a common approach is Variable Speed Limit Control, where the road is divided into zones and individual speed limits are used to increase or decrease the load on the cells. This paper proposes a solution in which individual cells make decisions cooperatively, in contrast to classical state machine-based methods. Thanks to the jointly formulated goal of the agents, a predictive control method is created that leads to a reduction in emissions due to avoided shock waves and reduced waiting times. This paper presents a solution that provides a universal solution across multiple application lengths, illustrating the power of deep learning. https://github.com/istvan-knab/Variable-speed-limit-control.

The convergence of Information Technology (IT), Operational Technology (OT), and Educational Technology (ET) has led to the emergence of the fourth industrial revolution. As a result, a new concept has emerged known as Digital Twins (DT), which is defined as “a virtual representation of various objects or systems that receive data from physical objects/systems to make changes and corrections”. In the aviation industry, numerous attempts have been made to utilize DT in the design, manufacturing, and condition monitoring of aircraft fleets. Among these research efforts, real-time, accurate, fast, and predictive condition monitoring methods play a crucial role in ensuring the safe and efficient performance of aircraft. Using DT for condition and fleet monitoring not only enhances the reliability and safety of aircraft but also reduces operational and maintenance costs. In this paper, the conducted studies on the applications of DT systems for condition monitoring of aircraft units and the aerospace sector are discussed and reviewed. The aim of this review paper is to analyse the current developments of DT systems in the aviation industry as well as explain the remaining challenges of DT systems. Then Finally, future trends of DT systems along with aircraft are presented. Among reviewed papers, most of them have used computational fluid dynamics, finite element methods, and artificial intelligence techniques for developing DT models for aircraft. At the same time, most of these analyses are dedicated to the failure and crack detection body of aircraft as well as engine fault detection. Life prediction is another popular application for using DT in aircraft units that could help the engineers predict the maintenance required for different parts of the aircraft. Finally, the application of DT in marine, power systems, and space programs has been also reviewed and the lessons learned from them have been translated to the aviation sector.

This paper proposes a novel method to predict the fresh weight of lettuce at the shipping stage in a plant factory using the early-stage growth images. It is well-established that the size and shape of plants correlate with their fresh weight. The proposed method captures chlorophyll fluorescence-based growth images daily and extracts geometric features such as projection area, edge length, and skeleton length. We design a regression model to predict the fresh weight using the dimensionality-reduced historical features. However, without considering growth statuses, the dimensionality reduction approach leads to decreased predictive performance for mature and slower-growing plants. In this paper, we generate a plant growth model that simulates the growth process by integrating multiple growth records based on the comparison of growth statuses. The proposed method then reduces the dimensionality by fitting historical features to the plant growth model to obtain future features. Experimental results demonstrate that the proposed method accurately predicts the fresh weight and achieves the coefficient of determination of 0.885, root mean square error of 8.790 g, and mean absolute error of 6.684 g when predicting the fresh weight ten days ahead using growth images from the past ten days.

The usage of artificial intelligence and machine learning in wireless communications is the stepping stone towards a technological breakthrough in the current limitations of wireless communication systems. The trend of future coding schemes towards 6G appears to be based on rateless schemes and machine learning. Channel coding is important when transmitting data or information reliably as it provides error-correcting purposes. However, there is still a demand for more research regarding machine learning for channel coding. There is also a lack of a specific term or classification for existing machine learning applications for channel coding. This paper explores and compiles current trending machine learning techniques for channel coding. We are also introducing and proposing a new type of machine learning classification for channel coding purposes, as well as surveying some of the papers that fall under the respective class. This paper also discusses current challenges and future machine learning trends for channel coding, which are expected to impact future wireless communications development, especially in channel coding advancements.

Situation Awareness (SA) is a process of sensing, understanding and predicting the environment and is an important component in complex systems. The reception of information from the environment tends to be continuous and of a multimodal nature. AI technologies provide a more efficient and robust support by subdividing the different stages of SA objectives into tasks such as data fusion, representation, classification, and prediction. This paper provides an overview of AI and multimodal methods used to build, enhance and evaluate SA in a variety of environments and applications. Emphasis is placed on enhancing perceptual integrity and persistence. Research indicates that the integration of artificial intelligence and multimodal approaches has significantly enhanced perception and comprehension in complex systems. However, there remains a research gap in projecting future situations and effectively fusing multimodal information. This paper summarizes some of the use cases and lessons learned where AI and multimodal techniques have been used to deliver SA. Future perspectives and challenges are proposed, including more comprehensive predictions, greater interpretability, and more advanced visual information.

The rapid evolution of cryptocurrency markets and the increasing complexity of trading strategies necessitate a comprehensive understanding of price-prediction models and their direct impact on trading efficacy. While extensive research has been conducted separately on price prediction methods and trading strategies, there remains a significant gap in studies explicitly correlating precise price forecasts with successful trading outcomes. This review paper addresses this gap by critically examining the role of accurate cryptocurrency price predictions in enhancing trading strategies. We conducted a systematic review of sufficient scholarly articles and web resources, focusing on the methodologies and effectiveness of various predictive models and their integration into cryptocurrency trading strategies. Our selection criteria ensured the inclusion of papers that demonstrate methodological rigor, relevance, and recent contributions to the field, spanning from economic theories and statistical models to advanced machine learning techniques. The findings reveal that precise price predictions significantly contribute to the development of adaptive and risk-managed trading strategies, which are crucial in the highly volatile cryptocurrency market. The review also identifies current challenges and proposes directions for future research, emphasizing the need for interdisciplinary approaches and ethical considerations in predictive modeling. This synthesis aims to bridge the existing research gap and guide future studies, thereby fostering more sophisticated and profitable trading strategies in the cryptocurrency domain.

This paper proposes an accurate harmonic identification strategy for microgrids and distributed power systems. The harmonic identification strategy is one of the complex tasks in microgrids due to the need of high computational burden in terms of memory and computational time. The complexity of the considered problem is resulted from solving the transcendental nonlinear equations that characterize harmonics especially in real time is considered as a highly challenged problem. The proposed identification strategy aims at detecting individual and total harmonic distortion levels that is generated from several harmonic sources. In the current paper, the Machine Learning Regression Analysis (MLRA) including location-specific data and the Artificial Neural Networks (ANNs) are proposed to identify the harmonic distortion. To enhance the identification and prediction performance, the standard IEEE 34-bus test feeder with verified harmonic sources power system is emulated for various scenarios using Electric Transient Analysis Program (ETAP). An extracting procedure for individual and total harmonic components are employed. The higher reduction level of error with approximately maximum median of 3.7127e $-^{15}$ % assures the accurate prediction of harmonic components. In addition, this work investigates the impact of several practical cases when the voltages of the renewable clean energy source arrays vary (e.g., solar cell, wind turbine, and EV), and a double-stage topology is needed to have the same amount of voltage at the input of the inverter before inversion.

This paper lays the foundations for Physical Layer Security (PLS) as a native 6G communication operation mode. 6G mobile network will enable intelligent applications, automated services, and low-latency connectivity to data networks. Security performance improvements become vital as 6G ubiquitously integrates with critical infrastructure. This paper presents 6G PLS frameworks that combine a Deep Neural Network (DNN) with a shared key-based PLS scheme at the decoding stage to achieve higher security performance. The paper analyzes and models an Artificial Intelligence (AI)/Machine Learning (ML) PLS Layer-1 security detection scheme that achieves an impressive approximately $10-15$ dB of improvement in Signal to Noise Ratio (SNR) performance over the linear PLS detection model in multiple fading environments modelled in COST 259 channel model and accompanying AWGN noise prior to error correction coding (ECC). An analysis of raw BER versus SNR is presented for multiple MIMO antenna sizes and codebook modulation orders. An analysis of the scheme shows that eavesdroppers cannot perfectly decode the secret information even under significantly improved SNR case scenarios. In contrast, legitimate receivers achieve nearly perfect secrecy at lower SNR cases by adopting the ML-based PLS model. In addition, the paper develops the theory and analysis, for the first time to our knowledge, of Alamouti Space-Time Coding (STC) over adjacent periods within a shared key-based PLS scheme with 2M transmit and 2M receive antennas. New modes of STC-based PLS show vast gains of approximately 15 dB over existing linear shared key-based PLS for multi-codebook modulation order in COST 259 environments and noise. Finally, the paper also proposes an operational overview of AI/ML integrated PLS with shared key-agreement protocol in an Open Radio Access Network (O-RAN) architecture model. The paper illustrates a secure, low latency, Layer-1 security O-RAN architecture scheme for 6G communication based upon “security Apps” for latency mitigation.

Unmanned aerial systems/vehicles (UAS/UAVs) are increasingly utilized for inspecting high-voltage (HV) Tx lines. Operating on batteries, these UAVs are equipped with various electrical sensors, microprocessors, and motors. All of these sensors are susceptible to electric (E) and magnetic (H) fields, which may affect sensing, communication, and control operations of UAS. This paper is first-of-its-kind large scale study that gathered field data (E/H) across 5 different Tx lines ( $69 kV_{AC}$ , $230 kV_{AC}$ , $345 kV_{AC}$ , $500 kV_{AC}$ , and $250 kV_{DC}$ ). Additional flight data on a microwave tower was also gathered. Multiple types of DJI UAVs (M2EA, M30, and M300) are used to carry E/H field sensors, and radio frequency (RF) sensors as payload for this study. Specific measurements include the E field in V/m, H field in mG, Battery voltage in V, Battery current in A, Battery percentage, Battery Temperature in °F, latitude, and longitude. The overall goal of the paper is to quantify E/H field distribution in HVTx lines which aided in realization of FAA rule making on how close UAVs could fly safely. The preliminary findings indicate that significantly higher E/H field level were reported near AC than DC Tx lines. The paper discusses conditions influencing E/H field strength during UAV operations. Additionally, hybrid machine learning (HML) models (RFR-SVM, RFR-KNN) were used to forecast battery drain. The results indicate that the hybrid RFR-KNN model yields lower MAPE values compared to stand-alone and other hybrid models.

The Open Radio Access Network (RAN) emerges as a revolutionary architecture promising unprecedented levels of openness, flexibility, and intelligence within radio access networks. Central to this innovation is the integration of Machine Learning (ML) and Artificial Intelligence (AI) within the RAN Intelligent Controller (RIC), aimed at optimizing network operations and enhancing control mechanisms. This paper undertakes a thorough examination of Open RAN, particularly focusing on its energy consumption aspects, which are pivotal for ensuring the sustainability of future wireless networks. In this paper, we review and compare Open RAN architecture with previous network architectures. In particular we focus on O-RAN Alliance specifications. Additionally, we explore the deployment of ML across various facets of Open RAN and highlights how to estimate the energy consumption of ML models. Through constructing explicit energy consumption models for key O-RAN components, we provide a granular analysis of their energy profiles. Finally we compare the energy dynamics of O-RAN against traditional RAN architectures, delineating the impact of virtualization and disaggregation on energy efficiency.

The integration of intelligent reflecting surfaces (IRS) into cooperative non-orthogonal multiple access (NOMA) systems revolutionizes wireless networks by enhancing signal strength, mitigating interference, and optimizing spectral efficiency. The cooperative NOMA (CNOMA) framework, empowered by IRS technology, further promises enhanced performance, robustness, and scalability for next-generation wireless networks as compared to NOMA only systems. This paper explores the synergy between IRS and NOMA to leverage cooperative techniques for superior wireless system design. Fundamental principles, technological advancements, and potential applications of IRS-assisted CNOMA systems are discussed, highlighting existing works. Both underlay and overlay NOMA principles are examined in conjunction with IRS in the paper. Spatial modulation-aided CNOMA is explored for multiple-input multiple-output (MIMO) systems, along with its advantages and practical challenges. Additionally, the paper discusses fundamental principles and technological advancements of IRS-assisted CNOMA systems, emphasizing solutions to potential challenges and the role of machine learning (ML)/deep learning (DL) in resource optimization like transmit power and IRS phase settings. Simulation results are presented to highlight the benefits of IRS-aided CNOMA system design. Finally, the paper outlines future directions and potential research topics in IRS-aided CNOMA.

According to the World Health Organization (WHO), some chronic diseases such as diabetes mellitus, stroke, cancer, cardiac vascular, kidney failure, and hypertension are essential for early prevention. One of the prevention that can be taken is to predict chronic diseases using machine learning based on personal medical record or general checkup result. The common prediction objective is to minimize the prediction error as low as possible. The most influencing chronic diseases prediction factors are the quality of data and the choice of predictor such as machine learning methods. The five main problems those lower data quality are outliers, missing values, feature selection, normalization, and imbalance. After we ensure the quality of data, the next task is to choose the best machine learning methods. The most influencing factor to consider when we choose the predictor its performance evaluation (accuracy, recall, precision, f1-score). Thus, predicting chronic disease aims to produce increased performance and solve problems in medical data. This paper presents a Systematic Literature Review (SLR) that offers a comprehensive discussion of research on chronic diseases prediction using machine learning and its data preprocessing handling. This paper covers machine learning methods discussion such as supervised learning, ensemble learning, deep learning, and reinforcement learning. The preprocessing handling we discuss includes missing values, outliers, feature selection, normalization, and imbalance. The final discussions of this paper are open issues, and the potential future works in improving the prediction performance for chronic diseases using a data preprocessing handling and machine learning methods.

In the rapidly evolving landscape of digital technology, the security of web and mobile applications stands paramount. As these platforms become increasingly integrated into our daily lives, the need for robust safety measures becomes imperative. This research paper delves into the intricate realm of web and mobile request safety, unraveling a multi-faceted exploration that combines traditional feature engineering with state-of-the-art machine learning models. Beginning with foundational models like TextCNN and TextRNN, we scrutinize their effectiveness in discerning the safety of requests. Advancing our investigation, we delve into the capabilities of sophisticated architectures, including Bidirectional LSTMs, DistilBERT, and RoBERTa. Beyond individual assessments, we introduce hybrid models that synergize the strengths of various approaches, establishing a comprehensive defense against emerging security threats. Throughout this research, we navigate the intricacies of model training, evaluation, and performance metrics. From accuracy and precision to recall and confusion matrices, each metric paints a nuanced picture of the efficacy of these models in ensuring the safety of web and mobile interactions. In a world where cyber threats loom large, the significance of this research lies not only in its technical contributions but also in its practical implications. By providing insights into innovative strategies for enhancing the security and resilience of digital applications, this paper contributes to the ongoing discourse on fortifying the digital infrastructure.

One of the most visible manifestations of the changes brought about by the digitization of everyday life is undoubtedly the spread of electronic commerce. It is difficult to think of the digital economy without considering transactions through electronic channels. In turn, the user interface (UI) is the key to e-commerce, as it is usually the first and primary point of contact between business and consumer. A key trend in e-commerce is the personalization of communications, which can improve the user experience, increase customer satisfaction and deliver tangible business benefits. Today, it is technically possible to base this personalization on an analysis of user behavior using artificial intelligence and machine learning techniques. A common form of personalization in e-commerce is the use of product recommendation systems, but the user interface can be tailored much more extensively. The approach described and discussed in this paper is a multivariant user interface that allows the layout to be tailored to the characteristics, attributes, and behaviors of customer groups generated using machine learning techniques. The results of the research carried out make it possible to verify the practicality of the proposed solution and provide an opportunity to identify development directions that take into account the potential of artificial intelligence. The application of the concept described in the paper is broad, covering all aspects of e-commerce design that require compromises when serving a single UI variant, but allow flexibility and customization for different users when serving a multivariant UI.

Federated Learning (FL) allows task initiators (servers) to utilize data from task participants (clients) to train machine learning models while protecting data privacy. However, in the FL system, when the client data are non-independently identically distributed (Non-IID), appropriate metrics are chosen to accurately evaluate the quality of the client data, accordingly to select a reasonable subset of clients, and thus ensure the accuracy of the FL aggregation model. In this paper, based on the experimental results, a data distribution evaluation model is proposed, which is based on two metrics: the volume of client data and its increment and the balance of global client data. This data distribution evaluation model enables more accurate evaluation of clients with Non-IID characteristics. Based on this evaluation model, this paper further proposes an FL client subset selection algorithm. This algorithm accurately evaluates the data value of each client, enabling the server to select the most valuable subset of clients before FL training, thus improving the accuracy of the federated learning aggregation model in scenarios with Non-IID client data. When training the FL aggregation model using the proposed method on datasets composed of CIFAR-10, Fashion-MNIST, and DEAP distributions, compared to the optimal baseline, the average precision scores increased by 5.99%, 4.79%, and 4.29% respectively. The improvement in accuracy is more pronounced in scenarios with Non-IID data, such as in the DEAP dataset distribution with the highest Non-IID degree, where the accuracy increased by 5.30% compared to the optimal baseline.

This research paper intends to provide real-life applications of Generative AI (GAI) in the cybersecurity domain. The frequency, sophistication and impact of cyber threats have continued to rise in today’s world. This ever-evolving threat landscape poses challenges for organizations and security professionals who continue looking for better solutions to tackle these threats. GAI technology provides an effective way for them to address these issues in an automated manner with increasing efficiency. It enables them to work on more critical security aspects which require human intervention, while GAI systems deal with general threat situations. Further, GAI systems can better detect novel malware and threatening situations than humans. This feature of GAI, when leveraged, can lead to higher robustness of the security system. Many tech giants like Google, Microsoft etc., are motivated by this idea and are incorporating elements of GAI in their cybersecurity systems to make them more efficient in dealing with ever-evolving threats. Many cybersecurity tools like Google Cloud Security AI Workbench, Microsoft Security Copilot, SentinelOne Purple AI etc., have come into the picture, which leverage GAI to develop more straightforward and robust ways to deal with emerging cybersecurity perils. With the advent of GAI in the cybersecurity domain, one also needs to take into account the limitations and drawbacks that such systems have. This paper also provides some of the limitations of GAI, like periodically giving wrong results, costly training, the potential of GAI being used by malicious actors for illicit activities etc.

Traffic collisions pose a significant global concern, necessitating innovative solutions to enhance safety. Post-collision, it becomes crucial for the police to determine the responsibilities of the involved parties, distinguishing between criminal acts and non-criminal incidents. Insurance companies also rely on such investigations to compensate victims. Assessing responsibility after a crash is a complex task requiring advanced knowledge of road rules. For straightforward scenarios like crashes with traffic lights, decisions are fast and easy. However, in situations such as crashes without any traffic signs, expert knowledge is essential. Automating such tasks demands innovative approaches, representing a necessity for the future of the automobile and insurance industries. Despite these critical needs, there has been limited research in the domain. This paper introduces a system that is capable of detecting vehicle collision and implements an original responsibility assessment process to assess drivers responsibilities. This paper is an extended version of our previous work. In the previous work, the system supported only three types of head-on/angle crash scenarios without traffic lights within three weather conditions, in addition to those with traffic lights. This study presents the improved system which now accommodates six different types of head-on/angle crash scenarios without traffic lights within six harsh weather conditions, such as foggy and snowy days. Additionally, extensive experiments are conducted with results showing that the system performs better than its previous version, mainly during nighttime without traffic lights (up to 93% accuracy against up to 82.5% obtained previously). Moreover, through case studies and comparisons with existing research, the effectiveness and superiority of the system are demonstrated.

The breath rate can now be monitored remotely due to the advancements in digital stethoscope sensor technology, signal processing, and machine learning. Automatic breathing rate classification, on the other hand, provides additional benefits in medical diagnostics. In this paper, a lightweight convolutional neural network is proposed for automatic breathing rate classification utilizing the piezoresistive sensor data. In the proposed work, the raw signals from the piezoresistive sensor are pre-processed using a continuous wavelet transform to generate the corresponding images. These images are then fed into a lightweight convolutional neural network, which efficiently classifies the breathing rate into six classes based on the number of breaths per minute. Through extensive results, we show that the proposed model results in a classification accuracy of 96.40% which is higher than all the benchmark models considered in this paper. We also evaluate the performance of the proposed model using edge computing devices such as Raspberry Pi, Nvidia AGX Xavier, and Nvidia Jetson Nano.

Alzheimer’s Disease (AD) accounts for the majority of dementia, and Mild Cognitive Impairment (MCI) is the early stage of AD. Early and accurate diagnosis of dementia plays a vital role in more targeted treatments and effectively halting disease progression. However, the clinical diagnosis of dementia requires various examinations, which are expensive and require a high level of expertise from the doctor. In this paper, we proposed a classification method based on multi-modal data including Electroencephalogram (EEG), eye tracking and behavioral data for early diagnosis of AD and MCI. Paradigms with various task difficulties were used to identify different severity of dementia: eye movement task and resting-state EEG tasks were used to detect AD, while eye movement task and delayed match-to-sample task were used to detect MCI. Besides, the effects of different features were compared and suitable EEG channels were selected for the detection. Furthermore, we proposed a data augmentation method to enlarge the dataset, designed an extra ERPNet feature extract layer to extract multi-modal features and used domain-adversarial neural network to improve the performance of MCI diagnosis. We achieved an average accuracy of 88.81% for MCI diagnosis and 100% for AD diagnosis. The results of this paper suggest that our classification method can provide a feasible and affordable way to diagnose dementia.

Rare event detection (RED) involves the identification and detection of events characterized by low frequency of occurrences, but of high importance or impact. This paper presents a Systematic Review (SR) of rare event detection across various modalities using Machine Learning (ML) and Deep Learning (DL) techniques. This review comprehensively outlines techniques and methods best suited for rare event detection across various modalities, while also highlighting future research prospects. To the extent of our knowledge, this paper is a pioneering SR dedicated to exploring this specific research domain. This SR identifies the employed methods and techniques, the datasets utilized, and the effectiveness of these methods in detecting rare events. Four modalities concerning RED are reviewed in this SR: video, sound, image, and time series. The corresponding performances for the different ML and DL techniques for RED are discussed comprehensively, together with the associated RED challenges and limitations as well as the directions for future research are highlighted. This SR aims to offer a comprehensive overview of the existing methods in RED, serving as a valuable resource for researchers and practitioners working in the respective field.