Search

As noted in Wikipedia, skin in the game refers to having 'incurred risk by being involved in achieving a goal', where 'skin is a synecdoche for the person involved, and game is the metaphor for actions on the field of play under discussion'. For exascale applications under development in the US Department of Energy Exascale Computing Project, nothing could be more apt, with the skin being exascale applications and the game being delivering comprehensive science-based computational applications that effectively exploit exascale high-performance computing technologies to provide breakthrough modelling and simulation and data science solutions. These solutions will yield high-confidence insights and answers to the most critical problems and challenges for the USA in scientific discovery, national security, energy assurance, economic competitiveness and advanced healthcare. This article is part of a discussion meeting issue 'Numerical algorithms for high-performance computational science'.

We demonstrate the advantage of using machine learning for surface enhanced Raman scattering (SERS) spectral analysis for quantitative detection of pyocyanin in Luria-Bertani media. Planar Au nanoparticle clusters were selfassembled on PS-b-PMMA diblock copolymer template using EDC crosslinking chemistry and electrohydrodynamic flow to fabricate SERS substrates. Resulting substrates produce uniform SERS response over large area with signal relative standard deviation of 10.8 % over 50 μm × 50 μm region. Taking advantage of the uniformity, 400 SERS spectra were collected at each pyocyanin concentration as training dataset. Tracking the intensity of pyocyanin 1350 cm-1 vibrational band shows linear regime beginning at 10 ppb. PLS analysis was also performed on the same training dataset. Without being explicitly "told" which spectrum to look for, PLS analysis recognizes the SERS spectrum of pyocyanin as its first loading vector even in the presence of other molecules in LB media. PLS regression enables quantitative detection at 1 ppb, 1 order of magnitude earlier than univariate regression. We hope this work will fuel a push toward wider adoption of more sophisticated machine learning algorithms for quantitative analysis of SERS spectra.

Objectives: Upper tract urothelial carcinoma (UTUC) is a rare, aggressive lesion, with early detection a key to its management. This study aimed to utilise computed tomographic urogram data to develop machine learning models for predicting tumour grading and staging in upper urothelial tract carcinoma patients and to compare these predictions with histopathological diagnosis used as reference standards. Methods: Protocol-based computed tomographic urogram data from 106 patients were obtained and visualised in 3D. Digital segmentation of the tumours was conducted by extracting textural radiomics features. They were further classified using 11 predictive models. The predicted grades and stages were compared to the histopathology of radical nephroureterectomy specimens. Results: Classifier models worked well in mining the radiomics data and delivered satisfactory predictive machine learning models. The multilayer panel showed 84% sensitivity and 93% specificity while predicting UTUC grades. The Logistic Regression model showed a sensitivity of 83% and a specificity of 76% while staging. Similarly, other classifier algorithms [e.g. Support Vector classifier (SVC)] provided a highly accurate prediction while grading UTUC compared to clinical features alone or ureteroscopic biopsy histopathology. Conclusion: Data mining tools could handle medical imaging datasets from small (<2 cm) tumours for UTUC. The radiomics-based machine learning algorithms provide a potential tool to model tumour grading and staging with implications for clinical practice and the upgradation of current paradigms in cancer diagnostics. Clinical Relevance: Machine learning based on radiomics features can predict upper tract urothelial cancer grading and staging with significant improvement over ureteroscopic histopathology. The study showcased the prowess of such emerging tools in the set objectives with implications towards virtual biopsy. [ABSTRACT FROM AUTHOR]

Despite the expert-level performance of artificial intelligence (AI) models for various medical imaging tasks, real-world performance failures with disparate outputs for various subgroups limit the usefulness of AI in improving patients' lives. Many definitions of fairness have been proposed, with discussions of various tensions that arise in the choice of an appropriate metric to use to evaluate bias; for example, should one aim for individual or group fairness? One central observation is that AI models apply "shortcut learning" whereby spurious features (such as chest tubes and portable radiographic markers on intensive care unit chest radiography) on medical images are used for prediction instead of identifying true pathology. Moreover, AI has been shown to have a remarkable ability to detect protected attributes of age, sex, and race, while the same models demonstrate bias against historically underserved subgroups of age, sex, and race in disease diagnosis. Therefore, an AI model may take shortcut predictions from these correlations and subsequently generate an outcome that is biased toward certain subgroups even when protected attributes are not explicitly used as inputs into the model. As a result, these subgroups became nonprivileged subgroups. In this review, the authors discuss the various types of bias from shortcut learning that may occur at different phases of AI model development, including data bias, modeling bias, and inference bias. The authors thereafter summarize various tool kits that can be used to evaluate and mitigate bias and note that these have largely been applied to nonmedical domains and require more evaluation for medical AI. The authors then summarize current techniques for mitigating bias from preprocessing (data-centric solutions) and during model development (computational solutions) and postprocessing (recalibration of learning). Ongoing legal changes where the use of a biased model will be penalized highlight the necessity of understanding, detecting, and mitigating biases from shortcut learning and will require diverse research teams looking at the whole AI pipeline. [ABSTRACT FROM AUTHOR]

The development of disease detection with progression modelling due to long-term exposure to toxicity is complex. This leads to unwrapping the underlying intricate molecular network of toxicity and is thus a crucial challenge for the researchers. Therefore, identifying a set of biomarkers to predict the risk of exposure is vital. Thus, this article aims to provide a holistic machine learning-based solution over various time-varying 'omic data to understand and explore the factors involved in the development of diseases. To address this issue, a flexible non-negative matrix factorization based multi-level self organizing map (FNMF-MLSOM) is developed. The proposed algorithm utilizes two open-source time series datasets: Type-2 diabetes mellitus and Huntington disease, namely. The flexible non-negative matrix factorization based self organization model introduced in this article provides a negative value acceptance constraint as well as the clustering on the basis matrix to keep the biological meaning of the data intact. Since microarray data have rich information, we applied the proposed method to obtain the progression-specific convoluted biomarker for precise feature extraction. Further, to validate the differentially expressed biomarkers, the proposed method is applied to the test samples to verify the mathematical validity as well as the biological significance of the biomarkers. [ABSTRACT FROM AUTHOR]

Highlights • Sexual selection have lower extinction rates and higher turnover rates, compared with random selection populations. • Random mate selection has higher speciation rates compared with sexual selection. • Increased body size leads to higher extinction rates compared to lower body size. Abstract It is not clear from empirical and simulation studies that populations with females who employ sexual selection have any evolutionary advantages over populations where mates are randomly selected. There is an ongoing debate regarding whether speciation rates and extinction rates differ significantly between sexual selection and random selection. Although there is evidence that sexual selection drives speciation in some animal species,the biological community remains divided regarding this relationship. Similarly, multiple studies point to a possible connection between sexual selection and extinction rates, although there is no clear consensus regarding this connection: Some studies suggest that sexual selection increases the extinction rate whereas others suggest that sexual selection actually shields populations from extinction. Using individual based computer simulations, we found a significant difference between sexual selection and random selection, with respect to speciation rates, extinction rates and species turnover rates: It turned out that speciation rates were significantly higher for random selection, possibly to help offset the higher extinction and turnover rates. Moreover, we used machine learning to generate rules to help predict rates of speciation and extinction both for sexual selection and random selection. Not only were our rules corroborated by empirical studies but they also help to resolve some disputes regarding the role of sexual selection with respect to speciation rates and extinction rates. [ABSTRACT FROM AUTHOR]

Small data in materials present significant challenges to constructing highly accurate machine learning models, severely hindering the widespread implementation of data‐driven materials intelligent design. In this study, the Dual‐Strategy Materials Intelligent Design Framework (DSMID) is introduced, which integrates two innovative methods. The Adversarial domain Adaptive Embedding Generative network (AAEG) transfers data between related property datasets, even with only 90 data points, enhancing material composition characterization and improving property prediction. Additionally, to address the challenge of screening and evaluating numerous alloy designs, the Automated Material Screening and Evaluation Pipeline (AMSEP) is implemented. This pipeline utilizes large language models with extensive domain knowledge to efficiently identify promising experimental candidates through self‐retrieval and self‐summarization. Experimental findings demonstrate that this approach effectively identifies and prepares new eutectic High Entropy Alloy (EHEA), notably Al14(CoCrFe)19Ni28, achieving an ultimate tensile strength of 1085 MPa and 24% elongation without heat treatment or extra processing. This demonstrates significantly greater plasticity and equivalent strength compared to the typical as‐cast eutectic HEA AlCoCrFeNi2.1. The DSMID framework, combining AAEG and AMSEP, addresses the challenges of small data modeling and extensive candidate screening, contributing to cost reduction and enhanced efficiency of material design. This framework offers a promising avenue for intelligent material design, particularly in scenarios constrained by limited data availability. [ABSTRACT FROM AUTHOR]

Capsule endoscopy (CE) represents an important groundbreaking advancement in gastrointestinal (GI) examinations, distinguished by its noninvasive, painless, and convenient nature, and has swiftly established itself as a crucial tool for diagnosing and treating digestive diseases. With the development of artificial intelligence (AI) and machine learning (ML), as AI and ML progress, the capabilities of CE have expanded beyond mere imaging within the GI tract; it is progressively evolving to encompass procedures such as biopsies and targeted drug delivery. This review systematically searched five reputable repositories—Scopus, PubMed, IEEE Xplore, ACM Digital Library, and ScienceDirect—for all original publications on CE from 2001 to 2024. The review provides an overview of the current status and identified limitations of CE, highlighting the significant role that AI and ML are projected to play in its future development. [ABSTRACT FROM AUTHOR]

Climate change is increasing the occurrence of extreme weather events, such as intense windstorms, with a trend expected to worsen due to global warming. The growing intensity and frequency of these events are causing a significant number of failures in power distribution grids. However, understanding the nature of extreme wind events and predicting their impact on distribution grids can help and prevent these issues, potentially mitigating their adverse effects. This study analyzes a structured method to predict distribution grid disruptions caused by extreme wind events. The method utilizes Machine Learning (ML) models, including K-Nearest Neighbors (KNN), Random Forest (RF), Support Vector Machine (SVM), Decision Trees (DTs), Gradient Boosting Machine (GBM), Gaussian Process (GP), Deep Neural Network (DNN), and Ensemble Learning which combines RF, SVM and GP to analyze synthetic failure data and predict power grid outages. The study utilized meteorological information, physical fragility curves, and scenario generation for distribution systems. The approach is validated by using five-fold cross-validation on the dataset, demonstrating its effectiveness in enhancing predictive capabilities against extreme wind events. Experimental results showed that the Ensemble Learning, GP, and SVM models outperformed other predictive models in the binary classification task of identifying failures or non-failures, achieving the highest performance metrics. [ABSTRACT FROM AUTHOR]

Lung cancer is an important global health problem, and it is defined by abnormal growth of the cells in the tissues of the lung, mostly leading to significant morbidity and mortality. Its timely identification and correct staging are very important for proper therapy and prognosis. Different computational methods have been used to enhance the precision of lung cancer classification, among which optimization algorithms such as Greylag Goose Optimization (GGO) are employed. These algorithms have the purpose of improving the performance of machine learning models that are presented with a large amount of complex data, selecting the most important features. As per lung cancer classification, data preparation is one of the most important steps, which contains the operations of scaling, normalization, and handling gap factor to ensure reasonable and reliable input data. In this domain, the use of GGO includes refining feature selection, which mainly focuses on enhancing the classification accuracy compared to other binary format optimization algorithms, like bSC, bMVO, bPSO, bWOA, bGWO, and bFOA. The efficiency of the bGGO algorithm in choosing the optimal features for improved classification accuracy is an indicator of the possible application of this method in the field of lung cancer diagnosis. The GGO achieved the highest accuracy with MLP model performance at 98.4%. The feature selection and classification results were assessed using statistical analysis, which utilized the Wilcoxon signed-rank test and ANOVA. The results were also accompanied by a set of graphical illustrations that ensured the adequacy and efficiency of the adopted hybrid method (GGO + MLP). [ABSTRACT FROM AUTHOR]

Background: As patients interact with a healthcare service system, patterns of service utilization (PSUs) emerge. These PSUs are embedded in the sparse high-dimensional space of longitudinal cross-continuum health service encounter data. Once extracted, PSUs can provide quality assurance/quality improvement (QA/QI) efforts with the information required to optimize service system structures and functions. This may improve outcomes for complex patients with chronic diseases. Method: Working with longitudinal cross-continuum encounter data from a regional health service system, various pattern detection analyses were conducted, employing (1) graph community detection algorithms, (2) natural language processing (NLP) clustering, and (3) a hybrid NLP–graph method. Result: These approaches produced similar PSUs, as determined from a clinical perspective by clinical subject matter experts and service system operations experts. Conclusions: The similarity in the results provides validation for the methodologies. Moreover, the results stress the need to engage with clinical or service system operations experts, both in providing the taxonomies and ontologies of the service system, the cohort definitions, and determining the level of granularity that produces the most clinically meaningful results. Finally, the uniqueness of each approach provides an opportunity to take advantage of the various analytical capabilities that each approach brings, which will be further explored in our future research. [ABSTRACT FROM AUTHOR]

The development of artificial intelligence (AI) and self-driving technology is expected to enhance intelligent transportation systems (ITSs) by improving road safety and mobility, increasing traffic flow, and reducing vehicle emissions in the near future. In an ITS, each autonomous vehicle acts as a node with its own local machine learning models, which can be updated using locally collected data. However, for autonomous vehicles to learn effective models, they must be able to learn from data sources provided by other vehicles and infrastructure, utilizing innovative learning methods to adapt to various autonomous driving scenarios. Distributed learning plays a crucial role in implementing these learning tasks in an ITS. This review provides a systematic overview of distributed learning in the field of ITSs. Within an ITS, vehicles can engage in distributed learning by interacting with peers through opportunistic encounters and clustering. This study examines the challenges associated with distributed learning, focusing on issues related to privacy and security in data intelligence sharing, communication quality and speed, and trust. Through a thorough analysis of these challenges, this study presents potential research avenues to address these issues, including the utilization of incentive mechanisms that rely on reputation, the adoption of rapid convergence techniques, and the integration of opportunistic federated learning with blockchain technology. [ABSTRACT FROM AUTHOR]

Animal communication impacts many kinds of behavior including mating and courtship, escaping from predators and foraging activity. In this article, our investigation focuses primarily on how alarm communication impacts foraging activity, using individual based computer simulations. We used this approach to help resolve a debate in the literature between the risk-allocation hypothesis, which predicts that over time, animals become habituated to alarm communication versus the hypothesis that alarm communication consistently decreases foraging activity to avoid predators. We found that in most cases, alarm communication did indeed decrease foraging activity whereas in other cases, alarm communication resulted in habituation and a gradual increase in foraging activity, suggesting that there is some truth to both hypotheses. Moreover, it is possible that a decrease in foraging as well as habituation in response to alarm communication both contribute to fitness, or more generally, that alarm communication contributes to fitness as opposed to non-communication. Among the communication runs, we found that although there were higher levels of fitness compared with non-communication runs, fitness was higher when communication results in decreased foraging activity vs. runs where communication results in increased foraging activity. Finally, we used a variational autoencoder based estimation of distribution algorithms in conjuction with C4.5 decision trees as a wrapper to discern the features that distinguish communication runs from non-communication runs. In general, communication runs tend to have relatively low population densities, whereas non-communication runs tend to have relatively high population densities, suggesting that the ability to communicate fear obviates the need for prey to stay in close proximity to one another in order to defend against predators. Also a high level of reproductive urgency was observed in individuals with communication ability when the level of fear of predators was low. • In most cases alarm communication decreased foraging activity. • Communication becomes more urgent when there are fewer conspecifics to defend against predators. • There were higher levels of fitness in Communication runs compared to non-communication runs • The fitness was higher when communication results in decreased foraging activity [ABSTRACT FROM AUTHOR]

This paper provides a comprehensive review of machine learning strategies and optimization formulations employed in energy management systems (EMS) tailored for plug-in hybrid electric vehicles (PHEVs). EMS stands as a pivotal component facilitating optimized power distribution, predictive and adaptive control strategies, component health monitoring, and energy harvesting, thereby enabling the maximal exploitation of resources through optimal operation. Recent advancements have introduced innovative solutions such as Model Predictive Control (MPC), machine learning-based techniques, real-time optimization algorithms, hybrid optimization approaches, and the integration of fuzzy logic with neural networks, significantly enhancing the efficiency and performance of EMS. Additionally, multi-objective optimization, stochastic and robust optimization methods, and emerging quantum computing approaches are pushing the boundaries of EMS capabilities. Remarkable advancements have been made in data-driven modeling, decision-making, and real-time adjustments, propelling machine learning and optimization to the forefront of enhanced control systems for vehicular applications. However, despite these strides, there remain unexplored research avenues and challenges awaiting investigation. This review synthesizes existing knowledge, identifies gaps, and underscores the importance of continued inquiry to address unanswered research questions, thereby propelling the field toward further advancements in PHEV EMS design and implementation. [ABSTRACT FROM AUTHOR]

The traditional design method for terahertz metasurface biosensors is cumbersome and time-consuming, requires expertise, and often leads to significant discrepancies between expected and actual values. This paper presents a novel approach for the fast, efficient, and convenient inverse design of THz metasurface sensors, leveraging convolutional neural network techniques based on deep learning. During the model training process, the magnitude data of the scattering parameters collected from the numerical simulation of the THz metasurface served as features, paired with corresponding surface structure matrices as labels to form the training dataset. During the validation process, the thoroughly trained model precisely predicted the expected surface structure matrix of a THz metasurface. The results demonstrate that the proposed algorithm realizes time-saving, high-efficiency, and high-precision inversion methods without complicated data preprocessing and additional optimization algorithms. Therefore, deep learning algorithms offer a novel approach for swiftly designing and optimizing THz metasurface sensors in biomedical detection, bypassing the complex and specialized design process of electromagnetic devices, and promising extensive prospects for their application in the biomedical field. [ABSTRACT FROM AUTHOR]

Novel architectures incorporating transport networks and artificial intelligence (AI) are currently being developed for beyond 5G and 6G technologies. Given that the interfacing mobile and transport network nodes deliver high transactional packet volume in downlink and uplink streams, 6G networks envision adopting diverse transport networks, including non-terrestrial types of transport networks such as the satellite network, High-Altitude Platform Systems (HAPS), and DOCSIS cable TV. Hence, there is a need to match the traffic to the transport network. This paper focuses on such a matching problem and defines a method that leverages machine learning and mixed-integer linear programming. Consequently, the proposed scheme in this paper is to develop a traffic steering capability based on types of transport networks, namely, optical, satellite, and DOCSIS cable. Novel findings demonstrate a more than 90% accuracy of steered traffic to respective types of transport networks for dedicated transport network resources. [ABSTRACT FROM AUTHOR]

Learning from high dimensional data has been utilized in various applications such as computational biology, image classification, and finance. Most classical machine learning algorithms fail to give accurate predictions in high dimensional settings due to the enormous feature space. In this article, we present a novel ensemble of classification trees based on weighted random subspaces that aims to adjust the distribution of selection probabilities. In the proposed algorithm base classifiers are built on random feature subspaces in which the probability that influential features will be selected for the next subspace, is updated by incorporating grouping information based on previous classifiers through a weighting function. As an interpretation tool, we show that variable importance measures computed by the new method can identify influential features efficiently. We provide theoretical reasoning for the different elements of the proposed method, and we evaluate the usefulness of the new method based on simulation studies and real data analysis. [ABSTRACT FROM AUTHOR]

The study of high-entropy (HE) alloys has seen dramatic growth in recent years as, in some cases, these systems can exhibit exceptional properties, including enhanced oxidation resistance, superior mechanical properties, and desirable magnetic properties. The identification of promising HE alloys is, however, extremely challenging due to the extraordinarily large number of distinct systems that may be fabricated from the available palette of elements. For this reason, machine learning strategies have been employed to reduce the size of the associated chemistry/composition space. In this review, we outline several computational strategies that have led to the identification of useful alloys and discuss the relative merits and shortcomings of these approaches. We also present short tutorials illustrating the use of selected computational approaches to HE characterization and design. [ABSTRACT FROM AUTHOR]

During the optimization process, a large number of data are generated through the search. Machine learning techniques and algorithms can be used to handle the generated data to contribute to the optimization process. The use of machine learning enhancing metaheuristics applied to optimization problems has been drawing attention due to their capacity to add domain knowledge during the search process. This knowledge can accelerate metaheuristics and lead to better and promising solutions. This work provides a systematic literature review of machine learning enhancing metaheuristics and summarizes the current state of the classification of the research field, main techniques and machine learning models, validations strategies, and real-world optimization problems that the approach was applied. Our keyword search found 1.960 papers, published in the last 10 years. After considering the inclusion and exclusion criteria and performing backward snowballing procedure, we have analyzed 111 primary studies. The results show the predominance of the use of surrogate-assisted evolutionary algorithms (SAEAs) for improving the efficiency of the optimization, and the use of estimation of distribution algorithms (EDAs) to increase the effectiveness of the optimization. The objective function value is the mostly applied evaluating criteria to validate the algorithm with other methods. The developed techniques of the studies found are applied in diverse real-world applications such as developing machine learning models, physics simulations with expensive function evaluation, and the variants of the classical job shop scheduling problem. We also discuss trends and opportunities of the research field. [ABSTRACT FROM AUTHOR]

Cluster analysis is a crucial technique in unsupervised machine learning, pattern recognition, and data analysis. However, current clustering algorithms suffer from the need for manual determination of parameter values, low accuracy, and inconsistent performance concerning data size and structure. To address these challenges, a novel clustering algorithm called the fully automated density-based clustering method (FADBC) is proposed. The FADBC method consists of two stages: parameter selection and cluster extraction. In the first stage, a proposed method extracts optimal parameters for the dataset, including the epsilon size and a minimum number of points thresholds. These parameters are then used in a density-based technique to scan each point in the dataset and evaluate neighborhood densities to find clusters. The proposed method was evaluated on different benchmark datasets and metrics, and the experimental results demonstrate its competitive performance without requiring manual inputs. The results show that the FADBC method outperforms well-known clustering methods such as the agglomerative hierarchical method, k-means, spectral clustering, DBSCAN, FCDCSD, Gaussian mixtures, and density-based spatial clustering methods. It can handle any kind of data set well and perform excellently. [ABSTRACT FROM AUTHOR]

Computational materials design (CMD) based on the first-principles electronic structure calculations is demonstrated for two topics related to the design of high-entropy alloys (HEAs). The first one is a construction of prediction model of elastic constants. By applying machine learning technique with the use of the linearly independent descriptor generation method to the database of elastic constants of 2555 BCC HEAs generated by the full potential Korringa-Kohn-Rostoker coherent potential approximation (FPKKR-CPA) method. The obtained model is used to predict new HEAs with high Young’s modulus. The second topic is a simulation of atomic arrangement in HEAs at finite temperature. In this simulation, HEAs are described by using the Potts-like model and the interaction parameters are determined based on the generalized perturbation method combined with the KKR-CPA method. Monte Carlo simulations for the models of CrMnFeCoNi and CrMnFeCoCu predict atomic arrangements which are consistent to the experimental observations. [ABSTRACT FROM AUTHOR]

Tuberculosis (TB) is one of the leading infectious causes of death worldwide. The effective management and public health control of this disease depends on early detection and careful treatment monitoring. For many years, the microscopy-based analysis of sputum smears has been the most common method to detect and quantify Mycobacterium tuberculosis (Mtb) bacteria. Nonetheless, this form of analysis is a challenging procedure since sputum examination can only be reliably performed by trained personnel with rigorous quality control systems in place. Additionally, it is affected by subjective judgement. Furthermore, although fluorescence-based sample staining methods have made the procedure easier in recent years, the microscopic examination of sputum is a time-consuming operation. Over the past two decades, attempts have been made to automate this practice. Most approaches have focused on establishing an automated method of diagnosis, while others have centred on measuring the bacterial load or detecting and localising Mtb cells for further research on the phenotypic characteristics of their morphology. The literature has incorporated machine learning (ML) and computer vision approaches as part of the methodology to achieve these goals. In this review, we first gathered publicly available TB sputum smear microscopy image sets and analysed the disparities in these datasets. Thereafter, we analysed the most common evaluation metrics used to assess the efficacy of each method in its particular field. Finally, we generated comprehensive summaries of prior work on ML and deep learning (DL) methods for automated TB detection, including a review of their limitations. [ABSTRACT FROM AUTHOR]

In the last decade, deep learning has reinvigorated the machine learning field. It has solved many problems in computer vision, speech recognition, natural language processing, and other domains with state-of-the-art performances. In these domains, the data is generally represented in the Euclidean space. Various other domains conform to non-Euclidean space, for which a graph is an ideal representation. Graphs are suitable for representing the dependencies and inter-relationships between various entities. Traditionally, handcrafted features for graphs are incapable of providing the necessary inference for various tasks from this complex data representation. Recently, there has been an emergence of employing various advances in deep learning for graph-based tasks (called Graph Neural Networks (GNNs)). This article introduces preliminary knowledge regarding GNNs and comprehensively surveys GNNs in different learning paradigms—supervised, unsupervised, semi-supervised, self-supervised, and few-shot or meta-learning. The taxonomy of each graph-based learning setting is provided with logical divisions of methods falling in the given learning setting. The approaches for each learning task are analyzed from theoretical and empirical standpoints. Further, we provide general architecture design guidelines for building GNN models. Various applications and benchmark datasets are also provided, along with open challenges still plaguing the general applicability of GNNs. [ABSTRACT FROM AUTHOR]

The COVID-19 pandemic has posed significant challenges to existing healthcare systems around the world. The urgent need for the development of diagnostic and therapeutic strategies for COVID-19 has boomed the demand for new technologies that can improve current healthcare approaches, moving towards more advanced, digitalized, personalized, and patient-oriented systems. Microfluidic-based technologies involve the miniaturization of large-scale devices and laboratory-based procedures, enabling complex chemical and biological operations that are conventionally performed at the macro-scale to be carried out on the microscale or less. The advantages microfluidic systems offer such as rapid, low-cost, accurate, and on-site solutions make these tools extremely useful and effective in the fight against COVID-19. In particular, microfluidic-assisted systems are of great interest in different COVID-19-related domains, varying from direct and indirect detection of COVID-19 infections to drug and vaccine discovery and their targeted delivery. Here, we review recent advances in the use of microfluidic platforms to diagnose, treat or prevent COVID-19. We start by summarizing recent microfluidic-based diagnostic solutions applicable to COVID-19. We then highlight the key roles microfluidics play in developing COVID-19 vaccines and testing how vaccine candidates perform, with a focus on RNA-delivery technologies and nano-carriers. Next, microfluidic-based efforts devoted to assessing the efficacy of potential COVID-19 drugs, either repurposed or new, and their targeted delivery to infected sites are summarized. We conclude by providing future perspectives and research directions that are critical to effectively prevent or respond to future pandemics. [ABSTRACT FROM AUTHOR]

Fantasy Sports allows individuals to assemble a virtual team to participate in free or paid tournaments and earn rewards. Selecting a good team forms a crucial decision in fantasy cricket. Existing team selection methods cater only to professional cricket and are not suited well to accommodate the differences between fantasy cricket and the on-field game. This paper proposes a two-step methodology for player assessment and team selection in fantasy cricket. Player assessment is carried out using recursive feature elimination in random forest, in which context relevant player metrics are considered and the selection of players is based on modified genetic algorithm. We illustrate the efficacy of the proposed method on Dream11, a popular fantasy sports application. The results show that the proposed method outshines the traditional team selection process in fantasy sports, which is based on hit and trial. Furthermore, we provide a typology to analyse the proposed algorithm along the dimensions of reward distribution and entry fee. [ABSTRACT FROM AUTHOR]

Regional climate models (RCMs) have a high computational cost due to their higher spatial resolution compared to global climate models (GCMs). Therefore, various downscaling approaches have been developed as a surrogate for the dynamical downscaling of GCMs. This study assesses the potential of using a cost‐efficient machine learning alternative to dynamical downscaling by using the example case study of emulating surface mass balance (SMB) over the Antarctic Peninsula. More specifically, we determine the impact of the training framework by comparing two training scenarios: (a) a perfect and (b) an imperfect model framework. In the perfect model framework, the RCM‐emulator learns only the downscaling function; therefore, it was trained with upscaled RCM (UPRCM) features at GCM resolution. This emulator accurately reproduced SMB when evaluated on UPRCM, but its predictions on GCM data conserved RCM‐GCM inconsistencies and led to underestimation. In the imperfect model framework, the RCM‐emulator was trained with GCM features and downscaled the GCM while exposed to RCM‐GCM inconsistencies. This emulator predicted SMB close to the truth, showing it learned the underlying inconsistencies and dynamics. Our results suggest that a deep learning RCM‐emulator can learn the proper GCM to RCM downscaling function while working directly with GCM data. Furthermore, the RCM‐emulator presents a significant computational gain compared to an RCM simulation. We conclude that machine learning emulators can be applied to produce fast and fine‐scaled predictions of RCM simulations from GCM data. Plain Language Summary: Over the last century, climate scientists have tried to deepen their understanding of the behavior of climate processes through two types of computer climate simulations: global (GCMs) and regional (RCMs) climate models. GCMs cover the whole planet but do not contain fine spatial details, whereas RCMs provide highly detailed information but cover small areas and come at a high additional computational cost. Therefore, we imitated regional models from global models using machine learning to facilitate their faster development. To test our machine learning framework, we focused on the Antarctic Peninsula and aimed to reproduce the surface mass balance (SMB) of ice formation and loss. We trained our model to learn the relationship between a group of low‐resolution images of climate variables and a high‐resolution image from SMB images in the same region. Our results show that the machine learning model is fast and could recreate regional images of ice sheet processes from global data almost identical to existing on‐site observations. This is a good start for further usage of machine learning emulators. In conclusion, we can make fast and detailed reproductions of SMB processes at regional scales from globally accessible climate data using machine learning. Key Points: We developed a computationally fast machine learning emulator to downscale a global climate model (GCM) to regional resolutionThe emulator reproduces regional high‐resolution surface mass balance predictions over the Antarctic Peninsula from a GCMThe imperfect model framework outperforms the perfect model framework in the application success of the deep learning emulator [ABSTRACT FROM AUTHOR]

The aim of this study was to explore the relationship between lipids with different structural features and lung cancer (LC) risk and identify prospective biomarkers of LC. Univariate and multivariate analysis methods were used to screen for differential lipids, and two machine learning methods were used to define combined lipid biomarkers. A lipid score (LS) based on lipid biomarkers was calculated, and a mediation analysis was performed. A total of 605 lipid species spanning 20 individual lipid classes were identified in the plasma lipidome. Higher carbon atoms with dihydroceramide (DCER), phosphatidylethanolamine (PE), and phosphoinositols (PI) presented a significant negative correlation with LC. Point estimates revealed the inverse associated with LC for the n-3 PUFA score. Ten lipids were identified as markers with an area under the curve (AUC) value of 0.947 (95%, CI: 0.879–0.989). In this study, we summarized the potential relationship between lipid molecules with different structural features and LC risk, identified a panel of LC biomarkers, and demonstrated that the n-3 PUFA of the acyl chain of lipids was a protective factor for LC. [ABSTRACT FROM AUTHOR]

The human face is considered the prime entity in recognizing a person's identity in our society. Henceforth, the importance of face recognition systems is growing higher for many applications. Facial recognition systems are in huge demand, next to fingerprint-based systems. Face-biometric has a highly dominant role in various applications such as border surveillance, forensic investigations, crime detection, access management systems, information security, and many more. Facial recognition systems deliver highly meticulous results in every of these application domains. However, the face identity threats are evenly growing at the same rate and posing severe concerns on the use of face-biometrics. This paper significantly explores all types of face recognition techniques, their accountable challenges, and threats to face-biometric-based identity recognition. This survey paper proposes a novel taxonomy to represent potential face identity threats. These threats are described, considering their impact on the facial recognition system. State-of-the-art approaches available in the literature are discussed here to mitigate the impact of the identified threats. This paper provides a comparative analysis of countermeasure techniques focusing on their performance on different face datasets for each identified threat. This paper also highlights the characteristics of the benchmark face datasets representing unconstrained scenarios. In addition, we also discuss research gaps and future opportunities to tackle the facial identity threats for the information of researchers and readers. [ABSTRACT FROM AUTHOR]

In this work, we operate on a small dataset available from the technical literature to predict the oxidation-induced mass change at 1000 °C of thousands of new alloy compositions using "Tree-based Pipeline Optimization Tool" , an automated machine learning (ML) method. The ML pipeline we develop is trained on the log10 of the mass change per unit area. This yields a mean absolute error of 0.34 on the test set's values, which span 3.5 decades. With additional insights from thermodynamic simulations, a set of seven alloys is selected, manufactured, and characterized. Of these, the oxidation behavior of five alloys is well-predicted by the ML-based model, while results for two alloys show orders of magnitude deviations from predictions. The results show that ML-based methods can be useful for predicting composition-dependent oxidation behavior, despite its many complexities. [ABSTRACT FROM AUTHOR]

This paper presents the behavior of a 102-year-old truss bridge under wind loading. To examine the wind-related responses of the historical bridge, state-of-the-art and traditional modeling methodologies are employed: a machine learning approach called random forest and three-dimensional finite element analysis. Upon training and validating these modeling methods using experimental data collected from the field, member-level forces and stresses are predicted in tandem with wind speeds inferred by Weibull distributions. The intensities of the in-situ wind are dominated by the location of sampling, and the degree of partial fixities at the supports of the truss system is found to be insignificant. Compared with quadrantal pressure distributions, uniform pressure distributions better represent the characteristics of wind-induced loadings. The magnitude of stress in the truss members is enveloped by the stress range in line with the occurrence probabilities of the characterized wind speed between 40% and 60%. The uneven wind distributions cause asymmetric displacements at the supports. [ABSTRACT FROM AUTHOR]

Multi-view clustering is a prominent area of interest in machine learning and data mining. However, most existing methods are confined to static multi-view data, posing challenges for achieving multi-view information fusion and clustering in a dynamic streaming context. We propose a multi-view streaming clustering with incremental anchor learning, which effectively partitions continuous chunks of multi-view data into meaningful clusters. Initially, a shared subspace representation is derived to reveal the intrinsic structure hidden across views, which is adapted to the evolving data distribution through incremental learning of anchors. Furthermore, the shared subspace representation, anchors, and clustering assignments are learned simultaneously in a unified framework, where their interactive negotiation avoids the suboptimal solution problem and significantly enhances overall clustering performance. Finally, extensive experiments on several real-world datasets demonstrate that the proposed method achieves superior multi-view clustering performance and efficiency in a streaming context. [ABSTRACT FROM AUTHOR]

Presently, cognitive Internet of Things (CIoT) with cloud computing (CC) enabled intelligent healthcare models are developed, which enables communication with intelligent devices, sensor modules, and other stakeholders in the healthcare sector to avail effective decision making. On the other hand, Alzheimer disease (AD) is an advanced and degenerative illness which injures the brain cells, and its earlier detection is necessary for suitable interference by healthcare professional. In this aspect, this paper presents a new Oriented Features from Accelerated Segment Test (FAST) with Rotated Binary Robust Independent Elementary Features (BRIEF) Detector (ORB) with optimal artificial neural network (ORB-OANN) model for AD diagnosis and classification on the CIoT based smart healthcare system. For initial pre-processing, bilateral filtering (BLF) based noise removal and region of interest (RoI) detection processes are carried out. In addition, the ORB-OANN model includes ORB based feature extractor and principal component analysis (PCA) based feature selector. Moreover, artificial neural network (ANN) model is utilized as a classifier and the parameters of the ANN are optimally chosen by the use of salp swarm algorithm (SSA). A comprehensive experimental analysis of the ORB-OANN model is carried out on the benchmark database and the obtained results pointed out the promising outcome of the ORB-OANN technique in terms of different measures. [ABSTRACT FROM AUTHOR]

Prediction of human physical activities has become a necessity for some applications that come with the development of wearable and portable hardware such as smartwatches and smartphones. The task of Human Activity Recognition (HAR) is to recognize human physical activities, e.g., walking, sitting, and running, using the data collected from sensors, e.g., accelerometers and gyroscope. HAR is commonly applied on smart systems, such as smartphones, to serve the understanding of a user's behaviors and provide assistance to the user because of the rapid development of ubiquitous computing technology in recent years. Thus, predicting activities, such as standing, walking, sitting, during the day have become a popular topic in machine and deep learning. The aim of this study is to predict the user's activities based on context information gathered by sensors such as gyroscopes and accelerometers. The conducted classification algorithms extract features from training data and learn a classification model based on the features to predict activity. In this paper, various classical machine and deep learning techniques have been studied and compared for human activity recognition. A comparative analysis is performed between techniques in order to select the classifier with the best recognition performance. Experimental results show that established Deep Neural Network (DNN) model achieved an accuracy of up to 96.81% and mean absolute error of up to 0.03 on publicly available UCI-HAR dataset. This method has given the best performance between conducted classification methods in this study to predict human activity. [ABSTRACT FROM AUTHOR]

Recommender systems are widely used to provide users with recommendations based on their preferences. With the ever-growing volume of information online, recommender systems have been a useful tool to overcome information overload. The utilization of recommender systems cannot be overstated, given its potential influence to ameliorate many over-choice challenges. There are many types of recommendation systems with different methodologies and concepts. Various applications have adopted recommendation systems, including e-commerce, healthcare, transportation, agriculture, and media. This paper provides the current landscape of recommender systems research and identifies directions in the field in various applications. This article provides an overview of the current state of the art in recommendation systems, their types, challenges, limitations, and business adoptions. To assess the quality of a recommendation system, qualitative evaluation metrics are discussed in the paper. [ABSTRACT FROM AUTHOR]

The accuracy and diversity of recommendation algorithms have always been the research hotspot of recommender systems. A good recommender system should not only have high accuracy and diversity, but also have adequate robustness against spammer attacks. However, the issue of recommendation robustness has received relatively little attention in the literature. In this paper, we systematically study the influences of different spammer behaviors on the recommendation results in various recommendation algorithms. We further propose an improved algorithm by incorporating the inner-similarity of user’s purchased items in the classic KNN approach. The new algorithm effectively enhances the robustness against spammer attacks and thus outperforms traditional algorithms in recommendation accuracy and diversity when spammers exist in the online commercial systems. [ABSTRACT FROM AUTHOR]

Recently, many online social networks, such as MySpace, Orkut, and Friendster, have faced inactivity decay of their members, which contributed to the collapse of these networks. The reasons, mechanics, and prevention mechanisms of such inactivity decay are not fully understood. In this work, we analyze decayed and alive subwebsites from the Stack Exchange platform. The analysis mainly focuses on the inactivity cascades that occur among the members of these communities. We provide measures to understand the decay process and statistical analysis to extract the patterns that accompany the inactivity decay. Additionally, we predict cascade size and cascade virality using machine learning. The results of this work include a statistically significant difference of the decay patterns between the decayed and the alive subwebsites. These patterns are mainly cascade size, cascade virality, cascade duration, and cascade similarity. Additionally, the contributed prediction framework showed satisfactorily prediction results compared to a baseline predictor. Supported by empirical evidence, the main findings of this work are (1) there are significantly different decay patterns in the alive and the decayed subwebsites of the Stack Exchange; (2) the cascade’s node degrees contribute more to the decay process than the cascade’s virality, which indicates that the expert members of the Stack Exchange subwebsites were mainly responsible for the activity or inactivity of the Stack Exchange subwebsites; (3) the Statistics subwebsite is going through decay dynamics that may lead to it becoming fully-decayed; (4) the decay process is not governed by only one network measure, it is better described using multiple measures; (5) decayed subwebsites were originally less resilient to inactivity decay, unlike the alive subwebsites; and (6) network’s structure in the early stages of its evolution dictates the activity/inactivity characteristics of the network. [ABSTRACT FROM AUTHOR]

Complementing the formal organizational structure of a business are the informal connections among employees. These relationships help identify knowledge hubs, working groups, and shortcuts through the organizational structure. They carry valuable information on how a company functions de facto. In the past, eliciting the informal social networks within an organization was challenging; today they are reflected by friendship relationships in online social networks. In this paper we analyze several commercial organizations by mining data which their employees have exposed on Facebook, LinkedIn, and other publicly available sources. Using a web crawler designed for this purpose, we extract a network of informal social relationships among employees of targeted organizations. Our results show that it is possible to identify leadership roles within the organization solely by using centrality analysis and machine learning techniques applied to the informal relationship network structure. Valuable non-trivial insights can also be gained by clustering an organization's social network and gathering publicly available information on the employees within each cluster. Knowledge of the network of informal relationships may be a major asset or might be a significant threat to the underlying organization. [ABSTRACT FROM AUTHOR]

Social photo sharing platforms on the Internet (e.g. Flickr) host billions of publicly accessible photos captured by millions of individual users from all over the world. These user-contributed and geo-tagged photo collections provide insights into human sociocultural life and provide important clues for understanding people's engagement and reaction to places and events around the world today. In this paper, we analyze over 2 million geo-tagged images uploaded by 12,000 individual Flickr users to investigate the photograph shooting patterns of different user groups; that is, tourist and local, Asian and European, and male and female users. Specifically, we make use of visual features extracted on single monocular images and their spatial configurations to infer 3D depth information of the photographs to establish the preferred shooting scale (close-up or far-distant) of the user groups. The results reveal which objects and scenes interest different groups of people and how these preferences change over space and time. As such, the research offers a new approach to the human sciences which study the individual, groups and society. [ABSTRACT FROM AUTHOR]

Legumes are a better source of proteins and are richer in diverse micronutrients over the nutritional profile of widely consumed cereals. However, when exposed to a diverse range of abiotic stresses, their overall productivity and quality are hugely impacted. Our limited understanding of genetic determinants and novel variants associated with the abiotic stress response in food legume crops restricts its amelioration. Therefore, it is imperative to understand different molecular approaches in food legume crops that can be utilized in crop improvement programs to minimize the economic loss. 'Omics'-based molecular breeding provides better opportunities over conventional breeding for diversifying the natural germplasm together with improving yield and quality parameters. Due to molecular advancements, the technique is now equipped with novel 'omics' approaches such as ionomics, epigenomics, fluxomics, RNomics, glycomics, glycoproteomics, phosphoproteomics, lipidomics, regulomics, and secretomics. Pan-omics—which utilizes the molecular bases of the stress response to identify genes (genomics), mRNAs (transcriptomics), proteins (proteomics), and biomolecules (metabolomics) associated with stress regulation—has been widely used for abiotic stress amelioration in food legume crops. Integration of pan-omics with novel omics approaches will fast-track legume breeding programs. Moreover, artificial intelligence (AI)-based algorithms can be utilized for simulating crop yield under changing environments, which can help in predicting the genetic gain beforehand. Application of machine learning (ML) in quantitative trait loci (QTL) mining will further help in determining the genetic determinants of abiotic stress tolerance in pulses. [ABSTRACT FROM AUTHOR]