Search

Present report demonstrates the application of Sb 0.3 Sn 0.7 Se 2 single crystal as a paper based flexible photodetector. Direct vapour transport grown bulk crystals of Sb 0.3 Sn 0.7 Se 2 has been converted to nanosheets by chemical assisted exfoliation process. The paper-based photodetector is fabricated and the switching action is studied. Tuning of photodetector has been carried out by 670 nm laser illumination for different bias voltages. Temporal photo-response is also studied under polychromatic light with different intensities in vacuum and open environment. The low temperature stability of photodetector has been studied for temperature range 300 K–180 K. Experimental results are obtained in terms of time resolved photocurrent under different illumination and atmospheric conditions. The flexibility and stability are also examined in detail for fabricated detector. Overall, the results suggest the application of Sb 0.3 Sn 0.7 Se 2 as a versatile flexible photodetector. Additionally, the machine learning (ML) model is trained and tested using an experimental photocurrent dataset that has a complex material design with variations in time, bias voltage, intensity, and temperature. The k-nearest neighbor algorithm exhibited outstanding performance, achieving the highest R2 value of 0.9986 when applied to a temperature dataset, with a test size of 0.4. Performance metrics such as mean absolute error and root mean squared error of various test sizes ranging from 0.4 to 0.6 are used to assess the model's accuracy and robustness in changing conditions. This comprehensive analysis not only establishes a platform for future experimental optimization of photodetector materials but also underscores the efficacy of ML regression techniques in developing high-performance photodetectors. [Display omitted] • Chemical assisted exfoliation process were used for the synthesis of SbSnSe2 nanosheets. • Synthesized nanosheets were decorated on paper for the fabrication of flexible photo-detector. • The photodetection application was demonstrated for different atmospheric conditions and different illuminations. • Low temperature stability of the detector was also studied. • KNN Machine learning model were applied to the experimental data for the prediction of photodetection behavior. [ABSTRACT FROM AUTHOR]

• Five Machine Learning (ML) models compared for wave height time series prediction. • Complex ML models do not outperform simple AR in wave height time series prediction. • Comment to related papers: signal decomposition in test set series is WRONG. Significant Wave Height (SWH) is crucial in many aspect of ocean engineering. The accurate prediction of SWH has therefore been of immense practical value. Recently, Artificial Intelligence (AI) time series prediction methods have been widely used for single-point short-term SWH time-series forecasting, resulting in many AI-based models claiming to achieve good results. However, the extent to which these complex AI models can outperform traditional methods has largely been overlooked. This study compared five different models - AutoRegressive (AR), eXtreme Gradient Boosting (XGB), Artificial Neural Network (ANN), Long Short-Term Memory (LSTM), and WaveNet - for their performance on SWH time series prediction at 16 buoy locations. Surprisingly, the results suggest that the differences of performance among different models are negligible, indicating that all these AI models have only "learned" the linear auto-regression from the data. Additionally, we noticed that many recent studies used signal decomposition method for such time series prediction, and most of them decomposed the test sets, which is WRONG. [ABSTRACT FROM AUTHOR]

[Display omitted] • 3D-ACM involves two rounds of Hilbert transform and tensor product operations. • It significantly enhances the equivalent frequency points and sample numbers. • The R-squared values for PLS-LR, KNN, RF and CNN models approach or equal 1. • These supervised models are comparable to unsupervised models such as PCA-LR. We have developed a novel 3D asynchronous correlation method (3D-ACM) designed for the classification and identification of Chinese handmade paper samples using Raman spectra and machine learning. The 3D-ACM approach involves two rounds of tensor product and Hilbert transform operations. In the tensor product process, the outer product of the spectral data from different samples within the same category is computed, establishing inner connections among all samples within that category. The Hilbert transform introduces a 90-degree phase shift, resulting in a true three-dimensional spectral data structure. This expansion significantly increases the number of equivalent frequency points and samples within each category. This enhancement substantially boosts spectral resolution and reveals more hidden information within the spectral data. To maximize the potential of 3D-ACM, we employed six machine learning models: principal component analysis (PCA) with linear regression (LR), support vector machine (SVM) with LR, k-Nearest Neighbors (KNN), random forest (RF), and convolutional neural network (CNN). When applied to the 3D-ACM data preprocessing method, R-squared values of PLS-LR, KNN, RF and CNN supervised models, approached or equaled 1. This indicates exceptional performance comparable to unsupervised models like PCA. 3D-ACM stands as a versatile mathematical technique not confined to spectral data. It also eliminates the necessity for additional experimental setups or external control conditions, distinct from traditional two-dimensional correlation spectroscopy. Moreover, it preserves the original experimental data, setting it apart from conventional data preprocessing methods. This positions 3D-ACM as a promising tool for future material classification and identification in conjunction with machine learning. [ABSTRACT FROM AUTHOR]

• Qualitative and quantitative detection of amine gases could be achieved by CSA. • A visible detection of beef freshness using the amine-responsive CSA was proposed. • ResNet34 had the best performance for beef freshness detection based on CSA. • T-SNE could further visualize and understand the classification process of DL. This study developed an innovative approach that combines a colourimetric sensor array (CSA) composed of twelve pH-response dyes with advanced algorithms, aiming to detect amine gases and assess the freshness of chilled beef. With the assistance of multivariate statistical analysis, the sensor array can effectively distinguish five amine gases and enable rapid quantification of trimethylamine vapour with a limit of detection (LOD) of 8.02 ppb and visually monitor the fresh levels of chilled beef. Moreover, the utilization of deep learning models (ResNet34, VGG16, and GoogleNet) for chilled beef freshness evaluation achieved an overall accuracy of 98.0 %. Furthermore, t -distributed stochastic neighbour embedding (t -SNE) visualized the feature extraction process and provided explanations to understand the classification process of deep learning. The results demonstrated that applying deep learning techniques in the process of pattern recognition of CSA can help in realizing the rapid, robust, and accurate assessment of chilled beef freshness. [ABSTRACT FROM AUTHOR]

Global food systems can benefit significantly from continuous monitoring of microbial food safety, a task for which tedious operations, destructive sampling, and the inability to monitor multiple pathogens remain challenging. This study reports significant improvements to a paper chromogenic array sensor - machine learning (PCA-ML) methodology sensing concentrations of volatile organic compounds (VOCs) emitted on a species-specific basis by pathogens by streamlining dye selection, sensor fabrication, database construction, and machine learning and validation. This approach enables noncontact, time-dependent, simultaneous monitoring of multiple pathogens (Listeria monocytogenes , Salmonella , and E. coli O157:H7) at levels as low as 1 log CFU/g with over 90% accuracy. The report provides theoretical and practical frameworks demonstrating that chromogenic response, including limits of detection, depends on time integrals of VOC concentrations. The paper also discusses the potential for implementing PCA-ML in the food supply chain for different food matrices and pathogens, with species- and strain-specific identification. [ABSTRACT FROM AUTHOR]

• A machine learning approach was used to investigate moisture penetration in paper materials. • The study evaluated the capabilities of both the Feedforward Neural Network (FNN) and the Recurrent Neural Network (RNN). • Numerically generated data was employed for network training. • A continuum model incorporating anisotropic properties, creep behavior, viscoelasticity, and moisture dependency was used for simulations. • The RNN demonstrated superior predictive capability compared to the FNN. In this work, we predicted the gradient of the deformational moisture dynamics in a sized commercial paper by observing the curl deformation in response to the one-sided water application. The deformational moisture is a part of the applied liquid which ends up in the fibers causing swelling and subsequent mechanical response of the entire fiber network structure. The adapted approach combines traditional experimental procedures, advanced machine learning techniques and continuum modeling to provide insights into the complex phenomenon relevant to ink-jet digital printing in which the sized and coated paper is often used, meaning that not all the applied moisture will reach the fibers. Key material properties including elasticity, plastic parameters, viscoelasticity, creep, moisture dependent behavior, along with hygroexpansion coefficients are identified through extensive testing, providing vital data for subsequent simulation using a continuum model. Two machine learning models, a Feedforward Neural Network (FNN) and a Recurrent Neural Network (RNN), are probed in this study. Both models are trained using exclusively numerically generated moisture profile histories, showcasing the value of such data in contexts where experimental data acquisition is challenging. These two models are subsequently utilized to predict moisture profile history based on curl experimental measurements, with the RNN demonstrating superior accuracy due to its ability to account for temporal dependencies. The predicted moisture profiles are used as inputs for the continuum model to simulate the associated curl response comparing it to the experiment representing "never seen" data. The result of comparison shows highly predictive capability of the RNN. This study melds traditional experimental methods and innovative machine learning techniques, providing a robust technique for predicting moisture gradient dynamics that can be used for both optimizing the ink solution and paper structure to achieve desirable printing quality with lowest curl propensities during printing. [ABSTRACT FROM AUTHOR]

Antibiotic residues accumulation in the environment endangers ecosystems and human health. There is an urgent need for a facile and efficient strategy to detect antibiotics. Here, we report a photoluminescent sensor array based on protein-stabilized gold nanoclusters (AuNCs) for the detection of two families of antibiotics, tetracyclines and quinolones. The nanoclusters were synthesized with bovine serum albumin (BSA) and ovalbumin (OVA), respectively. They had different interactions with seven kinds of antibiotics and exhibited diverse photoluminescence (PL) responses, which were analyzed by linear discriminant analysis and ExtraTrees algorithms. The sensor array performed well in both classification and quantification of seven antibiotics. And the quantitative results of all antibiotics obtained R2 of no less than 0.99 at 0–100 μM when using suitable regression models. Additionally, the sensor array was able to distinguish antibiotic mixtures and multiple interfering substances, and it also kept 100% classification accuracy in river water samples. Moreover, test paper assisted by a smartphone was applied for quick detection of antibiotics, with good performance in both HEPES buffer and river water. These studies reveal great potential for the point-of-use analysis of antibiotics in environmental monitoring. [Display omitted] • The first nanoclusters-based photoluminescent sensor array for quinolones is developed. • Seven antibiotics of tetracyclines and quinolones are successfully identified. • ExtraTrees algorithm is first employed for antibiotics detection. • Test paper assisted by a smartphone is used to identify antibiotics in river water. [ABSTRACT FROM AUTHOR]

Efficient control loop performance is pivotal in process industries to ensure optimal production, maintain product quality, and adhere to regulatory standards. Poorly tuned controllers can disrupt these objectives, necessitating accurate detection methods. This paper introduces a novel approach for detecting poor controller tuning through advanced techniques: the Gramian Angular Field (GAF) and Stack Auto-Encoder (SAE). Unlike manual methods, this automated system promptly identifies poorly tuned controllers, offering real-time monitoring and timely alerts to operators. The proposed methodology is substantiated through two case studies: the ISDB dataset and the pulp and paper dataset. The outcomes illustrate that the proposed approach correctly determines the appropriate outcome for the majority of the analyzed control loops across diverse industries. • New method detects poorly tuned controllers via Gramian angular field and SAE. • PV and OP images help SAE distinguish poor tuning from other oscillation causes. • Transfer learning has been used to improve methodology's effectiveness. • Tested on benchmark control loops, yielding accurate verdicts for most cases. [ABSTRACT FROM AUTHOR]

A label-free detection method for noninvasive biofluids enables rapid on-site disease screening and early-stage cancer diagnosis by analyzing metabolic alterations. Herein, we develop three-dimensional plasmonic hexaplex nanostructures coated on a paper substrate (3D-PHP). This flexible and highly absorptive 3D-PHP sensor is integrated with commercial saliva collection tube to create an efficient on-site sensing platform for lung cancer screening via surface-enhanced Raman scattering (SERS) measurement of human saliva. The multispike hexaplex-shaped gold nanostructure enhances contact with saliva viscosity, enabling effective sampling and SERS enhancement. Through testing patient salivary samples, the 3D-PHP sensor demonstrates successful lung cancer detection and diagnosis. A logistic regression-based machine learning model successfully classifies benign and malignant patients, exhibiting high clinical sensitivity and specificity. Additionally, important Raman peak positions related to different lung cancer stages are investigated, suggesting insights for early-stage cancer diagnosis. Integrating 3D-PHP senor with the conventional saliva collection tube platform is expected to offer promising practicality for rapid on-site disease screening and diagnosis, and significant advancements in cancer detection and patient care. [Display omitted] [ABSTRACT FROM AUTHOR]

• 252 chemicals were identified by SNTS in groundwater from four regions with diverse contamination histories. • A novel and robust systematic machine learning-based workflow for predicting chemical indicators was proposed. • The proposed workflow showed good predictive ability (Q2) of 0.897. • 51 chemical indicators for tracking groundwater contamination sources were suggested. In this study, the strong analytical power of gas chromatography coupled to a high resolution mass spectrometry (GC-HRMS) in suspect and non-target screening (SNTS) of organic micropollutants was combined with machine learning tools for proposing a novel and robust systematic environmental forensics workflow, focusing on groundwater contamination. Groundwater samples were collected from four different regions with diverse contamination histories (namely oil [OC], agricultural [AGR], industrial [IND], and landfill [LF]), and a total of 252 organic micropollutants were identified, including pharmaceuticals, personal care products, pesticides, polycyclic aromatic hydrocarbons, plasticizers, phenols, organophosphate flame retardants, transformation products, and others, with detection frequencies ranging from 3 % to 100 %. Amongst the SNTS identified compounds, a total of 51 chemical indicators (i.e., OC: 13, LF: 12, AGR: 19, IND: 7) which included level 1 and 2 SNTS identified chemicals were pinpointed across all sampling regions by integrating a bootstrapped feature selection method involving the bootfs algorithm and a partial least squares discriminant analysis (PLS-DA) model to determine potential prevalent contamination sources. The proposed workflow showed good predictive ability (Q2) of 0.897, and the suggested contamination sources were gasoline, diesel, and/or other light petroleum products for the OC region, anthropogenic activities for the LF region, agricultural and human activities for the AGR region, and industrial/human activities for the IND region. These results suggest that the proposed workflow can select a subset of the most diagnostic features in the chemical space that can best distinguish a specific contamination source class. [Display omitted] [ABSTRACT FROM AUTHOR]

It is critical to have an accurate prediction of surface roughness (Sa) in order to improve grinding productivity, reduce costs, and minimize the period of time required for trials and testing. Although many prediction methods have been developed, fewer studies have been conducted on the prediction of surface roughness in longitudinal ultrasonic vibration-assisted grinding (LUVAG). In this paper, a surface roughness prediction model algorithm based on ensemble learning of support vector machines (ELSVM) is proposed that can be used for surface roughness prediction of LUVAG alumina ceramics. This paper first details the development of the ELSVM surface roughness prediction model, which consists of four modules: the prepossessing module, the multi-algorithm regression module, the support vector machine algorithm (SVM) module, and the ensemble module. In addition, ELSVM was compared with four other machine learning methods based on experimental results for surface roughness prediction modeling. The error of ELSVM model was reduced by 6.3%, 7.9%, 8.9%, and 7.5%, respectively, compared to the individual prediction models such as I-AISPSO, I-AIS, SPSO, and KBaNN. Comparison results show that the ELSVM model has the lowest average ratio of Mean Absolute Error (MAE) for surface roughness prediction in LUVAG. • The artificial immune particle swarm optimization is improved. • The adaptive particle swarm optimization algorithm and the manual free algorithm are combined in parallel. • A surface roughness prediction model of support vector machine based on ensemble learning algorithm is proposed. • Longitudinal ultrasonic assisted grinding can reduce surface roughness. [ABSTRACT FROM AUTHOR]

• A model to explore the multiphysics characteristic of semiconductors is presented. • The model determines the thermal and mechanical properties of complex patterns. • Results show a strong correlation between thermal and mechanical properties. • Dielectric bands asymmetrically affect thermal and mechanical behavior. • Machine learning can predict one property from another. As semiconductor performance improves through advanced package designs, it becomes important to consider both thermal and mechanical properties. Better understanding their relationship enhances system design and optimization. This study has introduced a numerical method that takes into account both thermal and mechanical fluxes to construct thermal and mechanical property maps for practical application in semiconductor engineering. Furthermore, this paper investigated the relationship between the two properties using the constructed property maps in commercially available semiconductor packages. Owing to the complexity of packaging design patterns, a coupled isoparametric mapping and machine learning (ML) method was introduced to investigate their effects. The model then determines and investigates the anisotropic equivalent thermal and mechanical properties of the commercialized semiconductor packages with complex pattern designs, according to the package materials, volume fractions of each material, and design patterns. This approach pursues more sophistication and practicality compared to simplified composite structure property evaluation models. Additionally, all processes in the algorithm are automated based on ML methods, making them practically applicable in the semiconductor industry. The study shows that there is a strong correlation between the thermal and mechanical characteristics in the complex package patterns. This relationship was able to form a fairly clear category based on the shape of the dielectric band because the dielectric band had significantly different effects on the thermal and mechanical fluxes. The study discussed the physical implications of the similarities and differences between the two behaviors and validated the results of the equivalent properties through finite element (FE) analysis. Finally, the study cross-validates the relationship by showing that information from one flux behavior can be used to predict the other based on the ML method. Based on the results, this paper can provide a better understanding of thermal and mechanical fluxes in semiconductor package patterns for package design and optimization. [ABSTRACT FROM AUTHOR]

This paper uses machine learning approaches to predict the association between traffic volume, air pollution, and meteorological conditions. A key focus is on the interaction between these factors. The paper does this using hourly traffic volume, N O x , P M 2.5 , and weather data for Oslo, Norway. I considered a total of six datasets of the 2019 whole-year data to verify the prediction accuracy of the models. I find that the autoregressive integrated moving average model with exogenous input variables, and the autoregressive moving average dynamic linear model outperform the machine learning models in predicting air pollution. At the same time, I also explored the effect of sampling weather subsets on prediction accuracy. Finally, my study makes optimal policy recommendations for reducing air pollution from traffic volume, after considering the interaction and lagged effects of meteorology, time variables, traffic, and air pollution. [ABSTRACT FROM AUTHOR]

Advances in artificial intelligence (AI) and automated robotics will profoundly influence space operations. By utilising machine learning and deep learning approaches, AI-enabled systems may accomplish tasks as well as improve their own performance. These capabilities are useful in the often-remote settings of outer space and will grow in value as automated space operations become more widespread. As AI extends throughout the space domain, automated algorithms will take on many of the roles that have historically been handled by humans. Artificial intelligence is progressing from theory to implementation in the space environment by exposing new satellites and orbital autonomous vehicles to new data. Even though all initial computational parameters are provided, such systems' outputs can be very unpredictable, putting people, property, and the environment at risk. This paper investigates the application of United Nations space treaties, selected regional AI regulations, and various 'soft-law' instruments and industry initiatives focusing on responsible AI system development to space-based AI systems. Following that, reforms are proposed to clarify the practical relationship between AI systems and the international legal regime that governs space, as well as a 'bottom-up' regulatory approach to better facilitate the future development of regulation governing the use of AI by the global space sector. While this work does not purport to provide a conclusive resolution to these multifaceted matters, its objective is to underscore significant obstacles that arise at the convergence of space law and AI, serving as a preliminary foundation for subsequent discussions on this issue. • Advances in AI and automated robotics will have a profound impact on space operations. Automated algorithms will take on roles traditionally handled by humans as AI becomes more widespread in space. • However, the unpredictable outputs of AI systems can put people, property, and the environment at risk, raising questions about liability. • The paper investigates the application of UN space treaties, regional AI regulations, and industry initiatives to space-based AI systems. • Reforms are proposed to clarify the relationship between AI systems and the international legal regime governing space. • A 'bottom-up' regulatory approach is suggested to facilitate future regulation of AI in the global space sector. [ABSTRACT FROM AUTHOR]

With the rapid advancement of 3D sensors, there is an increasing demand for 3D scene understanding and an increasing number of 3D deep learning algorithms have been proposed. However, a large-scale and richly annotated 3D point cloud dataset is critical to understanding complicated road and urban scenes. Motivated by the need to bridge the gap between the rising demand for 3D urban scene understanding and limited LiDAR point cloud datasets, this paper proposes a richly annotated WHU-Urban3D dataset and an effective method for semantic instance segmentation. WHU-Urban3D stands out from existing datasets due to its distinctive features: (1) extensive coverage of both Airborne Laser Scanning and Mobile Laser Scanning point clouds, along with panoramic images; (2) containing large-scale road and urban scenes in different cities (over 3. 2 × 1 0 6 m 2 area), with richly point-wise semantic instance labels (over 200 million points); (3) inclusion of particular attributes (e.g., reflected intensity, number of returns) in addition to 3D coordinates. This paper also provides the performance of several representative baseline methods and outlines potential future works and challenges for fully exploiting this dataset. The WHU-Urban3D dataset is publicly accessible at https://whu3d.com/. [ABSTRACT FROM AUTHOR]

With the rapidly evolving geo-political landscape and unceasing advancements in technology, sustainable energy security is a very important topic. Research indicates that electrochemical energy systems are quite promising to solve many of energy conversion, storage, and conservation challenges while offering high efficiencies and low pollution. The paper provides an overview of electrochemical energy devices and the various optimization techniques used to evaluate them. The optimization techniques include linear regression, factorial design, the Taguchi method, artificial neural networks, filters, and a combination of such methods to improve these systems. To support the growing interest in research, the bulk of this study focuses on a review of the most promising and highly researched electrochemical energy devices, such as fuel cells, batteries, and supercapacitors. The paper also provides modest commentary on hydrogen production technologies, electrochemical reactors, and membrane separation technologies, amongst other technologies. Building on a previous paper by the authors of this paper on artificial intelligence in hybrid renewable energy systems with fuel cells, this work provides a comparative review of optimization techniques for supercapacitors by highlighting key findings based on model accuracy. A summary of the advantages and disadvantages of the different major optimization techniques is presented. The paper concludes that a combination of optimization techniques is used to overcome the drawbacks of individual techniques, with adaptive filters being the most widely studied. This paper presents studies on the Design of Experiments (DoE) with the goal of building a better understanding of the relationships that exist between different operating variables in various electrochemical devices. • The electrochemical systems can be optimized for better performance with high accuracy. • The optimization techniques used are either model-based or data-driven. • Linear regression and artificial neural networks are the most common techniques studied. • Hybrid models could be applied to enhance systems' accuracy and minimize errors. [ABSTRACT FROM AUTHOR]

Large deep learning models are impressive, but they struggle when real-time data is not available. Few-shot class-incremental learning (FSCIL) poses a significant challenge for deep neural networks to learn new tasks from just a few labeled samples without forgetting the previously learned ones. This setup can easily leads to catastrophic forgetting and overfitting problems, severely affecting model performance. Studying FSCIL helps overcome deep learning model limitations on data volume and acquisition time, while improving practicality and adaptability of machine learning models. This paper provides a comprehensive survey on FSCIL. Unlike previous surveys, we aim to synthesize few-shot learning and incremental learning, focusing on introducing FSCIL from two perspectives, while reviewing over 30 theoretical research studies and more than 20 applied research studies. From the theoretical perspective, we provide a novel categorization approach that divides the field into five subcategories, including traditional machine learning methods, meta learning-based methods, feature and feature space-based methods, replay-based methods, and dynamic network structure-based methods. We also evaluate the performance of recent theoretical research on benchmark datasets of FSCIL. From the application perspective, FSCIL has achieved impressive achievements in various fields of computer vision such as image classification, object detection, and image segmentation, as well as in natural language processing and graph. We summarize the important applications. Finally, we point out potential future research directions, including applications, problem setups, and theory development. Overall, this paper offers a comprehensive analysis of the latest advances in FSCIL from a methodological, performance, and application perspective. • In-depth survey on FSCIL methods, applications, performance. • Categorizing FSCIL into five subcategories for clear analysis. • Evaluating FSCIL research on benchmarks for strengths, weaknesses. • FSCIL applications in computer vision, NLP, and graph. [ABSTRACT FROM AUTHOR]

Hydrogen is crucial in the forthcoming low-carbon energy systems offering the potential to reduce CO 2 emissions and its versatility to be utilized across diverse energy sectors. Precise assessment of hydrogen thermo-physical characteristics is vital for the effective design and execution of numerous processes encompassing hydrogen production, transportation, storage, and utilization. In this paper, various distinct machine learning (ML) approaches, namely extreme gradient boosting (XGBoost), Random Forest (RF), Adaptive boosting (AdaBoost), and Light gradient boosting machine (LightGBM) were implemented to estimate thermo-physical properties, namely thermal conductivity, density, viscosity, and diffusivity in water based on pressure and temperature variables. For this purpose, comprehensive experimental data points covering a broad range of pressures and temperatures were acquired from the literature. The results demonstrated a strong agreement between the predictions generated by all proposed techniques and the experimental data. Furthermore, XGBoost yielded a root mean square error (RMSE) of 0.0085, 2.1548, 0.3343, and 1.5308 for the estimation of thermal conductivity, density, viscosity, and diffusivity, respectively, demonstrating superior accuracy in predicting all outcomes. In the sensitivity evaluation, it was observed that temperature, with absolute relevancy factor 0.88, 0.54, 0.246, and 0.92 had the most significant influence on the hydrogen thermo-physical properties in the order mentioned above. Furthermore, the trend analysis of the model indicated that all thermo-physical properties of hydrogen experience positive effects from temperature and pressure, except for density, which decreases as temperature increases. The leverage technique confirmed the statistical validity of both the experimental dataset used for modeling and the model's development. This study's results provide the efficient design and secure operation of hydrogen storage, refueling stations, transportation infrastructure, and production facilities. [Display omitted] • Hydrogen holds great importance in the forthcoming low-carbon energy. • XGBoost, RF, AdaBoost, and LightGBM are utilized for modeling the thermo-physical properties of hydrogen. • XGBoost demonstrates its superior accuracy in predicting thermal conductivity, density, viscosity, and diffusivity. • Temperature has the most significant influence on the hydrogen thermo-physical properties. [ABSTRACT FROM AUTHOR]

Due to the coupling of multiple physical fields inside proton exchange membrane fuel cell (PEMFC) and the complex and changeable external application scenarios, the faults of PEMFC system greatly affects its engineering application. The paper proposes an ensemble learning-based fault diagnosis method for PEMFC systems, utilizing five distinct algorithms to develop an integrated ensemble classifier for fault detection. The original sample dataset collected from fuel cells undergoes a process of deep feature extraction to obtain a composite feature sample dataset, which is then inputted into an ensemble classifier to establish a fault diagnosis model. The experimental data of four kinds of faults of 80 W fuel cell test bench, EC fuel cell system and 50 W water-cooled fuel cell test bench are analyzed by using this classification model. The results show that the accuracy and generalization of the ensemble learning algorithm for different fault diagnosis of PEMFC system in different application environments are verified. • Adopt multiple classification algorithms for fault modeling and diagnosis. • Good applicability to different faults is shown in different test scenarios. • Compared with a single algorithm, it has higher fault identification accuracy. [ABSTRACT FROM AUTHOR]

Airborne hyperspectral images can be used to map the land cover in large urban areas, thanks to their very high spatial and spectral resolutions on a wide spectral domain. While the spectral dimension of hyperspectral images is highly informative of the chemical composition of the land surface, the use of state-of-the-art machine learning algorithms to map the land cover has been dramatically limited by the availability of training data. To cope with the scarcity of annotations, semi-supervised and self-supervised techniques have lately raised a lot of interest in the community. Yet, the publicly available hyperspectral data sets commonly used to benchmark machine learning models are not totally suited to evaluate their generalization performances due to one or several of the following properties: a limited geographical coverage (which does not reflect the spectral diversity in metropolitan areas), a small number of land cover classes and a lack of appropriate standard train/test splits for semi-supervised and self-supervised learning. Therefore, we release in this paper the Toulouse Hyperspectral Data Set that stands out from other data sets in the above-mentioned respects in order to meet key issues in spectral representation learning and classification over large-scale hyperspectral images with very few labeled pixels. Besides, we discuss and experiment self-supervised techniques for spectral representation learning, including the Masked Autoencoder (He et al., 2022), and establish a baseline for pixel-wise classification achieving 85% overall accuracy and 77% F1 score. The Toulouse Hyperspectral Data Set and our code are publicly available at www.toulouse-hyperspectral-data-set.com and www.github.com/Romain3Ch216/tlse-experiments/ , respectively. [ABSTRACT FROM AUTHOR]

Condensation on surfaces in the presence of non-condensable gas (NCG) is a ubiquitous and critical phenomenon in many industrial fields. However, current empirical correlations for predicting condensation heat transfer coefficients in the presence of NCG may exhibit significant deviations from reality. Machine learning algorithms are now being dedicated to this field, but the models developed can only predict the condensation heat transfer coefficients of steam or vapor of a non-aqueous working fluid with similar physical properties to water in the presence of NCG. In the present study, a comprehensive theoretical analysis was conducted to investigate the total condensation heat transfer coefficients with NCG, and 16 dimensionless numbers were identified as input variables for machine learning models. Based on a filtered database consisting of 4377 data points extracted from 37 papers, the Spearman correlation coefficients were calculated to evaluate the relationship between the total heat transfer coefficients and the input variables, indicating the magnitude of the impact of the 16 dimensionless variables. Four machine learning models, namely Gradient Boosting Regression (GBR), Extreme Gradient Boosting (XGBoost), Random Forest Regression (RFR), and Multilayer Perceptron (MLP), were developed to predict the total heat transfer coefficients for water and non-aqueous working fluids. The mean absolute percentage errors for the four models were 1.38%, 1.63%, 3.00%, and 4.42%, respectively, with the GBR model exhibiting the highest degree of accuracy. The determination of the application scope of these models was conducted by analyzing the value ranges for each dimensionless parameter and its corresponding frequency distribution. • Four machine-learning models were developed to predict condensation heat transfer. • Sixteen dimensionless variables were determined as input parameters for these models. • These models can predict condensation heat transfer under different working conditions. • Conditions were diverse in types of wall geometry, convection, working fluid and NCG. [ABSTRACT FROM AUTHOR]

Risk assessment studies of local asteroid hazards traditionally simulate the physics of meteors with engineering models tailored to analyze tens-of-millions of scenarios. However, these simplified approaches still need to solve time-dependent ODEs to model the entry process and the resulting ground damage. With a computational cost of O(0.01 CPU.s) per scenario, simulating these large numbers of potential entry conditions can take several days on local computers. To improve computational efficiency, we propose in this paper an orthogonal approach based on machine learning models to predict the size of damaged areas given a list of entry parameters. We train 5 machine learning methods and compare the predictions to the outputs of NASA's state-of-the-art Probabilistic Asteroid Impact Risk (PAIR) model. PAIR's remarkable capability to handle a broad spectrum of entry conditions and a high volume of scenarios establishes it as our reference model, with PAIR estimates assumed ground truth and used to train, test, and validate the machine learning models. The ML models, initially trained with PAIR's primitive input variables, are then further enhanced with engineered features to improve the performance of interpretable models. We find that complex models like neural networks are well-suited to estimate blast hazards, while simpler, interpretable linear models can accurately assess thermal damage. For both types of hazards, the radii of damaged areas can be predicted with less than 10% average relative errors for radii greater than 100 km. The CPU time is decreased by a factor O(1 0 3 ) compared to the PAIR model, enabling the simulation of millions of scenarios within minutes on a local computer. Beyond serving as surrogate models, we incorporate the machine learning models into a comprehensive Shapley sensitivity analysis, yielding a ranking of entry parameters based on their contributions to ground damages. These results offer valuable insights for prioritizing observation missions. • Machine learning models enable accurate and rapid simulations of asteroid hazards. • Data-driven methods can assess the radius of damaged areas with 10% average errors. • Models provide insights on the relative importance of entry parameters. • Risk studies can be performed on local machines instead of supercomputers. [ABSTRACT FROM AUTHOR]

Detecting action units is an important task in face analysis, especially in facial expression recognition. This is due, in part, to the idea that expressions can be decomposed into multiple action units. To evaluate systems that detect action units, F1-binary score is often used as the evaluation metric. In this paper, we argue that F1-binary score does not reliably evaluate these models due largely to class imbalance. Because of this, F1-binary score should be retired and a suitable replacement should be used. We justify this argument through a detailed evaluation of the negative influence of class imbalance on action unit detection. This includes an investigation into the influence of class imbalance in train and test sets and in new data (i.e., generalizability). We empirically show that F1-micro should be used as the replacement for F1-binary. • We show that AU base rates have a large influence on detection across different architectures. • We show how different evaluation metrics are impacted by AU base rates. • We argue that F1-binary should not be used, for AU detection, and that F1-micro should replace it. [ABSTRACT FROM AUTHOR]

Although current semantic segmentation approaches have achieved impressive performance, their ability to incrementally learn new classes is limited. Moreover, pixel-by-pixel annotations are costly and time-consuming. Therefore, a new field called Weakly-supervised Incremental Learning for Semantic Segmentation (WILSS) has emerged, which learns new classes using image-level labels. However, image-level labels do not provide sufficient detail, and we discover that the state-of-the-art of WILSS suffers from confusion between old knowledge and new knowledge. To address this issue, we propose W eakly-supervised I ncremental learning for S emantic segmentation with Class H ierarchy (WISH), a method that considers the hierarchical structure of each class when determining which knowledge to trust in cases of confusion between old and new knowledge. Our method has achieved new state-of-the-art performances in all settings compared to the previous methods on the Pascal VOC and MS COCO datasets. • Our paper introduces WISH, utilizing hierarchy and image labels for class incremental learning in semantic segmentation. • We efficiently utilize hierarchy, boosting segmentation performances without added costs. • Our method outperforms existing methods in all configurations, showcasing hierarchical integration's effectiveness. [ABSTRACT FROM AUTHOR]

In many high-impact applications, it is important to ensure the quality of the output of a machine learning algorithm as well as its reliability in comparison to the complexity of the algorithm used. In this paper, we have initiated a mathematically rigorous theory to decide which models (algorithms applied on data sets) are close to each other in terms of certain metrics, such as performance and the complexity level of the algorithm. This involves creating a grid on the hypothetical spaces of data sets and algorithms so as to identify a finite set of probability distributions from which the data sets are sampled and a finite set of algorithms. A given threshold metric acting on this grid will express the nearness (or statistical distance) of each algorithm and data set of interest to any given application. A technically difficult part of this project is to estimate the so-called metric entropy of a compact subset of functions of infinitely many variables that arise in the definition of these spaces. [ABSTRACT FROM AUTHOR]

• A SMOTE approach was proposed to handle imbalanced data of hazards levels. • SMOTE was integrated with Gaussian process for the prediction of hazards levels. • A real-world example is used to illustrate and validate the proposed approach. • Results from different data generation approaches are compared with that of SMOTE. • Results from different machine learning methods are compared with that of GP. Cultural soil heritages (CSHs) are artifacts with historical, artistic, and scientific significance; however, they are vulnerable to various hazards, such as weathering, fractures, hollowing, collapses, and gullies. This is especially true for those CSHs exposed to the outdoors. Due to the large number of CSHs sites within China, managing and protecting these heritages with the aid of detailed on-site investigations is time-consuming and expensive. Consequently, evaluating the spatial distribution and degree of hazards developed in all these heritages becomes impractical. To address this issue, this paper developed a Gaussian process classification (GPC) method to predict the spatial distribution of typical hazards (i.e., weathering, fractures, hollowing, collapses, and gullies) and the development level of each hazard from eight environmental factors (e.g., annual relative humidity and annual sunshine time) and a limited number of investigation data. As the number of investigation data for different levels of each hazard is usually imbalanced and sparse, this study proposed a synthetic minority oversampling technique (SMOTE) with GPC to form the SMOTE-GPC method. A real-world example is used to illustrate this approach. Results from real-world data demonstrated that the proposed method achieved an F 1 score, precision, recall, and Cohen's kappa with values greater than 0.93 in both the training and testing datasets, indicating its good performance. [ABSTRACT FROM AUTHOR]

Efficient detection of refrigerant leakage is of utmost importance for industrial refrigeration systems due to its potential to cause substantial impacts on system performance and the environment. Existing research on fault detection and diagnosis in refrigeration systems primarily revolves around solutions based on experimental or laboratory data. However, in the industrial use case, achieving accurate and early detection poses significant challenges. This paper reports on the development of a novel refrigerant leak detection method for industrial vapor compression refrigeration systems. Our method leverages real-world data obtained from operational installations, enabling us to assess its reliability and applicability. The proposed data-driven approach involves predicting the fault-free liquid level in the installation receiver and comparing the actual and predicted levels. In this work, we place emphasis on features and model selection. Dedicated metrics combined with a model comparison method are proposed to evaluate and compare the performance of commonly used regression models with two sets of features to determine the most effective one. Furthermore, we provide insights into the results obtained from the deployment of the proposed method in real-world industrial installations. • Refrigerant leak is detected through the receiver liquid level in the installation. • A regression model can be used to predict the fault-free liquid level. • System performance indicators were proven to be effective in long-term predictions. • Extremely randomized forest emerged as the top performer. [ABSTRACT FROM AUTHOR]

Despite significant progress has been made towards crop yield prediction with remote sensing, there exist knowledge gaps on (1) the impacts of temporal resolution of imaging frequencies on yield prediction, (2) transferability of the models among different genotypes and test sites, and (3) translation of these research developments to crop breeding that benefit farmers. Existing research predominantly provides an on-site perspective, frequently missing the complexities of real-world applications. The objectives of this paper are to investigate the transferability and generalization capabilities of yield prediction models for crop breeding across test sites located in North and South Americas. Toward that goal, we tested different machine learning techniques including Random Forest Regressor (RF), Gradient Boosting Regression (GB), and Deep Neural Networks (DNN) for soybean yield prediction with experiments conducted in different climate and growth conditions. A novel transfer learning approach was proposed for genotype selection and categorizing soybean yield for screening high-yield varieties. Furthermore, we studied the effect of temporal resolution on yield prediction, focusing on the critical development stages and optimal aerial survey frequencies for precise yield prediction using large 31,404 sample data. Results demonstrated that the combined dataset of Argentina and United States representing different climate regimes provided the highest performance with an R2 of 0.76 using RF and GB algorithms. The classification approach was proven to be most useful for crop breeding as demonstrated by accurately identifying the high-yielding genotypes. Increasing temporal sampling of key phenological stages significantly improved yield prediction. Although transfer learning yielded promising outcomes across trials within Argentina the efficacy of transferring models from Argentina to the United States was limited, attributed to significant seasonal and climate variations. This study pioneered the use of transfer learning for model adaptability in real-world breeding scenarios, training and transferring models within the South and North Americas, providing actionable insights and strategies for the breeding community, aiming to facilitate improved decision-making for agricultural productivity. [ABSTRACT FROM AUTHOR]

With the ever increasing number of active satellites in space, the rising demand for larger formations of small satellites and the commercialization of the space industry (so-called New Space), the realization of manufacturing processes in orbit comes closer to reality. Reducing launch costs and risks, allowing for faster on-demand deployment of individually configured satellites as well as the prospect for possible on-orbit servicing for satellites makes the idea of realizing an in-orbit factory promising. In this paper, we present a novel approach to an in-orbit factory of small satellites covering a digital process twin, AI-based fault detection, and teleoperated robot-control, which are being researched as part of the "AI-enabled Cyber–Physical In-Orbit Factory" project. In addition to the integration of modern automation and Industry 4.0 production approaches, the question of how artificial intelligence (AI) and learning approaches can be used to make the production process more robust, fault-tolerant and autonomous is addressed. This lays the foundation for a later realization of satellite production in space in the form of an in-orbit factory. Central aspect is the development of a robotic AIT (Assembly, Integration and Testing) system where a small satellite could be assembled by a manipulator robot from modular subsystems. Approaches developed to improving this production process with AI include employing neural networks for optical and electrical fault detection of components. Force sensitive measuring and motion training helps to deal with uncertainties and tolerances during assembly. An AI-guided teleoperated control of the robot arm allows for human intervention while a Digital Process Twin represents process data and provides supervision during the whole production process. Approaches and results towards automated satellite production are presented in detail. • An automated production system for modular small satellites is proposed. • Central element is a force-sensitive robotic assembly, integration & testing system. • AI and machine learning approaches are applied to make the process more robust. • A digital process twin supervises the autonomous production. • A multi-modal shared control approach allows human teleoperation if required. [ABSTRACT FROM AUTHOR]

This research paper presents a thorough examination of shock wave train phenomena in various duct structures, utilizing a machine learning-optimized k-ω model. The focus of the study is to apply machine learning techniques to adapt the constant coefficients of the k-ω turbulence model, improving its accuracy and computational efficiency in capturing the dynamics of shock waves. An important aspect of this investigation is the impact of diverging sections within the duct, specifically how changes in the divergence angle, while maintaining a constant ratio of the exit area to the throat area, affect compressible flow parameters, shock wave positions, and other related characteristics of shock trains. The study systematically explores these effects and provides fresh insights into the behavior of shock wave trains under different divergence conditions. In a departure from the usual practices in supersonic research, this study compares the phenomena of shock wave trains in rectangular and circular duct geometries. The findings contribute valuable data and analyses to a field that has traditionally focused on rectangular configurations. The results indicate that the shape of the duct has a significant influence on the shock wave train, emphasizing the importance of considering geometric diversity in the study of supersonic flows and shock wave phenomena. This research sets the stage for further investigations into the interaction between duct geometry, shock wave propagation, and the dynamics of compressible flows. • Improved k-ω SST model accuracy to 95.5 % with machine learning. • Longer duct diverging sections intensify the first lambda shock. • Duct shape critically affects shock wave and boundary layer interactions. [ABSTRACT FROM AUTHOR]

Ionospheric scintillation is a phenomenon that influences radio waves from the Global Navigation Satellite System (GNSS) in the ionosphere, reducing the accuracy, integrity, and continuity of tracking and navigation applications. Automatic and accurate detection of scintillation events based on threat level is essential for space weather forecasting applications. Therefore, a multi-class classification of four categories based on scintillation intensities was developed. In this paper, a Synthetic Minority Oversampling Technique (SMOTE) based Super Learner (LR) machine learning classification algorithm is proposed with two months of GPS ionospheric amplitude scintillations S4 data obtained for 2015 from Hyderabad station (17.45°N, 78.47°E). The Synthetic Minority Oversampling Technique (SMOTE) is implemented to oversample the minority classes to balance the events in each class. Moreover, we also focused on annotating the data transmitted by all the visible satellites. Later, a machine learning method is proposed to achieve the automated detection of ionospheric scintillation to improve the detection performance and obtain the detected results for each category. The experimental results show that the proposed SL model approach considerably has detection accuracy on the classified data as 97.8% during the quiet day and 78.3% during storm day for the Hyderabad station. The confusion matrix results indicate the proposed algorithm's effectiveness for quiet and disturbed conditions. [ABSTRACT FROM AUTHOR]

Face recognition models often encounter various unseen domains and environments in real-world applications, leading to unsatisfactory performance due to the open-set nature of face recognition. Models trained on central datasets may exhibit poor generalization when faced with different candidates under varying illumination and blur conditions. In this paper, our goal is to enhance the generalization of face recognition models for diverse target conditions without relying on active or incremental learning. We propose an approach for face recognition that utilizes contrastive learning to synthesize positive and multiple negative samples. To address the combinatorial challenges posed by positive and negative samples, our framework incorporates a combination of contrastive regularizer loss and Arcface loss, along with an effective sampling strategy for batch model learning. We update the model weights by jointly back-propagating contrastive and ArcFace gradients. We validate our method on both generalized and standard face recognition benchmarks dataset namely IJB-B and IJB-C. Series of experimentation revealed the out-performance of proposed framework against other state-of-the-art methods. • Proposed generalized face recognition aimed to handle unknown target domains without model updates or fine-tuning. • Proposed contrastive learning-based approach to address problem of illumination and motion blur for registered candidates. • Employ augmentation techniques to generate positive and negative samples to mitigate computational complexities. • Integrate ArcFace and contrastive regularizer loss to learn distinctive face representation for each identity. • Performed series of experiments for proof convergence of proposed model on IJB-B and IJB-C datasets. [ABSTRACT FROM AUTHOR]

The accurate measurement of the thermal conductivity of hydrogen is crucial in efficiently integrating hydrogen into diverse applications that encompass storage, utilization, and production. This paper introduces strong intelligent algorithms capable of predicting the thermal conductivity of hydrogen-based binary gaseous mixtures consisting of carbon monoxide (CO), methane (CH 4), and carbon dioxide (CO 2), based on the mole fractions of these components, pressure, and temperature. To this end, four advanced intelligent models, namely Radial Basis Function (RBF), Cascade Forward Neural Network (CFNN), Multilayer Perceptron (MLP), and Generalized Regression Neural Network (GRNN) were developed. After that, an analysis of graphical and statistical errors was conducted to further examine the precision of the proposed models. The findings revealed a strong alignment between the suggested models and experimental values. Additionally, among the developed models, the GRNN technique was identified as the superior forecasting approach with the (RMSE = 0.00129) and the highest value of the coefficient of determination, (R2 = 0.9987). Moreover, the sensitivity analysis of the GRNN model indicates that hydrogen composition has the highest influence on the thermal conductivity of hydrogen-based gaseous mixtures, as evidenced by the relevancy factor value of 0.881. The assessment of group errors also demonstrates that the proposed models exhibit high accuracy when confronted with higher temperatures and pressures. Ultimately, the Leverage analysis was employed to evaluate the dependability of the new paradigm, revealing that more than 97% of the data falls within the applicable range of the paradigm. • Intelligent approaches estimate the thermal conductivity of H2-based mixtures. • GRNN technique outperforms other models for forecasting thermal conductivity. • Hydrogen composition has the highest influence on thermal conductivity. • Proposed models exhibit high accuracy at higher temperatures and pressures. [ABSTRACT FROM AUTHOR]

The establishment of a prediction model for the hydrogen absorption state of a solid-state hydrogen storage device is crucial for its practical application. In this paper, a solid-state hydrogen storage device filled with AB 2 hydrogen storage alloy is investigated. The experimental data are obtained by controlling the hydrogen absorption temperature and hydrogen absorption rate respectively. The Broyden-Fletcher-Goldfarb-Shanno algorithm (BFGS) is used to fit the solid-state hydrogen storage device experimental data based on the Richards model, and Linear Regression, Gradient Descent and Artificial Neural Network algorithms are used to predict the parameters of the Richards model. It is confirmed that the Linear Regression model has the best prediction effect after verified by the test set, with coefficients of determination R2 ≥ 0.96. This study provides an effective way to predict the hydrogen absorption state of solid-state hydrogen storage devices quickly. • The effects of temperature and rate on hydrogen absorption process of solid-state hydrogen storage device are studied. • Richards model is used to fit the hydrogen absorption curve shaped as an S-type. • The prediction model of hydrogen absorption process of the device is established. • Three machine learning algorithms are applied to establish the prediction model. [ABSTRACT FROM AUTHOR]

• In clinical needle EMG interpretation, conventional machine learning methods were more often studied than deep learning methods. • Most studies developed models for classification of contraction signals and a few addressed the classification of rest signals. • Current research has methodological shortcomings, which prevent clinical implementation, but are surmountable. This scoping review provides an overview of artificial intelligence (AI), including machine and deep learning techniques, in the interpretation of clinical needle electromyography (nEMG) signals. A comprehensive search of Medline, Embase and Web of Science was conducted to find peer-reviewed journal articles. All papers published after 2010 were included. The methodological quality of the included studies was assessed with CLAIM (checklist for artificial intelligence in medical imaging). 51 studies were identified that fulfilled the inclusion criteria. 61% used open-source EMGlab data set to develop models to classify nEMG signal in healthy, amyotrophic lateral sclerosis (ALS) and myopathy (25 subjects). Only two articles developed models to classify signals recorded at rest. Most articles reported high performance accuracies, but many were subject to bias and overtraining. Current AI-models of nEMG signals are not sufficient for clinical implementation. Suggestions for future research include emphasizing the need for an optimal training and validation approach using large datasets of clinical nEMG data from a diverse patient population. The outcomes of this study and the suggestions made aim to contribute to developing AI-models that can effectively handle signal quality variability and are suitable for daily clinical practice in interpreting nEMG signals. [ABSTRACT FROM AUTHOR]

Despite the growing availability of high-capacity computational platforms, implementation complexity still has been a great concern for the real-world deployment of neural networks. This concern is not exclusively due to the huge costs of state-of-the-art network architectures, but also due to the recent push towards edge intelligence and the use of neural networks in embedded applications. In this context, network compression techniques have been gaining interest due to their ability for reducing deployment costs while keeping inference accuracy at satisfactory levels. The present paper is dedicated to the development of a novel compression scheme for neural networks. To this end, a new form of ℓ 0 -norm-based regularization is firstly developed, which is capable of inducing strong sparseness in the network during training. Then, targeting the smaller weights of the trained network with pruning techniques, smaller yet highly effective networks can be obtained. The proposed compression scheme also involves the use of ℓ 2 -norm regularization to avoid overfitting as well as fine tuning to improve the performance of the pruned network. Experimental results are presented aiming to show the effectiveness of the proposed scheme as well as to make comparisons with competing approaches. [ABSTRACT FROM AUTHOR]

Decentralized deep learning algorithms leverage peer-to-peer communication of model parameters and/or gradients over communication graphs among the learning agents with access to their private data sets. The majority of the studies in this area focus on achieving high accuracy, with many at the expense of increased communication overhead among the agents. However, large peer-to-peer communication overhead often becomes a practical challenge, especially in harsh environments such as for an underwater sensor network. In this paper, we aim to reduce communication overhead while achieving similar performance as the state-of-the-art algorithms. To achieve this, we use the concept of Minimum Connected Dominating Set from graph theory that is applied in ad hoc wireless networks to address communication overhead issues. Specifically, we propose a new decentralized deep learning algorithm called minimum connected Dominating Set Model Aggregation (DSMA). We investigate the efficacy of our method for different communication graph topologies with a small to large number of agents using varied neural network model architectures. Empirical results on benchmark data sets show a significant (up to 100X) reduction in communication time while preserving the accuracy or in some cases, increasing it compared to the state-of-the-art methods. We also present an analysis to show the convergence of our proposed algorithm. • Proposing a communication-efficient protocol for decentralized learning systems. • Finding valuable connections in the network to reduce communication overhead. • Hybrid of federated learning and gossiping methods. • Verifying the method performance for different networks, datasets, and models. • Preserving accuracy while reducing communication, even in large-scale networks. [ABSTRACT FROM AUTHOR]

Building footprint information is crucial for comprehending global urban development processes. Deep learning algorithms have shown significant potential in building extraction from high spatial resolution imagery. However, the requirement for large-scale annotated data has been a limitation for applying deep learning methods to city-level or national-level building mapping. The dynamic change and distinct landscape variation of cities in different geographic locations further emphasizes the need for automatic building footprint extraction. In this paper, we propose a Self-supervised Knowledge Transfer (SKTrans) framework for extracting building footprints from unlabeled remote sensing images over large areas. To address building tone differences cross regions and time periods, a tone shift mechanism is introduced to reduce the contrast, brightness, and saturation differences between buildings. Furthermore, considering building style diversity across different cities, a knowledge integration module is proposed to develop a comprehensive depiction including tonal differences, multi-scale variance, structural attributes and semantic confusion between roads and buildings. The effectiveness of SKTrans is demonstrated on three public datasets of WHU, Massachusetts, and Inria. To further evaluate the feasibility of large-scale mapping for automatic building footprint, global generalization experiments cross regions and time periods, which is based on imagery from five countries of different continents were implemented. The comprehensive results demonstrate that SKTrans surpasses existing state-of-the-art methods, supporting the time-series city-level building mapping without labeled dataset. Additionally, the architectural landscape pattern changes can be quantitatively analyzed based on the generalization results to facilitate sustainable urban development. Overall, SKTrans provides a new insight to meet the needs of high-precision label-free building mapping and update from large-scale multi-style remote sensing images. [ABSTRACT FROM AUTHOR]

The term "mimicry attack" has been coined in computer security and used in adversarial machine learning: an attacker observes what a machine-learning system has learned and adjusts the malicious input so that it mimics a benign input. In this paper we extend this concept to image forensics, to allow an attacker modifying a manipulated image so that it appears pristine when analyzed by a target forensic detector. Recent work has shown that such attacks can be executed against detectors based on deep networks for hiding image tampering. We do more than that: our mimicry attack can force the target detector to identify arbitrary fictitious manipulations, while hiding the true ones. Accordingly, the user of the forensic detector is completely misled. From a methodological viewpoint, the proposed attack artificially alters the detector-specific intermediate representations according to the pixel distribution in the manipulated image, by applying a gradient-based optimization process. Experimental tests on different data sets and detectors demonstrate that our approach succeeds in jointly hiding manipulated areas and arbitrarily adding new ones, favorably comparing with the state-of-the-art in the first task. [Display omitted] • The proposed mimicry attack hides forgeries in images and introduces fictitious ones • It allows adversaries to introduce semantic biases and mislead the forensic detection • Its large-scale adoption may hinder the trust on forensic tools, akin to DoS attacks • Sequential detector-specific attacks enable tackling multiple detectors at once [ABSTRACT FROM AUTHOR]

Pressure swing adsorption(PSA) based gas separation technology is widely applied in chemical industries, especially in hydrogen purification. Its complex mechanism and non-steady state feature restrict the separation performance adjustment and application on efficient gas utilization. Process simulation, machine learning and modelling combined innovating gas management methodology is proposed in this paper to perform accurate gas separation for natural gas based syngas mixture. VB script program is employed to drive Aspen adsorption simulation, and then classical and regression tree(CART), polynomial regression(PNR) and surrogate model are comprehensively used to perform machine learning. Consequently, the coefficient matrix is obtained to quantify the separation performance. This method holds the potential for broader application in refineries and SPCs hybrid hydrogen network, facilitating the comprehensive utilization of hydrogen containing gases. Case study results show this methodology can conserve natural gas and gas productivity by 21.3 %, as well as carbon emission by 8211.6 Nm3/h. This work highlights the promising application of digital twin technology within the chemical industry, showcasing how process simulation, machine learning, and modeling can collectively revolutionize gas management and resource conservation. [Display omitted] • Novel PSA based gas allocation instead of gas separation technology. • Simulation, modelling and machine learning combined gas accommodating methodology. • An insight technology for integration of refinery and SPC. • A novel gas allocation method for zero waste of hydrogen. • An insight application of digital twin technology for gas efficient utilization. [ABSTRACT FROM AUTHOR]

Conducting in-situ measurements of planetary environments is a key scientific goal in the geoscience field for the extraterrestrial exploration, such as the Moon or Mars. Among these measurements, determining the thermal properties of regolith is particularly challenging due to the material's inhomogeneities. In this paper, an active thermal probing technique is proposed, utilizing a transient hot-wire technique with a needle-like probe structure integrated with temperature sensors and a heater. It is capable of measuring in-situ temperature variations and inner heat flux without disturbing the surrounding regolith. In addition, a machine-learning-based method is proposed for inverting in-situ thermophysical properties, including density, thermal conductivity, and heat capacity. An embedded artificial neural network model has been developed and trained with data sets from both finite element simulations and experiments. This machine learning model is capable of efficiently predicting temperature responses for different lunar regolith simulants. For total 12 parameters, the errors of ten cases are less than 10%, and other two cases are 12.23% and 13.17%. Finally, the proposed method is verified by heat measurement experiments under normal pressure and vacuum, and the variation of thermal properties with different compactness is obtained. The proposed probing technique provides novel ideas to measure the in-situ thermal properties of extraterrestrial regolith. [ABSTRACT FROM AUTHOR]

With the extensive utilization of fossil fuels, environmental concerns have been growing. Hydrogen-powered vehicles (HPVs) as well as hydrogen refueling stations (HRSs) are expected to proliferate to alleviate the pollution. In order to promote the economic valibity of HRSs, this paper proposes an optimal scheduling model for distributed HRSs to maximize HRSs' revenue during a day. HRSs serves both the mobility sector with fuel individually and the power system with reserve demand response collaboratively. The elaborated operating of electrolysers and relaxed control of hydrogen storage are involved to enhance practicability and flexibility of the scheduling. Considering that this model has multiple variables, intractable high-dimension nonlinear constraints, a transfer multi-agent reinforcement learning algorithm is developed. The algorithm consists of two parts: the one is the multi-agent optimization method by transforming multi-decision optimization into a multi-agent optimization; the other is to transfer the empirical knowledge of the source task to reinforcement learning according to the similarity between the source task and the target task, making it cut down the blindness of exploration and accelerate convergence. Numerical studies reveal that participating in the reserve services market collaboratively increases overall revenues of HRSs by up to 32.90%, and the computation time is sharply shortened by approximately 34 times compared to the basic reinforcement learning. • A joint operation model of distributed HRSs is proposed to lift revenue by 32.90%. • Relaxed control of hydrogen storage is utilized to minimize service deviation. • An evolutionary transfer reinforcement learning shortens solving time by 34 times. [ABSTRACT FROM AUTHOR]

Unmanned aerial vehicle-based hyperspectral imaging (HSI) and machine learning (ML) techniques are applied for crop field data collection such as yield, nitrogen content, leaf chlorophyll, biomass estimation, leaf area index (LAI), and biotic and abiotic stress. It is very difficult to identify with certainty best practices for specific applications and for targeted crop species because the results are frequently contradictory; models or indices sometimes perform well, but other times they do not. ML and HSI are useful and sophisticated techniques, but treating them as first-option methods may be detrimental in the long term. Favoring HSI in situations where a multispectral sensor performs equally well only delays end-user adoption of technology owing to cost and complexity. The past few years have seen increased interest in unmanned aerial vehicle (UAV)-based hyperspectral imaging (HSI) and machine learning (ML) in agricultural research, concomitant with an increase in published research on these topics. We provide an updated review, written for agriculturalists, highlighting the benefits in the retrieval of biophysical parameters of crops via UAVs relative to less sophisticated options. We reviewed >70 recent papers and found few consistent results between similar studies. Owing to their high complexity and cost, especially when applied to crops of low value, the benefits of most of the research reviewed are difficult to explain. Future effort will be necessary to distill research findings into lower-cost options for end-users. [ABSTRACT FROM AUTHOR]

Federated learning (FL) provides autonomy and privacy by design to participating peers, who cooperatively build a machine learning (ML) model while keeping their private data in their devices. However, that same autonomy opens the door for malicious peers to poison the model by conducting either untargeted or targeted poisoning attacks. The label-flipping (LF) attack is a targeted poisoning attack where the attackers poison their training data by flipping the labels of some examples from one class (i.e., the source class) to another (i.e., the target class). Unfortunately, this attack is easy to perform and hard to detect, and it negatively impacts the performance of the global model. Existing defenses against LF are limited by assumptions on the distribution of the peers' data and/or do not perform well with high-dimensional models. In this paper, we deeply investigate the LF attack behavior. We find that the contradicting objectives of attackers and honest peers on the source class examples are reflected on the parameter gradients corresponding to the neurons of the source and target classes in the output layer. This makes those gradients good discriminative features for the attack detection. Accordingly, we propose LFighter, a novel defense against the LF attack that first dynamically extracts those gradients from the peers' local updates and then clusters the extracted gradients, analyzes the resulting clusters, and filters out potential bad updates before model aggregation. Extensive empirical analysis on three data sets shows the effectiveness of the proposed defense regardless of the data distribution or model dimensionality. Also, LFighter outperforms several state-of-the-art defenses by offering lower test error, higher overall accuracy, higher source class accuracy, lower attack success rate, and higher stability of the source class accuracy. Our code and data are available for reproducibility purposes at https://github.com/NajeebJebreel/LFighter. [ABSTRACT FROM AUTHOR]

• Developed XGBoost model to predict refrigerant mass flow through electronic expansion valves. • Achieved improved prediction accuracy over existing models using two refrigerants. • XGBoost effectively captured complex nonlinear flow dynamics with 96.4–90.6 % accuracy. • A low RMSE of 2.2–4.23 % indicated high predictive performance of the developed model. • Advanced machine learning approach provides direction for more efficient refrigeration control. In an effort to enhance the prediction accuracy of the mass flow characteristics of refrigerants through electronic expansion valves (EEVs), this study develops a mass flow model using the XGBoost machine learning algorithm. Utilizing experimental data from open literature for refrigerants R1233zd(E) and R245fa, the model aims to accurately predict the mass flow coefficient in EEVs, a crucial aspect for improving the performance and efficiency of refrigeration systems. The model's performance is evaluated using the coefficient of determination (R2) on the training and test datasets, revealing a minimal performance gap and no overfitting issue. Remarkably, our model's predictions for R1233zd(E) and R245fa datasets align consistently with experimental data, with 96.4 % and 90.6 % of the predicted data deviating from the experimental ones within ±5 %, respectively. Further, the root-mean-square error (RMSE) and coefficient of variation of the root mean square error (CV-RMSE) values are quite low, with 2.2 % and 2.33 % for R1233zd(E), and 4.23 % and 4.47 % for R245fa, indicating high prediction accuracy. Compared to the original models, the paper approach with the XGBoost algorithm significantly improves prediction accuracy. This advancement provides a promising direction for developing more efficient and reliable control strategies for refrigeration systems in various applications. [ABSTRACT FROM AUTHOR]

The interpretability of ML models is important, but it is not clear what it amounts to. So far, most philosophers have discussed the lack of interpretability of black-box models such as neural networks, and methods such as explainable AI that aim to make these models more transparent. The goal of this paper is to clarify the nature of interpretability by focussing on the other end of the "interpretability spectrum". The reasons why some models, linear models and decision trees, are highly interpretable will be examined, and also how more general models, MARS and GAM, retain some degree of interpretability. It is found that while there is heterogeneity in how we gain interpretability, what interpretability is in particular cases can be explicated in a clear manner. [ABSTRACT FROM AUTHOR]

The first year of the COVID-19 pandemic constituted a major life stress event for many adolescents, associated with disrupted school, behaviors, social networks, and health concerns. However, pandemic-related stress was not equivalent for everyone and could have been influenced by pre-pandemic factors including brain structure and sleep, which both undergo substantial development during adolescence. Here, we analyzed clusters of perceived stress levels across the pandemic and determined developmentally relevant pre-pandemic risk factors in brain structure and sleep of persistently high stress during the first year of the COVID-19 pandemic. We investigated longitudinal changes in perceived stress at six timepoints across the first year of the pandemic (May 2020–March 2021) in 5559 adolescents (50 % female; age range: 11–14 years) in the United States (U.S.) participating in the Adolescent Brain Cognitive Development (ABCD) study. In 3141 of these adolescents, we fitted machine learning models to identify the most important pre-pandemic predictors from structural MRI brain measures and self-reported sleep data that were associated with persistently high stress across the first year of the pandemic. Patterns of perceived stress levels varied across the pandemic, with 5 % reporting persistently high stress. Our classifiers accurately detected persistently high stress (AUC > 0.7). Pre-pandemic brain structure, specifically cortical volume in temporal regions, and cortical thickness in multiple parietal and occipital regions, predicted persistent stress. Pre-pandemic sleep difficulties and short sleep duration were also strong predictors of persistent stress, along with more advanced pubertal stage. Adolescents showed variable stress responses during the first year of the COVID-19 pandemic, and some reported persistently high stress across the whole first year. Vulnerability to persistent stress was evident in several brain structural and self-reported sleep measures, collected before the pandemic, suggesting the relevance of other pre-existing individual factors beyond pandemic-related factors, for persistently high stress responses. • The paper reports findings from the ABCD study® on U.S. youth across the first COVID-19 pandemic year: May 2020 – March 2021. • 5% of adolescents reported persistently high stress levels during the first year of the pandemic. • Persistent stress was linked to pre pandemic brain structure, sleep difficulties, and more advanced puberty stages. • The findings inform preventive and treatment approaches for persistent stress during major life stress events like pandemics. [ABSTRACT FROM AUTHOR]

The dehydrogenation enthalpy, a crucial thermodynamic parameter, determines the difficulty of hydrogen desorption in V–Ti–Cr–Fe alloy, but it is unaffordable to predict and modify the dehydrogenation enthalpy and investigate the synergistic effects of structural and chemical parameters relying on conventional experimental methods. In this paper, an extreme gradient boosting (Xgboost) model with feature selection is developed, which can obtain a relative error under 15% in prediction on most of the data. Moreover, the genetic algorithm is utilized for optimizing alloy composition having suitable dehydrogenation enthalpy. The accuracy of prediction and optimization is demonstrated by the subsequent experimental validation, where the error between calculation and experiment is within 4.5%. Density function theory (DFT) is used to analyze the key features provided by machine learning, with results suggesting that the volumes, valence electrons, and bulk module factors have a significant impact. This research supplies an efficient way of using machine learning to predict and optimize the dehydrogenation enthalpy. In addition, DFT calculation also validates the feature importance outputted by machine learning and reveals the principle of feature effects at the microscopic aspect. • ML models predict precisely the dehydrogenation enthalpy of V–Ti–Cr–Fe alloy. • Properties calculation optimized by GA have good agreement with experiments. • Explanation of volume, electron and module factors affecting enthalpy using DFT. [ABSTRACT FROM AUTHOR]

Liquid hydrogen has attracted widespread attention due to industrial decarbonization as an excellent carrier of renewable energy accommodation. Accurate prediction of the heat transfer coefficient during flow boiling has become a fundamental prerequisite for enhancing the design rationality and safety of hydrogen devices since a slight heat leak will trigger flow boiling. A novel selection standard of the dataset for liquid hydrogen flow boiling is proposed, and an experimental database of liquid hydrogen is compiled, including 923 data points from 5 sources. The prediction performance of eight conventional empirical correlations and six machine learning models for liquid hydrogen flow boiling is compared and evaluated. Results show that the prediction ability of machine learning models for the heat transfer coefficient of liquid hydrogen flow boiling is much higher than that of traditional correlations. Fang correlation is recommended to use as the calculation benchmark due to simplicity and reliability if the inlet condition is obscure. The Extra tree model shows the most excellent evaluating performance among the six machine learning models with a mean absolute deviation of 11.44% and a fairly superior R 2 of 0.9543. The Froude number Fr exhibits an essential impact according to the feature importance analysis, which may need to be taken into account when developing new empirical correlations. An improved correlation for the flow boiling of liquid hydrogen is provided based on the Fang correlation, achieving a mean absolute deviation of 18.61% and a mean relative deviation of 16.49%. The results of this paper can offer some theoretical guidance for the design of a liquid hydrogen apparatus. [Display omitted] • A novel selection standard of data set for liquid hydrogen flow boiling is proposed. • Fang correlation is recommended as the calculation benchmark due to simplicity and reliability. • Extra tree model has the best prediction performance with MAD of 11.44% and R 2 of 0.9543. • Froude number Fr exhibits an essential impact on heat transfer. • An improved correlation is proposed with MAD of 18.61% and MRD of 16.49%. [ABSTRACT FROM AUTHOR]

Biomass-derived power generation presents various potential benefits compared to conventional fossil fuel-based power generation. Also, efficient integration of energy systems leads to higher sustainability and effectiveness. This paper presents a biomass-fueled energy system with two open and closed Brayton cycles in the waste heat, of which a humidification and dehumidification system is implemented. The usage of four different biomasses as the fuel of the gasifier is tested and put to comparison from the aspects of thermodynamics, economics, and environment. The Grassman diagram for each layout is presented to signpost the exact point of highest exergy destruction and irreversibility. The optimization based on machine learning techniques is conducted to pinpoint the exact optimum operational conditions. The results indicate that wood offers more sustainability compared to other biomasses, and the amount of energy efficiency, exergy efficiency, and freshwater flow rate produced for Wood biomass in the proposed system are 75.81 %, 36.98 %, and 0.4091 kg/s, respectively. Also, two optimization scenarios have been done for this study. In the initial optimal solution finding, effectivness, unit product cost, and carbon dioxide emission index are equal to 45.9 %, 12.6 $/GJ, and 0.7161 kg/kWh, correspondingly. In the latter scenario, the effectiveness, net power production, and production unit cost of products are equal to 45.9 %, 6580 kW, and 12.61 $/GJ, respectively. • A novel Tri-generation system based on supercritical CO 2 cycle and HDH cycle. • Energy, exergy, techno-economic, and environmental analyses are carried out. • Multi-objective optimization using Machine learning Algorithms is done. • Maximum exergy efficiency is obtained to be 40.39 % for MSW. • The minimum production unit cost of products is obtained to be 14.06 $/GJ for MSW. [ABSTRACT FROM AUTHOR]