Superior Generalization Research Articles

An extended variational autoencoder for cross-subject electromyograph gesture recognition

Surface electromyographic hand gesture recognition has gained significant attention in recent years, especially within the field of human–computer interfaces. However, cross-subject tasks remain challenging due to inherent individual differences. To address this, a novel approach for hand gesture recognition is proposed that leverages a subject-generalized variational autoencoder. This approach involves an extended variational autoencoder designed to disentangle input data into three distinct feature-specific representations. The primary classifier within the variational autoencoder focuses on gesture recognition, while two auxiliary classifiers work together to extract subject-specific and gesture-specific features. The gesture-specific features capture generalized characteristics applicable across all subjects, enabling direct application to new subjects. To enhance accuracy and stability, a competitive voting strategy is implemented. The effectiveness of the proposed method was evaluated using a dataset comprising six representative gestures performed by eight subjects. Comparative analysis with baseline models shows that our approach outperforms others, demonstrating superior generalization with an average accuracy of 90.52% in cross-subject validation.

Działalność Zgromadzenia Sióstr Karmelitanek Dzieciątka Jezus w ostatnim okresie życia m. Teresy Kierocińskiej (1945–1946)

The last period of Mother Teresa Kierocińska’s life was a difficult time for Poland, marked by chaos and instability. The years 1945–1946 were at the same time a period of intensive ministry of the Foundress of the Congregation as Superior General. After the end of hostilities, Mother Teresa and her sisters immediately joined the new tasks of the Church and responded to the needs of their surroundings.The congregation had to take urgent care of the premises for its activities. The sisters in Wolbrom were in a particularly difficult situation, being deprived of their own dwelling. In a short time, the Carmelite Sisters renovated the building assigned to them, at the same time carrying out care and educational activities. They ran two kindergartens and courses for girls. The work of the sisters in Polanka Wielka also required new organisation. There, within a few months, a semi-public chapel and rooms for a kindergarten were erected using the planks of a former German barrack. Between 1945 and 1946, work was carried out to adapt the building in Sosnowiec, where the various sectors of the educational establishment were located. In addition, care was taken to allocate an abandoned building in Wiejska Street to the state authorities for the purposes of an orphanage.Since 1945, in all Carmelite institutions of the Child Jesus, a wide range of apostolic activities have been resumed, including kindergartens, tailoring and embroidery courses, religious instruction, charitable services, and nursing. A retreat house in Czerna was resumed. The development of activities was accompanied by Mother Kierocińska’s concern for the spiritual development of the congregation. Mother Teresa took care of the formation of the professed sisters, as well as of the numerous group of candidates she accepted into the religious community. She was keen to deepen Carmelite spirituality and to ensure that the sisters were guided by the Discalced Carmelites. The new situation allowed for regular visits from the Carmelite fathers with the ministry of retreatants and confessors.At the beginning of July 1946, m. Teresa fell ill with peritonitis and, after suffering for several days, passed to eternity on 12 July 1946 in the presence of the sisters. This was shortly before the scheduled General Chapter at which she was to resign as Superior General. The funeral ceremony took place on 15 July and was presided over by Bishop Stanisław Czajka. After the Holy Mass celebrated in the Church of the Assumption of the Blessed Virgin Mary in Sosnowiec, a numerous funeral procession with the body of the deceased went to the cemetery in Józef Mirecki Avenue. The funeral became a real manifestation of the population who wanted to honor the “Mother of Workers and Zagłębie”.

High-noise solar panel defect identification method based on the improved EfficientNet-V2

As a crucial element in photovoltaic power generation systems, the condition of solar panels significantly impacts the efficiency of power generation. The ability to accurately and promptly detect defects in solar panels is essential for enhancing system performance. This study introduces a novel model for identifying defects in photovoltaic modules, leveraging an enhanced version of EfficientNet-V2. This model aims to address challenges in identifying defects in infrared images of solar panels under conditions of high-noise and low-model efficiency. To address the challenges of high image noise and blur, this article initially presents a methodology that combines the Db4 wavelet transform with a blind deconvolution algorithm for comprehensive preprocessing of the original image. Furthermore, this study optimizes the model's feature representation capabilities by implementing key transformations within the EfficientNet-V2 network framework. Notably, we replaced the traditional SE block with the more efficient channel attention (ECA) mechanism module. Due to its lightweight structure and effective performance, ECA substantially improves the model's capacity to extract complex and abstract image features, while also accelerating the training process's convergence speed and enhancing overall computational efficiency. At the classifier level, this paper innovatively integrates the XGBoost ensemble learning algorithm into the model, substituting the conventional softmax classifier used in traditional convolutional neural network (CNN). With its superior generalization capabilities, robust nonlinear modeling skills, and efficient computational characteristics, XGBoost can more accurately detect minute defects in solar panels based on the deep features produced by EfficientNet-V2, thereby significantly improving the accuracy and robustness of defect detection. Simulation results demonstrate that the proposed model structure outperforms traditional CNN models in terms of accuracy and stability, underscoring the efficacy of the enhanced EfficientNet-V2 model in detecting solar panel defects under high-noise conditions.

An efficient corn leaf disease prediction using Adaptive Color Edge Segmentation with Resnext101 model

Corn leaf disease prediction is essential in ensuring agricultural productivity and food security. Although classical segmentation and classification techniques have been used to identify diseases, they often fail to deal with complex leaf structures. Thus, deep learning has attracted growing attention for automatically learning discriminative features from data. This paper proposes a new approach to predicting corn leaf disease using deep learning and classical segmentation techniques for improved accuracy and efficiency. The proposed work presents an adaptive colour edge segmentation method improved by a hill-climbing search for robust feature extraction according to the diversified nature of corn leaves. With the help of stochastic hill climbing, the segmentation process is optimized by adaptively changing parameters to get better segmentation accuracy. After that, features are extracted and provided as input into a ResNeXt-101 model for disease classification. The novel activation function in ResNeXt-101 combines the Adam optimizer with a Gaussian-based GELU to enhance nonlinearity to the smooth activation of GELU. The results obtained from experiments show that the proposed approach is practical. The ResNeXt-101 model has achieved overwhelming performance, with 99% for testing accuracy and a minimum loss of 0.01, proving its superior generalization capability and robustness. The proposed work was performed with no overfitting and free from local minima during the training process.

Hierarchical Weight Averaging for Deep Neural Networks.

Despite simplicity, stochastic gradient descent (SGD)-like algorithms are successful in training deep neural networks (DNNs). Among various attempts to improve SGD, weight averaging (WA), which averages the weights of multiple models, has recently received much attention in the literature. Broadly, WA falls into two categories: 1) online WA, which averages the weights of multiple models trained in parallel, is designed for reducing the gradient communication overhead of parallel mini-batch SGD and 2) offline WA, which averages the weights of one model at different checkpoints, is typically used to improve the generalization ability of DNNs. Though online and offline WA are similar in form, they are seldom associated with each other. Besides, these methods typically perform either offline parameter averaging or online parameter averaging, but not both. In this work, we first attempt to incorporate online and offline WA into a general training framework termed hierarchical WA (HWA). By leveraging both the online and offline averaging manners, HWA is able to achieve both faster convergence speed and superior generalization performance without any fancy learning rate adjustment. Besides, we also analyze the issues faced by the existing WA methods, and how our HWA addresses them, empirically. Finally, extensive experiments verify that HWA outperforms the state-of-the-art methods significantly.

Utilizing multiple inputs autoregressive models for bearing remaining useful life prediction

Accurate prediction of the remaining useful life (RUL) of rolling bearings is crucial in industrial production, yet existing models often struggle with limited generalization capabilities due to their inability to fully process all vibration signal patterns. We introduce a novel multi-input autoregressive model to address this challenge in RUL prediction for bearings. Our approach uniquely integrates vibration signals with previously predicted RUL values, employing feature fusion to output current window RUL values. Through autoregressive iterations, the model attains a global receptive field, effectively overcoming the limitations in generalization. Furthermore, we innovatively incorporate a segmentation method and multiple training iterations to mitigate error accumulation in autoregressive models. Empirical evaluation on the PMH2012 dataset demonstrates that our model, compared to other backbone networks using similar autoregressive approaches, achieves significantly lower root mean square error (RMSE) and Score. Notably, it outperforms traditional autoregressive models that use label values as inputs and non-autoregressive networks, showing superior generalization abilities with a marked lead in RMSE and Score metrics.

Analyzing the heterogenous effects of factors on high-range speeding likelihood of taxi speeders: Does explainable deep learning provides more insights than random parameter approach?

The random parameters Generalized Linear Model (GLM) is frequently used to model speeding characteristics and capture the heterogenous effects of factors. However, this statistical approach is seldom employed for prediction and generalization due to the challenge of transferring its predefined errors. Recently, the emergence of explainable AI techniques has illuminated a new path for analyzing factors associated with risky driving behaviors. Despite this, there remains a gap that comparing results from machine and deep learning (ML/DL) approaches with those from random parameters GLM. This study aims to apply the random parameter GLM and explainable deep learning to evaluate the heterogenous effects of factors on the taxis’ high-range speeding likelihood. Initially, a Beta GLM with random parameters (BGLM-RP) is developed to model the high-range speeding likelihood among taxi drivers. Additionally, XGBoost, a simple convolutional neural network (Simple-CNN), a deeper CNN (DCNN), and a deeper CNN with self-attention (DCNN-SA) are developed. The quantified explanations and illustrations of the factors’ heterogenous effects from ML/DL models are derived from pseudo coefficients by decomposing factors’ SHapley Additive exPlanations (SHAP) values. All the developed statistical, ML, and DL models are compared in terms of mean absolute errors and mean square errors on testing and full data. Results show that DCNN-SA excels in prediction on testing data, indicating its superior generalization capabilities, while BGLM-RP outperforms other models on full data. The DCNN-SA can reveal the heterogenous effects of factors for both in-sample and out-of-sample data, which is not possible for the random parameter GLM. However, BGLM-RP can reveal larger magnitudes of the factors’ heterogenous effects for in-sample data. The signs and significances are identical between the varying coefficients from BGLM-RP and the pseudo coefficients from the ML/DL models, demonstrating the validity and rationale of using the proposed explanation framework to quantify the factors’ effects in ML/DL models. The study also discusses the contributions of various factors to the high-range speeding likelihood of taxi drivers.

Continuous, Learned Imputation of Missing Values in Parkinson's Disease.

Parkinson's disease management requires accurate clinical scores but suffers from missing data. Leveraging self-supervised learning, we demonstrate superior generalization capabilities across populations compared to other well-established imputation techniques (MIWAE, MissForest, MICE). With the ability to employ the method already during the data collection and not afterward, the technology allows more robust data collection in clinical reality.

Confinement strength prediction of corroded rectangular concrete columns using BP neural networks and support vector regression

The application of machine learning in predicting the mechanical performance of reinforced concrete columns subjected to reinforcement corrosion was investigated in this study. Traditional models, often relying on empirical formulas or simplified assumptions, are known to struggle in capturing all variables and relationships in complex problems, leading to inadequate prediction accuracy. Through theoretical analysis, the confinement effect of stirrups on core concrete and its performance deterioration caused by reinforcement corrosion were determined. To improve the prediction accuracy and generalization ability, the error Back Propagation Neural Network (BPNN) and Support Vector Regression (SVR) models were employed. Parameter effects in the BPNN model were interpreted using the SHapley Additive exPlanations (SHAP) and compared with traditional parameter sensitivity analyses. The investigation indicates that the machine learning models have a high degree of fit and low error levels, particularly within the training set. After analyzing using the SHAP approach, it is found that the corrosion ratio of the stirrup and longitudinal reinforcement have a significant effect on the confinement strength of the concrete, whereas the effects of the stirrup configuration and size effect are negligible, which is different from conventional parametric sensitivity analysis. As to the stability analysis of model predictions, the SVR model shows lower volatility and higher prediction stability compared with the BPNN model, despite the latter occasionally providing superior generalization ability. This research underscores the significant advantages offered by machine learning models in addressing longstanding challenges associated with strength prediction, and thus presents new perspectives for future study.

A 3D boundary-guided hybrid network with convolutions and Transformers for lung tumor segmentation in CT images

—Accurate lung tumor segmentation from Computed Tomography (CT) scans is crucial for lung cancer diagnosis. Since the 2D methods lack the volumetric information of lung CT images, 3D convolution-based and Transformer-based methods have recently been applied in lung tumor segmentation tasks using CT imaging. However, most existing 3D methods cannot effectively collaborate the local patterns learned by convolutions with the global dependencies captured by Transformers, and widely ignore the important boundary information of lung tumors. To tackle these problems, we propose a 3D boundary-guided hybrid network using convolutions and Transformers for lung tumor segmentation, named BGHNet. In BGHNet, we first propose the Hybrid Local-Global Context Aggregation (HLGCA) module with parallel convolution and Transformer branches in the encoding phase. To aggregate local and global contexts in each branch of the HLGCA module, we not only design the Volumetric Cross-Stripe Window Transformer (VCSwin-Transformer) to build the Transformer branch with local inductive biases and large receptive fields, but also design the Volumetric Pyramid Convolution with transformer-based extensions (VPConvNeXt) to build the convolution branch with multi-scale global information. Then, we present a Boundary-Guided Feature Refinement (BGFR) module in the decoding phase, which explicitly leverages the boundary information to refine multi-stage decoding features for better performance. Extensive experiments were conducted on two lung tumor segmentation datasets, including a private dataset (HUST-Lung) and a public benchmark dataset (MSD-Lung). Results show that BGHNet outperforms other state-of-the-art 2D or 3D methods in our experiments, and it exhibits superior generalization performance in both non-contrast and contrast-enhanced CT scans.

An analytic formulation of convolutional neural network learning for pattern recognition

Training convolutional neural networks (CNNs) using back-propagation (BP) is a time-consuming and resource-intensive process, primarily due to the need to iterate over the dataset multiple times. In contrast, analytic learning aims to train neural networks in a single epoch, offering a potential solution to these challenges. However, existing studies of analytic learning have been limited to multilayer perceptrons (MLPs). In this article, we propose an analytic formulation for convolutional neural network learning (ACnnL), which represents a significant advancement towards non-iterative learning paradigms for CNNs. Our formulation demonstrates that ACnnL extends the principles of MLP regularization constraints. From the implicit regularization and network interpretability viewpoints, we provide insights into why CNNs often exhibit superior generalization capabilities. The ACnnL is validated by conducting classification tasks on benchmark datasets such as MNIST, FashionMNIST, CIFAR10, CIFAR100 and Tiny-ImageNet. It is encouraging that the ACnnL trains CNNs in a significantly fast manner with reasonably close prediction accuracies to those using BP. In particular, a 5-layer vanilla CNN trained by ACnnL gave an accuracy of 0.9931, 0.9155, 0.7049 and 0.4628 for these datasets. The ACnnL achieves training speeds that are approximately 17 times faster than BP on GPU and 113 times faster than BP on CPU, while maintaining competitive prediction accuracies. Moreover, our experiments disclose a unique advantage of ACnnL under the small-sample scenario when training data are scarce or expensive. In a nutshell, an analytic method which deals well with small-sample size data has been put forward for the first time for fast CNN training with inherent network interpretability.

Gaussian process regression to predict dryout incipience quality of saturated flow boiling in mini/micro-channels

In the present study, a Gaussian process regression (GPR) model is developed to predict the dryout incipience quality for flow boiling in mini / micro – channels based on a consolidated database obtained from Purdue University Boiling and Two – Phase Flow Laboratory (PU – BTPFL) consisting of 997 points from 26 sources. The database includes 13 different working fluids over a wide range of operating conditions: hydraulic diameter (0.51 – 6.0 mm), mass velocities (29 – 2303 kg/ m2 s), liquid – only Reynolds number (125 – 53,770), boiling number (0.31 – 44.3 × 10-4), and reduced pressure (0.005 – 0.78). The inputs to the model were liquid – only Weber number (Wefo), reduced pressure (PR), boiling number (Bo), heated to frictional perimeter ratio(PH/PF), capillary number (Ca) and density ratio (ρg/ρf), and the output was dryout incipience quality. The database was randomly divided into training data to learn GPR kernel (covariance) parameters using maximum likelihood estimation, and test data to evaluate the prediction accuracy of the outputs based on mean absolute error (MAE). Six – different types of covariance functions were tested, and GPR model with automatic relevance detection (ARD) rational quadratic covariance function showed the best overall performance. A performance comparison of approved GPR model was made with an existing highly reliable universal correlation to predict dryout incipience quality for mini / micro – channels. Results show that the developed GPR model exhibits superior generalization ability with an overall MAE of 6.03 %, and a significant reduction of 51.76 % in MAE compared with the universal correlation. Overall, the GPR model was found to be a data efficient machine learning technique for predicting dryout incipience quality for flow boiling in mini / micro – channels based on a consolidated database.

Soil Cadmium Prediction and Health Risk Assessment of an Oasis on the Eastern Edge of the Tarim Basin Based on Feature Optimization and Machine Learning

Soil heavy metal pollution poses a serious threat to food security, human health, and soil ecosystems. Based on 644 soil samples collected from a typical oasis located at the eastern margin of the Tarim Basin, a series of models, namely, multiple linear regression （LR）, neural network （BP）, random forest （RF）, support vector machine （SVM）, and radial basis function （RBF）, were built to predict the soil heavy metal content. The optimal prediction result was obtained and utilized to analyze the spatial distribution features of heavy metal contamination and relevant health risks. The outcomes demonstrated that： ① The average Cd content in the study area was 0.14 mg·kg-1, which was 1.17 times the soil background value of Xinjiang, making it the primary factor of soil heavy metal contamination in the area. Additionally, the carcinogenicity risk coefficients of Cd for both adults and children were less than 10-4, indicating that there were no significant long-term health risks for humans in the area. ② The estimation accuracies of the five inversion models were compared, and the validation set of the RF model had an R2 value of 0.763 7, which was the highest among the five models. Additionally, the RMSE, MAE, and MBE of the RF model were the smallest among the five models. Therefore, the predicted values of the RF model were most consistent with the measured values of the soil Cd content. The predicted map of soil Cd distribution derived from the RF model coincided best with the interpolation map. ③ The RF model outperformed the other four models in predicting health risks associated with the soil Cd element for both adults and children, resulting in better prediction results. Comparatively, the predicted values of the LR model in the validation set varied greatly, leading to unreliable results. It was demonstrated that the RF was the best model for predicting soil Cd content and evaluating health risks in the study area, considering its superior generalization capability and anti-overfitting ability.

MSRA-Net: multi-channel semantic-aware and residual attention mechanism network for unsupervised 3D image registration

Objective. Convolutional neural network (CNN) is developing rapidly in the field of medical image registration, and the proposed U-Net further improves the precision of registration. However, this method may discard certain important information in the process of encoding and decoding steps, consequently leading to a decline in accuracy. To solve this problem, a multi-channel semantic-aware and residual attention mechanism network (MSRA-Net) is proposed in this paper. Approach. Our proposed network achieves efficient information aggregation by cleverly extracting the features of different channels. Firstly, a context-aware module (CAM) is designed to extract valuable contextual information. And the depth-wise separable convolution is employed in the CAM to alleviate the computational burden. Then, a new multi-channel semantic-aware module (MCSAM) is designed for more comprehensive fusion of up-sampling features. Additionally, the residual attention module is introduced in the up-sampling process to extract more semantic information and minimize information loss. Main results. This study utilizes Dice score, average symmetric surface distance and negative Jacobian determinant evaluation metrics to evaluate the influence of registration. The experimental results demonstrate that our proposed MSRA-Net has the highest accuracy compared to several state-of-the-art methods. Moreover, our network has demonstrated the highest Dice score across multiple datasets, thereby indicating that the superior generalization capabilities of our model. Significance. The proposed MSRA-Net offers a novel approach to improve medical image registration accuracy, with implications for various clinical applications. Our implementation is available at https://github.com/shy922/MSRA-Net.

Fast flow field prediction of pollutant leakage diffusion based on deep learning.

Predicting pollutant leakage and diffusion processes is crucial for ensuring people's safety. While the deep learning method offers high simulation efficiency and superior generalization, there is currently a lack of research on predicting pollutant leakage and diffusion flow field using deep learning. Therefore, it is necessary to conduct further studies in this area. This paper introduces a two-level network method to model the flow characteristics of pollutant diffusion. The proposed method in this study demonstrates a significant enhancement in flow field prediction accuracy compared to traditional deep learning methods. Moreover, it improves computational efficiency by over 800 times compared to traditional computational fluid dynamics (CFD) methods. Unlike conventional CFD methods that require grid expansion to calculate all operation conditions, the deep learning method is not confined by grid limitations. While deep learning methods may not entirely replace CFD methods, they can serve as a valuable supplementary tool, expanding the versatility of CFD methods. The findings of this research establish a robust foundation for incorporating deep learning methods in addressing pollutant leakage and diffusion challenges.

Research on the Maturity Detection Method of Korla Pears Based on Hyperspectral Technology

In this study, hyperspectral imaging technology with a wavelength range of 450 to 1000 nanometers was used to collect spectral data from 160 Korla pear samples at various maturity stages (immature, semimature, mature, and overripe). To ensure high-quality data, multiple preprocessing techniques such as multiplicative scatter correction (MSC), standard normal variate (SNV), and normalization were employed. Based on these preprocessed data, a custom convolutional neural network model (CNN-S) was constructed and trained to achieve precise classification and identification of the maturity stages of Korla pears. Additionally, a BP neural network model was used to determine the characteristic wavelengths for maturity assessment based on the sugar content feature wavelengths. The results demonstrated that the BP model, based on sugar content feature wavelengths, effectively discriminated the maturity stages of the pears. Specifically, the comprehensive recognition rates for the training, testing, and validation sets were 98.5%, 93.5%, and 90.5%, respectively. Furthermore, the combination of hyperspectral imaging technology and the custom CNN-S model significantly enhanced the detection performance of pear maturity. Compared to traditional CNN models, the CNN-S model improved the accuracy of the test set by nearly 10%. Moreover, the CNN-S model outperformed existing techniques based on partial least squares discriminant analysis (PLS-DA) and support vector machine (SVM) in capturing hyperspectral data features, showing superior generalization capability and detection efficiency. The superior performance of this method in practical applications further supports its potential in smart agriculture technology, providing a more efficient and accurate solution for agricultural product quality detection. Additionally, it plays a crucial role in the development of smart agricultural technology.

Predicting Wall Pressure in Shock Wave/Boundary Layer Interactions with Convolutional Neural Networks

Within the dynamic realm of variable-geometry shock wave/boundary layer interactions, the wall parameters of the flow field undergo real-time fluctuations. The conventional approach to sensing these changes in wall pressure through sensor measurements is encumbered by a cumbersome process, leading to diminished efficiency and an inability to provide swift predictions of wall parameters. This paper introduces a data-driven methodology that leverages non-contact schlieren imaging to predict wall pressure within the flow field, a technique that holds promise for informing the optimized design of variable-geometry systems. A sophisticated deep learning framework, predicated on Convolutional Neural Networks (CNN), has been engineered to anticipate alterations in wall pressure stemming from high-speed shock wave/boundary layer interactions. Utilizing an impulsive wind tunnel with a Mach number of 6, we have procured a sequence of schlieren images and corresponding wall pressure measurements, capturing the continuous variations induced by an attack angle from a shock wave generator. These data have been instrumental in compiling a comprehensive dataset for the training and evaluation of the CNN. The CNN model, once trained, has adeptly deduced the distribution of wall pressure from the schlieren imagery. Notwithstanding, it was observed that the CNN’s predictive prowess is marginally diminished in regions where pressure variations are most pronounced. To assess the model’s generalization capabilities, we have segmented the dataset according to different temporal intervals for network training. Our findings indicate that while the generalization of all models crafted was less than optimal, Model 4 demonstrated superior generalization. It is thus suggested that augmenting the training set with additional samples and refining the network architecture will be a worthwhile endeavor in subsequent research initiatives.

Bayesian optimization-based Model-Agnostic Meta-Learning: Application to predict maximum cyclic moment resistance of steel bolted T-stub connections

Accurately assessing the maximum moment resistance of steel bolted T-stub connections under cyclic loading is crucial for designing earthquake-resistant structures with such connections. Traditional methods based on design standards like ASCE 41-17 often lack precision. Recently, supervised machine learning techniques, particularly Artificial Neural Network (ANN), have been explored. However, conventional ANNs require substantial data for generalization, which is limited for steel bolted T-stub connections. To address these challenges, this study explores the feasibility of using the Model-Agnostic Meta-Learning (MAML) to predict the maximum cyclic moment resistance of steel bolted T-stub connections. MAML adapts task-specific model parameters rapidly and transfers knowledge across tasks to fine-tune global model parameters, potentially enhancing prediction accuracy with limited data. The MAML model is first optimized using Bayesian optimization with a Gaussian Process model to identify ideal hyperparameters. The optimized MAML model is then compared with two ANN models (one with optimized hyperparameters, another matching MAML's neural network architecture) and ASCE 41-17 method. Results demonstrate the optimized MAML model's superior generalization capabilities, offering a promising approach for steel bolted T-stub connections.

Benchmarking building energy consumption for space heating using an empirical Bayesian approach with urban-scale energy model

Calibration of an urban-scale building energy model is crucial for advancing building energy efficiency codes and policies to reduce carbon dioxide emissions whereas the accuracy of the model is challenged by performance gaps resulting from sparse data and uncertainties in modeling. An empirical Bayesian (EB) approach, incorporated a probability-based mixed-effects energy model and Monte Carlo Markov chain simulation, was developed to predict and benchmark building energy use intensity (EUI) for space heating, using the metering data of 2020 stochastically sampled district heating substations from a municipal utility database. The limitations, implementations and performances of different modeling strategy, including EB as well as Maximum Likelihood Estimation (MLE), Maximum A Posteriori (MAP) and Fully Bayes (FB) were compared. The results show that EB strategy achieved superior prediction performance than other strategies on the novel test dataset. The preparation, calibration, and validation efforts for implementing the EB approach were demonstrated. The proposed methodology is recommended as a robust approach for benchmarking urban-scale building EUI, distinguished from other machine learning approaches by requiring fewer training samples, better interpreting uncertainties in the data used, and exhibiting superior generalization capabilities. The proposed methodology was successfully demonstrated through the urban-scale benchmarking of space heating EUI in Beijing.

Deep learning with improved hybrid neuro-turning for predictive control of flux based on experimental DCMD module design of water desalination system

This study explores two scenarios for optimizing the predictive control of flux in water desalination systems: experimental design of direct contact membrane distillation (DCMD) and artificial intelligence (AI) models. Deep learning networks, specifically Long Short-Term Memory (LSTM), and standalone AI models, including hybrid Adaptive Neuro-Fuzzy Inference System (ANFIS) and Artificial Neural Network (ANN), were evaluated. Evaluation criteria, along with 2D graphical visualization, were used to assess model accuracy. ANFIS-C4 emerged as the standout model, achieving perfect predictive skills with 100 % accuracy and low RMSE of 0.0522 in the training phase, and maintaining high performance in testing with 99.73 % accuracy and RMSE of 0.7121. Similarly, ANN-C4 demonstrated near-perfect accuracy with 100 % accuracy and RMSE of 0.0643 in the training phase and also maintained excellent performance in the testing phase. Despite requiring multiple inputs, ANFIS and ANN show remarkably high capability in predictive accuracy. Among LSTM models, LSTM-C3 achieved the highest training DC of 91.35 %, indicating robust performance in capturing training data variability. LSTM-C2 showed superior generalization with a testing DC of 0.8539, despite using only two input parameters. The narrow distribution of predicted values around the median in violin plots for ANFIS-C4 and ANN-C4, along with their low RMSE, validates their superior performance. The exceptional results of ANFIS-C4 and ANN-C4 highlight their potential in water desalination applications, supporting Sustainable Development Goal 6 (SDGs-6) and Environmental Protection Agency (EPA) objectives for sustainable water management. This research underscores the importance of advanced computational models in enhancing the efficiency and reliability of water desalination systems, contributing to global efforts in sustainable water resource management.