Network Learning Research Articles

“I sit here thinking I can do this”—Developing justice‐centered ambitious science teaching identities in professional learning communities

AbstractExplicitly attending to justice in science teaching and learning is long overdue. Here, we examined the professional teacher identity development of 13 science teachers as they collaborated in networked professional learning communities (PLCs) to implement and revise a culture‐setting unit focused on the science of COVID and engaging in collaborative inquiry cycles to identify and refine justice‐centered ambitious science teaching (JuST) practices in classrooms. These JuST practices are conceptualized as a synthesis of justice‐centered pedagogies and ambitious science teaching. To accomplish our research aims, we drew on qualitative methods, where we relied on transcribed video recordings of 21 PLC meetings and three transcribed end‐of‐year focus group interviews of two PLCs (i.e., 13 secondary science teachers, mostly white) across a year‐long period. As a result of our analyses, we identified how professional learning arranged through PLCs, culture‐setting units, and collaborative inquiries supported professional JuST identity development by, among other affordances, providing space for the critical and emotional work of learning to discuss race, affording teachers strategies for getting to know their students and the assets they bring to classrooms, and recognition and positioning of teachers as professionals capable of identifying or developing, refining, and contributing to knowledge about JuST science teaching and learning. Challenges identified included, among others, identifying how the culture‐setting units could be effectively integrated into existing curriculum maps and the uneven implementation support from administrators. In the end, what is revealed helps better conceptualize how engaging teachers in PLCs around tasks like curriculum implementation and refinement or collaborative inquires support professional JuST identity development and how such experiences can be more carefully negotiated.

Advancing remote sensing: a unified deep learning approach with pretrained and custom architectures for high-precision classification

Advancement of remote sensing is vital for accurate land cover mapping and ecological surveillance. This research proposes a cutting-edge deep learning framework that integrates a meticulously designed customary network architecture with the effectiveness of pre-trained models, namely GoogleNet, VGG16, and InceptionV3. Our approach captures the complex features of RGB satellite images throughout various land cover categories using diverse EuroSAT datasets. Our system achieves an impressive 99.40% test accuracy by implementing feature extraction through effective models with a customized deep learning network in an intuitive manner. High class-wise accuracies are observed, ranging from 96.00% to 100.00%, whereas F1-scores, precision, and recall all converge at 99.40%. The aforementioned results demonstrate our approach’s perspective to significantly improve analysis while also confirming its superiority. These results not only confirm the performance of our algorithm but also show that it could be used to redesign satellite visual analysis behavior. These results open the window to further improve and accurate remote sensing systems, giving a helpful viewpoint for researchers as well as professionals.

Identification and Counting of Pirapitinga Piaractus brachypomus Fingerlings Fish Using Machine Learning.

Identifying and counting fish are crucial for managing stocking, harvesting, and marketing of farmed fish. Researchers have used convolutional networks for these tasks and explored various approaches to enhance network learning. Batch normalization is one technique that improves network stability and accuracy. This study aimed to evaluate machine learning for identifying and counting pirapitinga Piaractus brachypomus fry with different batch sizes. The researchers used one thousand photographic images of Pirapitinga fingerlings, labeled with bounding boxes. They trained the adapted convolutional network model with batch normalization layers added at the end of each convolution block. They set the training to one hundred and fifty epochs and tested batch sizes of 5, 10, and 20. Furthermore, they measured network performance using precision, recall, and mAP@0.5. Models with smaller batch sizes performed less effectively. The training with a batch size of 20 achieved the best performance, with a precision of 96.74%, recall of 95.48%, mAP@0.5 of 97.08%, and accuracy of 98%. This indicates that larger batch sizes improve accuracy in detecting and counting pirapitinga fry across different fish densities.

Music Generation Using Dual Interactive Wasserstein Fourier Acquisitive Generative Adversarial Network

Music composition, an intricate blend of human creativity and emotion, presents substantial challenges when generating melodies from lyrics which hinders effective learning in neural networks and the inadequate depiction of harmonic structure that fails to encapsulate the complex relationships between lyrics and melodies. The existing methods often struggle to balance emotional depth and structural coherence, leading to compositions that lack both the intended emotional resonance and musical consistency. To overcome these issues, this research introduces a novel approach named Dual Interactive Wasserstein Fourier Acquisitive Generative Adversarial Network (DIWFA-GAN), which integrates innovative techniques like swish activation functions and the Giant Trevally Optimizer (GTO) for parameter optimization. Meanwhile, the GTO, inspired by the movement patterns of the Giant Trevally fish, provides efficient and effective parameter optimization, improving the model’s convergence speed and accuracy. Comparative analysis against recent existing models reveals superior performance for both the LMD-full MIDI and Reddit MIDI datasets, with impressive metrics including inception scores of 9.36 and 2.98, Fréchet inception distances of 35.29 and 135.54 and accuracies of 99.98% and 99.95%, respectively. The DIWFA-GAN significantly outperforms existing models in generating high-fidelity melodies, as evidenced by superior inception scores, Fréchet inception distances, and accuracies on both datasets.

Intrusion Detection with Federated Learning and Conditional Generative Adversarial Network in Satellite-Terrestrial Integrated Networks

Intrusion Detection with Federated Learning and Conditional Generative Adversarial Network in Satellite-Terrestrial Integrated Networks

Deep-Learning-Based Amplitude Variation with Angle Inversion with Multi-Input Neural Networks

Deep-learning-based (DL-based) seismic inversion has emerged as one of the state-of-the-art research areas in exploration geophysics with the development of artificial intelligence technology. Due to its good portability and high computational efficiency, this method has emerged as a data-driven approach for estimating subsurface properties. However, most of the current DL-based methods rely solely on seismic data, lacking the incorporation of prior information. In addition, these methods are usually performed trace-by-trace, resulting in insufficient horizontal constraints. These limitations make traditional methods less robust, particularly when dealing with high noise levels or limited data. To address these challenges, we propose a multi-input deep learning network for pre-stack inversion, which combines data-driven and model-driven approaches for optimization. The proposed method separately extracts features from the model and data, merging them to improve feature utilization. Moreover, by adopting a 2-D training unit, rather than a trace-by-trace approach, the method improves the horizontal continuity of the results. Tests on synthetic and real seismic data confirmed the robustness and improved stability of the proposed method, even under challenging conditions. This dual-driven approach significantly enhances the reliability of seismic inversion.

ShiftTransUNet: An Efficient Deep Learning Model for Medical Image Segmentation Using ShiftViT Framework

Deep learning has significantly advanced the field of medical image segmentation. However, the complexity of network structures often leads to high computational demands, limiting their practical efficiency. To enhance the efficiency of image segmentation, this paper introduces an innovative, concise, and lightweight deep learning network. First, to reduce model complexity, we replaced the attention mechanism in the traditional vision transformer (ViT) structure with a shift operation, creating the ShiftViT architecture. This substitution significantly decreased computation and the number of parameters while preserving model performance. Second, to retain and enhance fine-grained features and facilitate more precise information transfer across different layers, we employed a full-scale progressive skip connection strategy. This approach effectively integrates multi-scale feature information, further enhancing model performance. Additionally, to further reduce network complexity, inspired by the independence of probabilities, we opted for depth-wise separable convolution over traditional convolution. This enhances the relative independence between layers. Together, these modifications achieved superior segmentation results on both the Synapse and Automated Cardiac Diagnostic Challenge (ACDC) datasets compared to mainstream models, with substantial advantages in terms of computational efficiency and parameter count. The proposed approach represents an effective solution for medical image applications with limited computational resources and holds great promise for clinical practice.

Advanced AI Technologies for Defect Prevention and Yield Optimization in PCB Manufacturing

The integration of Advanced Artificial Intelligence (AI) technologies in Printed Circuit Board (PCB) manufacturing has revolutionized the industry, offering significant improvements in defect prevention and yield optimization. As PCB manufacturing processes become increasingly complex and intricate, traditional methods of defect detection and quality assurance are often insufficient to meet the high demand for precision and efficiency. This paper explores the application of cutting-edge AI methodologies, including machine learning, computer vision, and neural networks, to address critical challenges in PCB production. The primary focus of this research is on the role of AI in enhancing automated defect detection systems and optimizing production yield. AI-based systems such as predictive maintenance, real-time process monitoring, and intelligent decision-making have been shown to drastically reduce defects while improving overall production throughput. Machine learning algorithms can identify subtle defects in real time, often undetectable by conventional methods, while neural networks can analyze historical data to predict potential failures before they occur. Additionally, AI-driven optimization techniques help manufacturers adjust production parameters dynamically, ensuring higher yields and minimizing waste. Through a combination of theoretical analysis and case studies, this paper highlights the effectiveness of AI-driven solutions in PCB manufacturing. Results from industry applications indicate significant improvements in both quality control and yield rates, providing a competitive edge to manufacturers adopting these technologies. Furthermore, this paper discusses future trends, including the integration of AI with the Internet of Things (IoT), edge computing, and sustainable manufacturing practices, which will further enhance the capabilities of AI systems in this field. The findings suggest that advanced AI technologies are not only capable of overcoming existing challenges in defect prevention but also hold the potential to reshape the future of PCB manufacturing by offering highly adaptive, precise, and efficient solutions. This research underscores the transformative impact of AI in modernizing manufacturing processes, making it an indispensable tool for the future of the PCB industry.

Utilising Artificial Intelligence to Predict Membrane Behaviour in Water Purification and Desalination

Water scarcity is a critical global issue, necessitating efficient water purification and desalination methods. Membrane separation methods are environmentally friendly and consume less energy, making them more economical compared to other desalination and purification methods. This survey explores the application of artificial intelligence (AI) to predict membrane behaviour in water purification and desalination processes. Various AI platforms, including machine learning (ML) and artificial neural networks (ANNs), were utilised to model water flux, predict fouling behaviour, simulate micropollutant dynamics and optimise operational parameters. Specifically, models such as convolutional neural networks (CNNs), recurrent neural networks (RNNs) and support vector machines (SVMs) have demonstrated superior predictive capabilities in these applications. This review studies recent advancements, emphasising the superior predictive capabilities of AI models compared to traditional methods. Key findings include the development of AI models for various membrane separation techniques and the integration of AI concepts such as ML and ANNs to simulate membrane fouling, water flux and micropollutant behaviour, aiming to enhance wastewater treatment and optimise treatment and desalination processes. In conclusion, this review summarised the applications of AI in predicting the behaviour of membranes as well as their strengths, weaknesses and future directions of AI in membranes for water purification and desalination processes.

Attentive graph neural network models for the prediction of blood brain barrier permeability.

The blood brain barrier's (BBB) unique endothelial cells and tight junctions selectively regulate passage of molecules to the central nervous system (CNS) to prevent pathogen entry and maintain neural homeostasis. Various neurological conditions and neurodegenerative diseases benefit from small molecules capable of BBB penetration (BBBP) to elicit a therapeutic effect. Predicting BBBP often involves in silico assessment of molecular properties such as lipophilicity (log P ) and polar surface area (PSA) using the CNS multiparameter optimization (MPO) method. This study curated an open-source dataset to benchmark rigorously machine learning (ML) and neural network (NN) models with each other and with MPO for predicting BBBP. Our analysis demonstrated that AI models, especially attentive NNs using stereochemical features, significantly outperform MPO in predicting BBBP. An attentive graph neural network (GNN), we refer to as CANDID-CNS™, achieved a 0.23-0.26 higher AUROC score than MPO on full test sets, and a 0.17-0.19 higher score on stereoisomers filtered subsets. Regarding stereoisomers that differ in BBBP, which MPO cannot distinguish, attentive GNNs correctly classify these with AUROC and MCC metrics comparable to or better than MPO's AUROC and MCC on less difficult test molecules. These findings suggest that integrating attentive GNN models into pharmaceutical drug discovery processes can substantially improve prediction rates, and thereby reduce the timeline, cost, and increase probability of success of designing brain penetrant therapeutics for the treatment of a wide variety of neurological and neurodegenerative diseases.

Enhancing UAV Swarm Tactics with Edge AI: Adaptive Decision Making in Changing Environments

This paper presents a drone system that uses an improved network topology and MultiAgent Reinforcement Learning (MARL) to enhance mission performance in Unmanned Aerial Vehicle (UAV) swarms across various scenarios. We propose a UAV swarm system that allows drones to efficiently perform tasks with limited information sharing and optimal action selection through our Efficient Self UAV Swarm Network (ESUSN) and reinforcement learning (RL). The system reduces communication delay by 53% and energy consumption by 63% compared with traditional MESH networks with five drones and achieves a 64% shorter delay and 78% lower energy consumption with ten drones. Compared with nonreinforcement learning-based systems, mission performance and collision prevention improved significantly, with the proposed system achieving zero collisions in scenarios involving up to ten drones. These results demonstrate that training drone swarms through MARL and optimized information sharing significantly increases mission efficiency and reliability, allowing for the simultaneous operation of multiple drones.

Deep learning architecture for scatter estimation in cone-beam computed tomography head imaging with varying field-of-measurement settings.

X-ray scatter causes considerable degradation in the cone-beam computed tomography (CBCT) image quality. To estimate the scatter, deep learning-based methods have been demonstrated to be effective. Modern CBCT systems can scan a wide range of field-of-measurement (FOM) sizes. Variations in the size of FOM can cause a major shift in the scatter-to-primary ratio in CBCT. However, the scatter estimation performance of deep learning networks has not been extensively evaluated under varying FOMs. Therefore, we train the state-of-the-art scatter estimation neural networks for varying FOMs and develop a method to utilize FOM size information to improve performance. We used FOM size information as additional features by converting it into two channels and then concatenating it to the encoder of the networks. We compared our approach for a U-Net, Spline-Net, and DSE-Net, by training them with and without the FOM information. We utilized a Monte Carlo-simulated dataset to train the networks on 18 FOM sizes and test on 30 unseen FOM sizes. In addition, we evaluated the models on the water phantoms and real clinical CBCT scans. The simulation study demonstrates that our method reduced average mean-absolute-percentage-error for U-Net by 38%, Spline-Net by 40%, and DSE-net by 33% for the scatter estimation in the 2D projection domain. Furthermore, the root-mean-square error on the 3D reconstructed volumes was improved for U-Net by 43%, Spline-Net by 30%, and DSE-Net by 23%. Furthermore, our method improved contrast and image quality on real datasets such as water phantom and clinical data. Providing additional information about FOM size improves the robustness of the neural networks for scatter estimation. Our approach is not limited to utilizing only FOM size information; more variables such as tube voltage, scanning geometry, and patient size can be added to improve the robustness of a single network.

Differential diagnosis of congenital ventricular septal defect and atrial septal defect in children using deep learning–based analysis of chest radiographs

BackgroundChildren with atrial septal defect (ASD) and ventricular septal defect (VSD) are frequently examined for respiratory symptoms, even when the underlying disease is not found. Chest radiographs often serve as the primary imaging modality. It is crucial to differentiate between ASD and VSD due to their distinct treatment.PurposeTo assess whether deep learning analysis of chest radiographs can more effectively differentiate between ASD and VSD in children.MethodsIn this retrospective study, chest radiographs and corresponding radiology reports from 1,194 patients were analyzed. The cases were categorized into a training set and a validation set, comprising 480 cases of ASD and 480 cases of VSD, and a test set with 115 cases of ASD and 119 cases of VSD. Four deep learning network models—ResNet-CBAM, InceptionV3, EfficientNet, and ViT—were developed for training, and a fivefold cross-validation method was employed to optimize the models. Receiver operating characteristic (ROC) curve analyses were conducted to assess the performance of each model. The most effective algorithm was compared with the interpretations provided by two radiologists on 234 images from the test group.ResultsThe average accuracy, sensitivity, and specificity of the four deep learning models in the differential diagnosis of VSD and ASD were higher than 70%. The AUC values of ResNet-CBAM, IncepetionV3, EfficientNet, and ViT were 0.87, 0.91, 0.90, and 0.66, respectively. Statistical analysis showed that the differential diagnosis efficiency of InceptionV3 was the highest, reaching 87% classification accuracy. The accuracy of InceptionV3 in the differential diagnosis of VSD and ASD was higher than that of the radiologists.ConclusionsDeep learning methods such as IncepetionV3 based on chest radiographs in the study showed good performance for differential diagnosis of congenital VSD and ASD, which may be able to assist radiologists in diagnosis, education, and training, and reduce missed diagnosis and misdiagnosis.

Low-Priori Intervention Transfer Learning Network for Large-Span Fault Diagnosis

Low-Priori Intervention Transfer Learning Network for Large-Span Fault Diagnosis

Symposium 3: Content and Language Integrated Learning

European policies mandate encouraging plurilingualism in a digitally enhanced world. This mandate is placing increased demands on higher educational practitioners and institutions to prepare today’s learners with new linguistic skills. Content and Language Integrated Learning (CLIL) policy appears to resonate strongly with European aspirations and goals of educating citizens and promoting lingual diversity, pluriculturism, and mobility within the European Union. Whilst offering a potential solution through the interweaving of content and language in a dual-focused educational approach, CLIL is at risk of becoming a ‘buzz word’ without evidence-based research on emerging CLIL practices. This paper suggests a framework for practice-based research in the initial steps of CLIL implementation into HE curriculum and considers expansive learning theory as a theoretical and analytical framework to advance knowledge creation. The deliberate construction of a transciplinary networked learning community is advanced as the outcome and vehicle as the first initiative for CLIL implementation. The partnering and convergence of the knowledge expertise of language experts and subject experts in collaborative reflective practice enhances networked learning within and beyond the institutional boundaries, professional development and learner multiliteracies, including languages, culture, content and digital media. The context for this study is within tertiary architectural education in France where students study architecture in the first language, French, and Language and Communication Skills in the additional language of English as a separate discipline. This lack of convergence appears at odds with the emerging trend and evolution of transdisiplinarity in architectural education and practice where academia and associated professions of architecture, design and engineering increasingly teach, practice and research collaboratively. This desk-based research first examines the significance of CLIL in the European context, its variants, along with the challenges and drawbacks in crossing disciplinary boundaries. The implications for language and disciplinary practitioners and their role are discussed. Transdisciplinary collaborative work, teaching and learning can bridge language and knowledge barriers between the different disciplines in and through the fusion of language learning of, for, and through the languages of architectural practice and content, leading to innovation in curriculum development. Relational agency, in other words calling on the capacity of individuals to jointly work and learn with other practitioners, pedagogies, theories and resources distributed within institutional settings, given that a supportive learning community is possible, can lead to enhanced professional agency, in other words the capacity to act effectively informed by appropriate professional knowledge.. This paper concludes that further research is needed on relational agency within collective activities, such as networked learning communities to advance CLIL implementation.

Give me a hint: an explicit prior based image denoising

The performance of an image denoising method highly depends on its ability to produce smoothed homogeneous regions and preserve fine-grained texture details in the restored image. Existing denoising methods rely on implicit learning of Convolutional Neural Networks (CNNs) to restore an image. We show that the implicitly learned representations are limited in capacity, resulting in a sub-optimal pixel estimation of homogeneous regions and loss in texture details. In this work, we argue that explicitly introducing low and high frequency information enhances the representational capacity of denoising CNNs. To this end, we propose XPrior that provides the network with two explicit priors, i.e., a Gaussian smoothed image and an Edge map. This simple yet effective integration of priors with the noisy image enhances the model’s ability to restore images with high perceived quality. The proposed approach can easily be plugged into the existing denoising methods. Comprehensive experimental results show the superiority of the proposed approach that achieves a remarkable performance gain of 1.23 dB and 1.49 dB average PSNR on BSD68 dataset over DRUNet (Zhang in IEEE Trans Pattern Anal Mach Intell 10.1109/TPAMI.2021.3088914, 2021) and DnCNN (Zhang in IEEE Trans Image Process 26:3142–3155, 2017), respectively.

Pest and Disease Video Classification with Convolutional Neural Network and Transfer Learning

The important field crops of agriculture are affected due to attack of various pests and diseases which leads to reduction in crop production. Early classification and identification of pests and diseases in plant helps farmers to take mitigation steps. To address this issue with computer vision based techniques, convolutional neural network (CNN) based deep learning models were studied for classification of pests and diseases videos. Six different CNN models were developed. Two approaches namely from scratch learning and transfer learning were used. Data augmentation techniques such as reflection, scaling, rotation, and translation were also applied to prevent the network from overfitting. The classification accuracy of 99.19%, 99.08% and 98.80% was attained in VGG19, DENSENET201 and CNN 5 Layer model. The results demonstrated that CNN models with good architecture can classify pests and diseases with good performance.

Noise attenuation in distributed acoustic sensing data using a guided unsupervised deep learning network

Distributed acoustic sensing (DAS) is a promising technology introducing a new paradigm in the acquisition of high-resolution seismic data. However, DAS data often show weak signals compared with the background noise, especially in challenging installation environments. In this study, we develop a new approach to denoise DAS data that leverages an unsupervised deep learning (DL) model, eliminating the need for labeled training data. The input DAS data undergo band-pass filtering to eliminate high-frequency content. Subsequently, a continuous wavelet transform (CWT) is performed, and the finest scale is used to guide the DL model in reconstructing the DAS signal. First, we extract 2D patches from the band-pass filtered data and the CWT scale of the data. Then, these patches are converted using an unrolling mechanism into 1D vectors to form the input of the DL model. A self-attention layer is included in each layer to extract the spatial relation between the band-pass filtered data and the CWT scale. Through an iterative process, the DL model tunes its parameters to suppress DAS noise, with the band-pass filtered data serving as the target for the network. The denoising performance of our framework is validated using field examples from the San Andreas Fault Observatory at Depth and Frontier Observatory for Research in Geothermal Energy data sets, where the data are recorded by a fiber-optic cable. Comparative analyses against three benchmark methods reveal the robust denoising performance of our framework.

Deep Learning‐Coupled Metabolic Heat Integrated Sensing System for Noninvasive Continuous Monitoring of Blood Glucose

The dynamic and continuous monitoring of blood glucose (BG) concentration is crucial for the health management of diabetic patients. Despite its importance, significant challenges remain in the development of effective BG monitoring technologies. Metabolic heat conformation (MHC) offers a promising solution due to its noninvasiveness and reliability. However, progress in MHC technology is hindered by the complexities of multisensor integration and the intricate correlation between MHC and BG. Herein, a wearable BG continuous monitoring device based on metabolic heat integrated sensing (MHIS) is developed, combined with a deep learning network to enable continuous BG detection using MHC principles. The MHIS device integrates a miniaturized circuit and intelligent secondary signal processing, allowing for the simultaneous acquisition and integrated operation of various physiological parameters, including metabolic heat production. By integrating a gate recurrent unit neural network, a model is established to facilitate continuous BG monitoring. The wearable device has a certain accuracy, and when analyzed in comparison with commercial noninvasive glucose meters, the mean absolute relative error meets international standards. The deep learning‐enhanced MHIS system proposed in this work enables noninvasive BG monitoring, paving the way for advancements in personalized healthcare management and offering new opportunities in digital healthcare consultation.

Classification of melt pool states for defect detection in laser directed energy deposition using FixConvNeXt model

Abstract Laser directed energy deposition (L-DED) has emerged as a promising technique for rapid prototyping due to its cost-effectiveness and efficiency. However, the intricate and multi-scale physics of the process hinder its widespread application. This paper addresses the challenge by focusing on real-time identification of melt pool states to detect defects early and minimize resource wastage. To achieve this, a FixConvNeXt model was developed for fast and accurate monitoring of melt pool states. This model was trained using 5000 melt pool images captured during the printing of single-track deposits from a charge-coupled device. To evaluate its performance, FixConvNeXt was compared with other models using various metrics. Experimental results demonstrated that FixConvNeXt achieved superior performance in accurately identifying melt pool states with 99.1% accuracy, while also reducing computation burden and processing time. The mechanism of classification by FixConvNeXt was explained using gradient-weighted class activation mapping. The research findings highlight the potential application of online process monitoring in L-DED. This study lays the foundation for future development of an efficient deep learning network for automatic defect detection and feedback control.