Search Results (219)

Background: Distant supervision employs external knowledge bases to automatically match with text, allowing for the automatic annotation of sentences. Although this method effectively tackles the challenge of manual labeling, it inevitably introduces noisy labels. Traditional approaches typically employ sentence-level attention mechanisms, assigning lower weights to noisy sentences to mitigate their impact. But this approach overlooks the critical importance of information flow between sentences. Additionally, previous approaches treated an entire bag as a single classification unit, giving equal importance to all features within the bag. However, they failed to recognize that different dimensions of features have varying levels of significance. Method: To overcome these challenges, this study introduces a novel network that incorporates sentence interaction and a bag-level feature enhancement (ESI-EBF) mechanism. We concatenate sentences within a bag into a continuous context, allowing information to flow freely between them during encoding. At the bag level, we partition the features into multiple groups based on dimensions, assigning an importance coefficient to each sub-feature within a group. This enhances critical features while diminishing the influence of less important ones. In the end, the enhanced features are utilized to construct high-quality bag representations, facilitating more accurate classification by the classification module. Result: The experimental findings from the New York Times (NYT) and Wiki-20m datasets confirm the efficacy of our suggested encoding approach and feature improvement module. Our method also outperforms state-of-the-art techniques on these datasets, achieving superior relation extraction accuracy. Full article

Metal additive manufacturing (MAM) has advanced significantly, yet accurately predicting clad characteristics from processing parameters remains challenging due to process complexity and data scarcity. This study introduces a novel hybrid machine learning (ML) framework that integrates validated multi-physics computational fluid dynamics simulations with experimental data, enabling prediction of clad characteristics unattainable through conventional methods alone. Our approach uniquely incorporates physics-aware features, such as volumetric energy density and linear mass density, enhancing process understanding and model transferability. We comprehensively benchmark ML models across traditional, ensemble, and neural network categories, analyzing their computational complexity through Big O notation and evaluating both classification and regression performance in predicting clad geometries and process maps. The framework demonstrates superior prediction accuracy with sub-second inference latency, overcoming limitations of purely experimental or simulation-based methods. The trained models generate processing maps with 0.95 AUC (Area Under Curve) accuracy that directly guide MAM parameter selection, bridging the gap between theoretical modeling and practical process control. By integrating physics-based simulations with ML techniques and physics-aware features, our approach achieves an R² of 0.985 for clad geometry prediction and improved generalization over traditional methods, establishing a new standard for MAM process modeling. This research advances both theoretical understanding and practical implementation of MAM processes through a comprehensive, physics-aware machine learning approach. Full article

Purpose: Sex classification is a major benchmark of previous work in learning on the structural connectome, a naturally occurring brain graph that has proven useful for studying cognitive function and impairment. While graph neural networks (GNNs), specifically graph convolutional networks (GCNs), have gained popularity lately for their effectiveness in learning on graph data, achieving strong performance in adult sex classification tasks, their application to pediatric populations remains unexplored. We seek to characterize the capacity for GNN models to learn connectomic patterns on pediatric data through an exploration of training techniques and architectural design choices. Methods: Two datasets comprising an adult BRIGHT dataset (N = 147 Hodgkin’s lymphoma survivors and N = 162 age similar controls) and a pediatric Human Connectome Project in Development (HCP-D) dataset (N = 135 healthy subjects) were utilized. Two GNN models (GCN simple and GCN residual), a deep neural network (multi-layer perceptron), and two standard machine learning models (random forest and support vector machine) were trained. Architecture exploration experiments were conducted to evaluate the impact of network depth, pooling techniques, and skip connections on the ability of GNN models to capture connectomic patterns. Models were assessed across a range of metrics including accuracy, AUC score, and adversarial robustness. Results: GNNs outperformed other models across both populations. Notably, adult GNN models achieved 85.1% accuracy in sex classification on unseen adult participants, consistent with prior studies. The extension of the adult models to the pediatric dataset and training on the smaller pediatric dataset were sub-optimal in their performance. Using adult data to augment pediatric models, the best GNN achieved comparable accuracy across unseen pediatric (83.0%) and adult (81.3%) participants. Adversarial sensitivity experiments showed that the simple GCN remained the most robust to perturbations, followed by the multi-layer perceptron and the residual GCN. Conclusions: These findings underscore the potential of GNNs in advancing our understanding of sex-specific neurological development and disorders and highlight the importance of data augmentation in overcoming challenges associated with small pediatric datasets. Further, they highlight relevant tradeoffs in the design landscape of connectomic GNNs. For example, while the simpler GNN model tested exhibits marginally worse accuracy and AUC scores in comparison to the more complex residual GNN, it demonstrates a higher degree of adversarial robustness. Full article

This study aims to identify the common indices having an impact on the SPV of BRICS countries during crises. To address this, the monthly data retrieved from the database of the Global Economic Monitor (GEM), World Bank, IMF International Financial Statistics data, and OECD in the period of January 2000 to December 2023 are analyzed in two phases. In the first phase, DM classification techniques are applied to the data to identify the best common classification technique in order to use this technique in the second phase to compare the results with Multiple Linear Regression (MLR) results. In the second phase, to account for the global financial crisis and COVID-19 crisis, the sample period is divided into two sub-periods. For those sub-periods, MLR and the best classification technique that was found in the first phase are utilized to find the common indices that have an impact on the stock price volatility during individual and both crises. The findings indicate that the Random Tree method commonly classified the data among the seven classification techniques. Regarding MLR results, no common indices were identified during the global financial crisis or the COVID-19 crisis. However, based on Random Tree classifications, the CPI price percent, National Currency, and CPI index for all items were common during the global financial crisis, whereas only the CPI price percent was common during the COVID-19 crisis. While some common indices were observed in individual crises for specific countries, no indices were consistently found across both crises. This variation is attributed to the unique nature of each crisis and the diverse economic and socio-political structures of different countries. These findings provide valuable insights for financial institutions and investors to refine financial and policy decisions based on the specific characteristics of each crisis and the indices affecting each country. Full article

Artificial neural networks are parametric machine learning models that have been applied successfully to an extended series of classification and regression problems found in the recent literature. For the effective identification of the parameters of the artificial neural networks, a series of optimization techniques have been proposed in the relevant literature, which, although they present good results in many cases, either the optimization method used is not efficient and the training error of the network is trapped in sub-optimal values, or the neural network exhibits the phenomenon of overfitting which means that it has poor results when applied to data that was not present during the training. This paper proposes an innovative technique for constructing the weights of artificial neural networks based on appropriate BNF grammars, used in the evolutionary process of Grammatical Evolution. The new procedure locates an interval of values for the parameters of the artificial neural network, and the optimization method effectively locates the network parameters within this interval. The new technique was applied to a wide range of data classification and adaptation problems covering a number of scientific areas and the experimental results were more than promising. Full article

CNNs have become one of the most commonly used computational tools in the past two decades. One of the primary downsides of CNNs is that they work as a “black box”, where the user cannot necessarily know how the image data are analyzed, and therefore needs to rely on empirical evaluation to test the efficacy of a trained CNN. This can lead to hidden biases that affect the performance evaluation of neural networks, but are difficult to identify. Here we discuss examples of such hidden biases in common and widely used benchmark datasets, and propose techniques for identifying dataset biases that can affect the standard performance evaluation metrics. One effective approach to identify dataset bias is to perform image classification by using merely blank background parts of the original images. However, in some situations, a blank background in the images is not available, making it more difficult to separate foreground or contextual information from the bias. To overcome this, we propose a method to identify dataset bias without the need to crop background information from the images. The method is based on applying several image transforms to the original images, including Fourier transform, wavelet transforms, median filter, and their combinations. These transforms are applied to recover background bias information that CNNs use to classify images. These transformations affect the contextual visual information in a different manner than it affects the systemic background bias. Therefore, the method can distinguish between contextual information and the bias, and can reveal the presence of background bias even without the need to separate sub-image parts from the blank background of the original images. The code used in the experiments is publicly available. Full article

Increased electrical activity in the brain causes epilepsy, which causes seizures, resulting in various medical complications that can sometimes be fatal. Doctors use electroencephalography (EEG) for the profiling and diagnosis of epilepsy. According to the World Health Organization (WHO), approximately 50 million people worldwide have epilepsy, making it one of the most common neurological disorders globally. This number represents about 0.7% of the global population. The conventional method of EEG analysis employed by medical professionals is a visual investigation that is time-consuming and requires expertise because of the variability in EEG signals. This paper describes a method for detecting epileptic seizures in EEG signals by combining signal processing and machine learning techniques. SVM and other machine learning techniques detect anomalies in the input EEG signal. To extract features, DWT is used for decomposition to sub-bands. The proposed method aims to improve the accuracy of the machine learning model while using as few features as possible. The classification results show an accuracy of 100% with just one feature, mean absolute value, from datasets A and E. With additional features, the overall accuracy remains high at 99%, with specificity and sensitivity values of 97.2% and 99.1%, respectively. These results outperform previous research on the same dataset, demonstrating the effectiveness of our approach. This research contributes to developing more accurate and efficient epilepsy diagnosis systems, potentially improving patient outcomes. Full article

The Ecuadorian electricity sector, until April 2024, presented losses of 15.64% (6.6% technical and 9.04% non-technical), so it is important to detect the areas that potentially sub-register energy in order to reduce Non-Technical Losses (NTLs). The “Empresa Eléctrica de Ambato Sociedad Anónima” (EEASA), as a distribution company, has, to reduce NTLs, incorporated many smart meters in special clients, generating a large amount of data that are stored. This historical information is analyzed to detect anomalous consumption that is not easily recognized and is a significant part of the NTLs. The use of machine learning with appropriate clustering techniques and deep learning neural networks work together to detect abnormal curves that record lower readings than the real energy consumption. The developed methodology uses three k-means validation indices to classify daily energy curves based on the days of the week and holidays that present similar behaviors in terms of energy consumption. The developed algorithm groups similar consumption patterns as input data sets for learning, testing, and validating the densely connected classification neural network, allowing for the identification of daily curves described by customers. The results obtained from the system detected customers who sub-register energy. It is worth mentioning that this methodology is replicable for distribution companies that store historical consumption data with Advanced Measurement Infrastructure (AMI) systems. Full article

Solar energy plays a crucial role in mitigating greenhouse gas emissions in the context of global climate change. However, its deployment for green electricity generation can significantly influence regional climate and vegetation dynamics. While prior studies have examined the impacts of solar power plants on vegetation, the accuracy of these assessments has often been constrained by the availability of publicly accessible multispectral, high-resolution remotely sensed imagery. Given the abundant solar energy resources and the ecological significance of the Tibetan Plateau, a thorough evaluation of the vegetation effects associated with solar power installations is warranted. In this study, we utilize sub-meter resolution imagery from the GF-2 satellite to reconstruct the fractional vegetation cover (FVC) at the Gonghe solar thermal power plant through image classification, in situ sampling, and sliding window techniques. We then quantify the plant’s impact on FVC by comparing data from the pre-installation and post-installation periods. Our findings indicate that the Gonghe solar thermal power plant is associated with a 0.02 increase in FVC compared to a surrounding control region (p < 0.05), representing a 12.5% increase relative to the pre-installation period. Notably, the enhancement in FVC is more pronounced in the outer ring areas than near the central tower. The observed enhancement in vegetation growth at the Gonghe plant suggests potential ecological and carbon storage benefits resulting from solar power plant establishment on the Tibetan Plateau. These findings underscore the necessity of evaluating the climate and ecological impacts of renewable energy facilities during the planning and design phases to ensure a harmonious balance between clean energy development and local ecological integrity. Full article

In recent years, super-resolution technology has gained widespread attention in the field of remote sensing. Despite advancements, current methods often employ uniform reconstruction techniques across entire remote sensing images, neglecting the inherent variability in spatial frequency distributions, particularly the distinction between high-frequency texture regions and smoother areas, leading to computational inefficiency, which introduces redundant computations and fails to optimize the reconstruction process for regions of higher complexity. To address these issues, we propose the Perception-guided Classification Feature Intensification (PCFI) network. PCFI integrates two key components: a compressed sensing classifier that optimizes speed and performance, and a deep texture interaction fusion module that enhances content interaction and detail extraction. This network mitigates the tendency of Transformers to favor global information over local details, achieving improved image information integration through residual connections across windows. Furthermore, a classifier is employed to segment sub-image blocks prior to super-resolution, enabling efficient large-scale processing. The experimental results on the AID dataset indicate that PCFI achieves state-of-the-art performance, with a PSNR of 30.87 dB and an SSIM of 0.8131, while also delivering a 4.33% improvement in processing speed compared to the second-best method. Full article

This investigation investigates the influence of a variety of data augmentation techniques on sentiment analysis in low-resource languages, with a particular emphasis on Bulgarian, Croatian, Slovak, and Slovene. The following primary research topic is addressed: is it possible to improve sentiment analysis efficacy in low-resource languages through data augmentation? Our sub-questions look at how different augmentation methods affect performance, how effective WordNet-based augmentation is compared to other methods, and whether lemma-based augmentation techniques can be used, especially for Croatian sentiment tasks. The sentiment-labelled evaluations in the selected languages are included in our data sources, which were curated with additional annotations to standardise labels and mitigate ambiguities. Our findings show that techniques like replacing words with synonyms, masked language model (MLM)-based generation, and permuting and combining sentences can only make training datasets slightly bigger. However, they provide limited improvements in model accuracy for low-resource language sentiment classification. WordNet-based techniques, in particular, exhibit a marginally superior performance compared to other methods; however, they fail to substantially improve classification scores. From a practical perspective, this study emphasises that conventional augmentation techniques may require refinement to address the complex linguistic features that are inherent to low-resource languages, particularly in mixed-sentiment and context-rich instances. Theoretically, our results indicate that future research should concentrate on the development of augmentation strategies that introduce novel syntactic structures rather than solely relying on lexical variations, as current models may not effectively leverage synonymic or lemmatised data. These insights emphasise the nuanced requirements for meaningful data augmentation in low-resource linguistic settings and contribute to the advancement of sentiment analysis approaches. Full article

Tropical cyclones (TCs) are associated with severe weather phenomena, making accurate wind field retrieval crucial for TC monitoring. SAR’s high-resolution imaging capability provides detailed information for TC observation, and wind speed calculations require wind direction as prior information. Therefore, utilizing SAR images to retrieve TC wind fields is of significant importance. This study introduces a novel approach for retrieving wind direction from SAR images of TCs through the classification of TC sub-images. The method utilizes a transfer learning-based Inception V3 model to identify wind streaks (WSs) and rain bands in SAR images under TC conditions. For sub-images containing WSs, the Mexican-hat wavelet transform is applied, while for sub-images containing rain bands, an edge detection technique is used to locate the center of the TC eye and subsequently the tangent to the spiral rain bands is employed to determine the wind direction associated with the rain bands. Wind direction retrieval from 10 SAR TC images showed an RMSD of 19.52° and a correlation coefficient of 0.96 when compared with ECMWF and HRD observation wind directions, demonstrating satisfactory consistency and providing highly accurate TC wind directions. These results confirm the method’s potential applications in TC wind direction retrieval. Full article

Tumors in the brain can be life-threatening, making early and precise detection crucial for effective treatment and improved patient outcomes. Deep learning (DL) techniques have shown significant potential in automating the early diagnosis of brain tumors by analyzing magnetic resonance imaging (MRI), offering a more efficient and accurate approach to classification. Deep convolutional neural networks (DCNNs), which are a sub-field of DL, have the potential to analyze rapidly and accurately MRI data and, as such, assist human radiologists, facilitating quicker diagnoses and earlier treatment initiation. This study presents an ensemble of three high-performing DCNN models, i.e., DenseNet169, EfficientNetB0, and ResNet50, for accurate classification of brain tumors and non-tumor MRI samples. Our proposed ensemble model demonstrates significant improvements over various evaluation parameters compared to individual state-of-the-art (SOTA) DCNN models. We implemented ten SOTA DCNN models, i.e., EfficientNetB0, ResNet50, DenseNet169, DenseNet121, SqueezeNet, ResNet34, ResNet18, VGG16, VGG19, and LeNet5, and provided a detailed performance comparison. We evaluated these models using two learning rates (LRs) of 0.001 and 0.0001 and two batch sizes (BSs) of 64 and 128 and identified the optimal hyperparameters for each model. Our findings indicate that the ensemble approach outperforms individual models, having 92% accuracy, 90% precision, 92% recall, and an F1 score of 91% at a 64 BS and 0.0001 LR. This study not only highlights the superior performance of the ensemble technique but also offers a comprehensive comparison with the latest research. Full article

A comprehensive review of the application of different ceramics for MEMS devices is presented. Main ceramics materials used for MEMS systems and devices including alumina, zirconia, aluminum Nitride, Silicon Nitride, and LTCC are introduced. Conventional and new methods of fabricating each material are explained based on the literature, along with the advantages of the new approaches, mainly additive manufacturing, i.e., 3D-printing technologies. Various manufacturing processes with relevant sub-techniques are detailed and the ones that are more suitable to have an application for MEMS devices are highlighted with their properties. In the main body of this paper, each material with its application for MEMS is categorized and explained. The majority of works are within three main classifications, including the following: (i) using ceramics as a substrate for MEMS devices to be mounted or fabricated on top of it; (ii) ceramics are a part of the materials used for an MEMS device or a monolithic fabrication of MEMS and ceramics; and finally, (iii) using ceramics as packaging solution for MEMS devices. We elaborate on how ceramics may be superior substitutes over other materials when delicate MEMS-based systems need to be assembled or packaged by a simpler fabrication process as well as their advantages when they need to operate in harsh environments. Full article

Situational awareness detection and characterization of mental states have a vital role in medicine and many other fields. An electroencephalogram (EEG) is one of the most effective tools for identifying and analyzing cognitive stress. Yet, the measurement, interpretation, and classification of EEG sensors is a challenging task. This study introduces a novel machine learning-based approach to assist in evaluating situational awareness detection using EEG signals and spiking neural networks (SNNs) based on a unique spike continuous-time neuron (SCTN). The implemented biologically inspired SNN architecture is used for effective EEG feature extraction by applying time–frequency analysis techniques and allows adept detection and analysis of the various frequency components embedded in the different EEG sub-bands. The EEG signal undergoes encoding into spikes and is then fed into an SNN model which is well suited to the serial sequence order of the EEG data. We utilize the SCTN-based resonator for EEG feature extraction in the frequency domain which demonstrates high correlation with the classical FFT features. A new SCTN-based 2D neural network is introduced for efficient EEG feature mapping, aiming to achieve a spatial representation of each EEG sub-band. To validate and evaluate the performance of the proposed approach, a common, publicly available EEG dataset is used. The experimental results show that by using the extracted EEG frequencies features and the SCTN-based SNN classifier, the mental state can be accurately classified with an average accuracy of 96.8% for the common EEG dataset. Our proposed method outperforms existing machine learning-based methods and demonstrates the advantages of using SNNs for situational awareness detection and mental state classifications. Full article