Search

Due to the large intra-class differences between the same categories and the scale imbalance between different categories in the remote sensing image dataset, the semantic segmentation task presents the problem of small-scale object information loss, the imbalance between foreground and background, and simultaneously the background dominates, which seriously affects the performance of the network model. To solve the above problems, this paper proposes an efficient bilateral branch depth neural network model based on the U-Net depth neural network, named BBU-Net. Firstly, one branch of the network learns the distribution characteristics of the original data, and the other focuses on difficult samples. Then the two branches improve the representation and classification ability of the neural network by accumulating learning strategies. Finally, considering the geometric diversity of remote sensing images, this paper adopts test time augmentation and reflection padding strategies and proposes a balanced weighted loss function named CombineLoss to alleviate the imbalance in the training process. The depth neural network proposed in this paper was first tested on the Inria Aerial Image Labeling Dataset, and 87.53% of mean intersection over union and 97.4% of mean pixel accuracy were obtained, respectively. At the same time, to verify the model's complexity, the model proposed in this paper is compared with the neural network based on integrated learning. The comparison results show that the spatial complexity of the network proposed in this paper is much lower than the neural network obtained by integrated learning, and the parameters are also much smaller than the neural network based on integrated learning. Then use the satellite building dataset I in the WHU Building Dataset and mainstream semantic segmentation methods for multiple groups of comparative experiments. The experimental results show that the method proposed in this paper can effectively extract the semantic information of remote sensing images, significantly improve the imbalance of remote sensing image data, improve the performance of the network model, and achieve a good semantic segmentation effect, which fully proves the effectiveness of this method. [ABSTRACT FROM AUTHOR]

Power line inspection is an important means to eliminate hidden dangers of power lines. It is a difficult research problem how to solve the low accuracy of power line inspection based on deep neural network (DNN) due to the problems of multi‐view‐shape, small‐size object. In this paper, an automatic detection model based on Feature visual clustering network (FVCNet) for power line inspection is established. First, an unsupervised clustering method for power line inspection is proposed, and applied to construct a detection model which can recognize multi‐view‐shape objects and enhanced object features. Then, the bilinear interpolation method is used to Feature enhancement method, and the enhanced high‐level semantics and low‐level semantics are fused to solve the problems of small object size and single sample. In this paper, FVCNet is applied to the MS‐COCO 2017 data set and self‐made power line inspection data set, and the test accuracy is increased to 61.2% and 82.0%, respectively. Compared with other models, especially for the categories that are greatly affected by multi‐view‐shape, the test accuracy has been improved significantly. [ABSTRACT FROM AUTHOR]

One key frontier of artificial intelligence (AI) is the ability to comprehend research articles and validate their findings, posing a magnanimous problem for AI systems to compete with human intelligence and intuition. As a benchmark of research validation, the existing peer-review system still stands strong despite being criticized at times by many. However, the paper vetting system has been severely strained due to an influx of research paper submissions and increased conferences/journals. As a result, problems, including having insufficient reviewers, finding the right experts, and maintaining review quality, are steadily and strongly surfacing. To ease the workload of the stakeholders associated with the peer-review process, we probed into what an AI-powered review system would look like. In this work, we leverage the interaction between the paper's full text and the corresponding peer-review text to predict the overall recommendation score and final decision. We do not envisage AI reviewing papers in the near future. Still, we intend to explore the possibility of a human–AI collaboration in the decision-making process to make the current system FAIR. The idea is to have an assistive decision-making tool for the chairs/editors to help them with an additional layer of confidence, especially with borderline and contrastive reviews. We use a deep attention network between the review text and paper to learn the interactions and predict the overall recommendation score and final decision. We also use sentiment information encoded within peer-review texts to guide the outcome further. Our proposed model outperforms the recent state-of-the-art competitive baselines. We release the code of our implementation here: https://github.com/PrabhatkrBharti/PEERRec.git. [ABSTRACT FROM AUTHOR]

The number of papers on machine learning and deep neural networks applied to intrusion detection systems (IDS) is ever-increasing. Differently from existing work on the topic, this paper explores the effect of training-time randomness of deep neural networks, which is overlooked by the related literature. Training-time randomness is regulated by the seed of the pseudorandom number generator, and affects the performance of IDS models. The seed selection is studied in conjunction with other critical learning parameters: to the best of our knowledge, there are no similar studies in IDS. The experiments are done with a recent and widely consolidated intrusion detection benchmark, which is used to train and test a neural network under different combinations of seeds and parameters both in supervised and semi-supervised learning modes. The results are inferred by a mixture of explorative analysis, design of experiments, and analysis of variance. According to the results, the choice of the seed yields either excellent or scarce detection metrics; more importantly, the seed selection might be as relevant as the other major learning parameters assessed. [ABSTRACT FROM AUTHOR]

This paper trains a multilayered neural network (MLNN) for solving fractional delay differential equations (FDDEs), including nonlinear and singular types. The proposed methodology involves replacing the unknown functions in the equations with a truncated power series expansion. Subsequently, a collection of algebraic equations is solved utilizing an iterative minimization technique that leverages the capabilities of the MLNN architecture. The outcomes demonstrate that the MLNN architecture provides the required accuracy and strong stability compared to several numerical methods. [ABSTRACT FROM AUTHOR]

Accurate three-dimensional (3D) measurement for large field of view (FOV) is currently a significant research field. Accordingly, system calibration is crucial to ensure accuracy. However typical calibration methods often involve the use of large calibration objects, which is not only expensive but also difficult to achieve sufficient accuracy. A novel method based on a dual-brand deep neural network (DNN) is proposed for the system calibration. Taking advantage of the concept of "divide and conquer", the FOV is divided into sub-regions with a part of overlapping regions by a small calibration object, which forms a large calibration object covering the whole FOV. Then the sub-regions are fused into a global framework and further optimized by the proposed dual-brand DNN. The proposed method reduces the need for calibration objects while improving the calibration accuracy and generalization ability in large FOV. A series of experiments have been designed to prove the effectiveness and robustness of the proposed method. • A high-accuracy calibration method of a 3D measurement system with large field-of-view is proposed. The highlights of the paper are: • A high-accuracy calibration method for large field-of-view is proposed. • The large field-of-view is divided into multiple sub-regions. • Sub-regions are unified to a uniform coordinate system by self-constraint. • A dual-branch DNN is optimized using the multi-objective loss function. [ABSTRACT FROM AUTHOR]

The driver authority decision for driver‐initiated takeover is closely related to vehicle driving safety. In this paper, a lateral driver‐automation driver authority decision method considering the safety of the intended functionality is proposed to improve vehicle driving safety. Based on systems‐theoretic processes analysis, functional safety analysis of driver initiative takeover is carried out to clarify the safety of the intended functionality issues caused by the unreasonable decision. Then, a method for defining the safe driving area is developed to assess the safety level of the driving area around the ego vehicle. Taking the driver‐vehicle state parameters and vehicle–vehicle state parameters as input, a deep neural network model is constructed to determine the intention of the driver to take over the vehicle. The lateral driver‐automation driver authority decision method is designed to make the driver‐machine take‐over decisions, which considers the safety of the intended functionality. The effectiveness of the proposed method is evaluated via numerical simulation and hardware‐in‐the‐loop experiments. The results show that the designed method not only improves the control ability of the driver to the maximum extent but also integrates the perception ability of the driver into the vehicle control system to improve further vehicle driving safety. [ABSTRACT FROM AUTHOR]

Recurrent neural networks (RNNs) are widely utilized in neural network research to capture spatiotemporal features in video data. However, their effectiveness heavily relies on the spatial features upon which they trained. This paper introduces innovative ensembles of features for constructing frame-wise structures by employing impactful neural network models with innovative training pipelines. These features are designed to enhance the recognition of hand gesture videos using RNN by leveraging temporal information. Recognizing hand gestures from videos is a complex task that presents considerable challenges. One notable challenge is the overlap in gesture motion, where different gesture categories exhibit similar hand poses within a single video clip. To overcome this issue, we were motivated to develop extensive and diverse features that offer a more comprehensive description of the gesture video clips, thereby mitigating recognition problems caused by images overlapping. Overall, our efforts to generate diverse features have yielded promising results in enhancing the recognition of hand gestures from videos, particularly in scenarios where overlap poses a significant challenge. We have combined the extracted features from a deep neural network trained from scratch with features obtained from various standard neural networks (Self-Organizing Map, Radial Base Function) that are trained to enhance the deep-trained features. The mutual arrangement for combining the shared features has configured new frame-wise image features. Furthermore, we have provided a performance comparison of the newly constructed frame-wise features through time-sharing to train RNN for recognition. The proposed models have been evaluated on two-hand gesture video datasets, where a preserving gesture sequence is crucial due to overlapping motions. Our work demonstrates a significant improvement in performance for both datasets. [ABSTRACT FROM AUTHOR]

Chemical warfare agents pose a serious threat due to their extreme toxicity, necessitating swift the identification of chemical gases and individual responses to the identified threats. Fourier transform infrared (FTIR) spectroscopy offers a method for remote material analysis, particularly in detecting colorless and odorless chemical agents. In this paper, we propose a deep neural network utilizing a semi-supervised autoencoder (SSAE) for the classification of chemical gases based on FTIR spectra. In contrast to traditional methods, the SSAE concurrently trains an autoencoder and a classifier attached to a latent vector of the autoencoder, enhancing feature extraction for classification. The SSAE was evaluated on laboratory-collected FTIR spectra, demonstrating a superior classification performance compared to existing methods. The efficacy of the SSAE lies in its ability to generate denser cluster distributions in latent vectors, thereby enhancing gas classification. This study established a consistent experimental environment for hyperparameter optimization, offering valuable insights into the influence of latent vectors on classification performance. [ABSTRACT FROM AUTHOR]

The paper introduces a computer vision methodology for detecting pitting corrosion in gas pipelines. To achieve this, a dataset comprising 576,000 images of pipelines with and without pitting corrosion was curated. A custom-designed and optimized convolutional neural network (CNN) was employed for binary classification, distinguishing between corroded and non-corroded images. This CNN architecture, despite having relatively few parameters compared to existing CNN classifiers, achieved a notably high classification accuracy of 98.44%. The proposed CNN outperformed many contemporary classifiers in its efficacy. By leveraging deep learning, this approach effectively eliminates the need for manual inspection of pipelines for pitting corrosion, thus streamlining what was previously a time-consuming and cost-ineffective process. [ABSTRACT FROM AUTHOR]

Establishing a reliable credit card fraud detection model has become a primary focus for academia and the financial industry. The existing anti-fraud methods face challenges related to low recall rates, inaccurate results, and insufficient causal modeling ability. This paper proposes a credit card fraud detection model based on counterfactual data enhancement of the triplet network. Firstly, we convert the problem of generating optimal counterfactual explanations (CFs) into a policy optimization of agents in the discrete–continuous mixed action space, thereby ensuring the stable generation of optimal CFs. The triplet network then utilizes the feature similarity and label difference of positive example samples and CFs to enhance the learning of the causal relationship between features and labels. Experimental results demonstrate that the proposed method improves the accuracy and robustness of the credit card fraud detection model, outperforming existing methods. The research outcomes are of significant value for both credit card anti-fraud research and practice while providing a novel approach to causal modeling issues across other fields. [ABSTRACT FROM AUTHOR]

The price of carbon emission rights in the market fluctuates greatly due to various factors from economy, finance, and climate. For an enterprise that needs to conduct carbon trading activities, it is extremely necessary to fully grasp the price of carbon trading in the future period. Thus in this paper, a deep neural network-based assistive decision model for financial risk prediction in carbon trading market is proposed for this purpose. Specifically, the dynamic risk spillover effects of domestic and international carbon trading markets are studied, and a frontier time-varying model is utilized to measure the risk spillover effects. Then, the deep neural network is used for quantitative research and to construct an intelligent decision scheme that outputs financial risk prediction results. Four perspectives, energy price, climate environment, carbon market price, and macroeconomics, are selected as the input to analyze the influence factors of carbon emission rights price. Finally, several linear regression models are adopted as the baseline methods for comparison, and experimental results show that the proposed method can achieve better prediction performance compared with baseline methods. [ABSTRACT FROM AUTHOR]

The health management of railway signal equipment in the high-speed railway is a key link between intelligent operation and maintenance. Accurately predicting the health state of compensation capacitors is of great significance to ensure the reliable work of track circuits. This paper proposes an improved deep neural network algorithm focusing on the problem of long-term accurate health forecasts for compensation capacitors. First, establishing a transmission state model for degradation mechanism mining, the difference function that can quantitatively evaluate features is defined by piecewise processing cab signaling receiving voltage. Introducing the degradation model, predictive driving under both model and data is implemented. Then, the convolutional neural networks and bidirectional long–short-term memory are combined and improved to construct a novel artificial intelligence combination strategy, while parameters are optimized based on the sparrow search algorithm. Finally, facing the conditional repair of compensation capacitors, we set a reasonable threshold for the occurrence of hidden dangers to complete fault warning. This novel and practical approach effectively explores the procedure of prognosis and health management, while the refined maintenance will better utilize current monitoring information, helping the intelligence and accuracy of safety control decision-making. [ABSTRACT FROM AUTHOR]

Existing research in metasurface design was based on trial-and-error high-intensity iterations and requires deep acoustic expertise from the researcher, which severely hampered the development of the metasurface field. Using deep learning enabled the fast and accurate design of hypersurfaces. Based on this, in this paper, an integrated learning approach was first utilized to construct a model of the forward mapping relationship between the hypersurface physical structure parameters and the acoustic field, which was intended to be used for data enhancement. Then a dual-feature fusion model (DFCNN) based on a convolutional neural network was proposed, in which the first feature was the high-dimensional nonlinear features extracted using a data-driven approach, and the second feature was the physical feature information of the acoustic field mined using the model. A convolutional neural network was used for feature fusion. A genetic algorithm was used for network parameter optimization. Finally, generalization ability verification was performed to prove the validity of the network model. The results showed that 90% of the integrated learning models had an error of less than 3 dB between the real and predicted sound field data, and 93% of the DFCNN models could achieve an error of less than 5 dB in the local sound field intensity. [ABSTRACT FROM AUTHOR]

In this paper, we propose a method for efficient target classification based on the spatial features of the point cloud generated by using a high-resolution radar sensor. The frequency-modulated continuous wave radar sensor can estimate the distance and velocity of a target. In addition, the azimuth and elevation angle of the target can be estimated by using a multiple-input and multiple-output antenna system. Using the estimated distance, velocity, and angle, the 3D point cloud of target can be generated. From the generated point cloud, we extract the point cloud for each individual target using the density-based spatial clustering of application with noise method and a camera mounted on the radar sensor. Then, we define the convex hull boundaries that enclose these point clouds in both 3D and 2D spaces obtained by orthogonally projecting onto the x y , y z , and z x planes. Using the vertices of convex hull, we calculate the volume of the targets and the areas in 2D spaces. Several feature points, including the calculated spatial information, are numerized and configured into feature vectors. We design an uncomplicated deep neural network classifier based on minimal input information to achieve fast and efficient classification performance. As a result, the proposed method achieved an average accuracy of 97.1%, and the time required for training was reduced compared to the method using only point cloud data and the convolutional neural network-based method. [ABSTRACT FROM AUTHOR]

People with color vision deficiency (CVD) have difficulty in distinguishing differences between colors. To compensate for the loss of color contrast experienced by CVD individuals, a lot of image recoloring approaches have been proposed. However, the state-of-the-art methods suffer from the failures of simultaneously enhancing color contrast and preserving naturalness of colors [without reducing the Quality of Vision (QOV)], high computational cost, etc. In this paper, we propose an image recoloring method using deep neural network, whose loss function takes into consideration the naturalness and contrast, and the network is trained in an unsupervised manner. Moreover, Swin transformer layer, which has long-range dependency mechanism, is adopted in the proposed method. At the same time, a dataset, which contains confusing color pairs to CVD individuals, is newly collected in this study. To evaluate the performance of the proposed method, quantitative and subjective experiments have been conducted. The experimental results showed that the proposed method is competitive to the state-of-the-art methods in contrast enhancement and naturalness preservation and has a real-time advantage. The code and model will be made available at https://github.com/Ligeng-c/CVD_swin. [ABSTRACT FROM AUTHOR]

The 3D point cloud deep neural network (3D DNN) has achieved remarkable success, but its black-box nature hinders its application in many safety-critical domains. The saliency map technique is a key method to look inside the black-box and determine where a 3D DNN focuses when recognizing a point cloud. Existing point-wise point cloud saliency methods are proposed to illustrate the point-wise saliency for a given 3D DNN. However, the above critical points are alternative and unreliable. The findings are grounded on our experimental results which show that a point becomes critical because it is responsible for representing one specific local structure. However, one local structure does not have to be represented by some specific points, conversely. As a result, discussing the saliency of the local structure (named patch-wise saliency) represented by critical points is more meaningful than discussing the saliency of some specific points. Based on the above motivations, this paper designs a black-box algorithm to generate patch-wise saliency map for point clouds. Our basic idea is to design the Mask Building-Dropping process, which adaptively matches the size of important/unimportant patches by clustering points with close saliency. Experimental results on several typical 3D DNNs show that our patch-wise saliency algorithm can provide better visual guidance, and can detect where a 3D DNN is focusing more efficiently than a point-wise saliency map. Finally, we apply our patch-wise saliency map to adversarial attacks and backdoor defenses. The results show that the improvement is significant. [ABSTRACT FROM AUTHOR]

Many test coverage metrics have been proposed to measure the deep neural network (DNN) testing effectiveness, including structural coverage and nonstructural coverage. These test coverage metrics are proposed based on the fundamental assumption: They are correlated with test effectiveness. However, the fundamental assumption is still not validated sufficiently and reasonably, which brings question on the usefulness of DNN test coverage. This paper conducted a revisiting study on the existing DNN test coverage from the test effectiveness perspective, to effectively validate the fundamental assumption. Here, we carefully considered the diversity of subjects, three test effectiveness criteria, and both typical and state‐of‐the‐art test coverage metrics. Different from all the existing studies that deliver negative conclusions on the usefulness of existing DNN test coverage, we identified some positive conclusions on their usefulness from the test effectiveness perspective. In particular, we found the complementary relationship between structural and nonstructural coverage and identified the practical usage scenarios and promising research directions for these existing test coverage metrics. [ABSTRACT FROM AUTHOR]

One of the predominant health issues affecting Saudi Arabia and leading to many complications is Type 2 diabetes (T2D). Early detection and significant preventative measures lead to curbing and controlling the health issue. There are fewer datasets in the literature for the detection of T2D in the Saudi population. Past studies using Saudi data have favoured machine learning algorithms to classify T2D. Although the application of this data in machine learning is evident, no studies exist in the literature that compare this data, especially those related to deep learning algorithms. This study's objective is to use specific Saudi data to develop multiple deep learning models that could be used to detect T2D. The research uses a Deep Neural Network (DNN), an Autoencoder (AE), and a Convolutional Neural Network (CNN) to create predictive models and compare their performance with a traditional machine learning classifier used on the same dataset that outperformed other machine learning algorithms such as a Decision Forest (DF). Various metrics were used to evaluate the effectiveness of the models, such as accuracy, precision, recall, F1 score and area under the ROC curve (AUC) where the ROC acts as a receiver operating characteristic curve. There are two cases in this paper: (i) uses all features of the dataset and (ii) uses six of the ten features, such as DF. In case (i), the results were shown that AE outperformed other models with the highest accuracy for imbalanced and balanced data 81.12% and 79.16%, respectively. The results for case (ii) showed that AE scored the highest 81.01% accuracy with imbalanced data compared to DF and DF achieved the highest accuracy of 82.1% with balanced data. As a result, both cases explored in this study revealed that AE has a constant superior performance if imbalanced data is used. In contrast, DF demonstrated the highest accuracy when a balanced dataset was used with a feature set reduction. They help to identify the undiagnosed T2D, and they are essential for professionals in Saudi Arabia in the health sector to promote health connections, identify risks and contain or improve their diabetes management. [ABSTRACT FROM AUTHOR]

During the operation of HVAC (Heating, Ventilation, and Air-Conditioning) systems, precise energy consumption prediction plays an important role in achieving energy savings and optimizing system performance. However, the HVAC system is a complex and dynamic system characterized by a large number of variables that exhibit significant changes over time. Therefore, it is inadequate to rely on a fixed offline model to adapt to the dynamic changes in the system that consume tremendous computation time. To solve this problem, a deep neural network (DNN) model based on Just-in-Time learning with hyperparameter R (RJITL) is proposed in this paper to predict HVAC energy consumption. Firstly, relevant samples are selected using Euclidean distance weighted by Spearman coefficients. Subsequently, local models are constructed using deep neural networks supplemented with optimization techniques to enable real-time rolling energy consumption prediction. Then, the ensemble JITL model mitigates the influence of local features, and improves prediction accuracy. Finally, the local models can be adaptively updated to reduce the training time of the overall model by defining the update rule (hyperparameter R) for the JITL model. Experimental results on energy consumption prediction for the HVAC system show that the proposed DNN-RJITL method achieves an average improvement of 5.17% in accuracy and 41.72% in speed compared to traditional methods. [ABSTRACT FROM AUTHOR]

Adversarial examples are often used to test and evaluate the security and robustness of image classification models. Though adversarial attacks under white-box setting can achieve a high attack success rate, due to overfitting, the success rate of black-box attacks is relatively low. To this end, this paper proposes diversified input strategies to improve the transferability of adversarial examples. In this method, various transformation methods are applied to randomly transform the original image multiple times, thereby generating a batch of transformed images. Then, in the process of back-propagation, the loss function gradient of the transformed images is calculated, and a weighted average of the obtained gradient values is performed to generate adversarial perturbation, which is iteratively added to the original image to generate adversarial examples. Meanwhile, by increasing the variety of data augmentation transformation types and the number of input images, the proposed method effectively alleviates overfitting and improves the transferability of adversarial examples. Extensive experiments on the ImageNet dataset indicate that the proposed method can perform black-box attacks better than benchmark methods, with an average of 97.2% success rate attacking multiple models simultaneously. [ABSTRACT FROM AUTHOR]

In recent years, deep learning has revolutionized natural language processing (NLP) by enabling the development of models that can learn complex representations of language data, leading to significant improvements in performance across a wide range of NLP tasks. Deep learning models for NLP typically use large amounts of data to train deep neural networks, allowing them to learn the patterns and relationships in language data. This is in contrast to traditional NLP approaches, which rely on hand-engineered features and rules to perform NLP tasks. The ability of deep neural networks to learn hierarchical representations of language data, handle variable-length input sequences, and perform well on large datasets makes them well-suited for NLP applications. Driven by the exponential growth of textual data and the increasing demand for condensed, coherent, and informative summaries, text summarization has been a critical research area in the field of NLP. Applying deep learning to text summarization refers to the use of deep neural networks to perform text summarization tasks. In this survey, we begin with a review of fashionable text summarization tasks in recent years, including extractive, abstractive, multi-document, and so on. Next, we discuss most deep learning-based models and their experimental results on these tasks. The paper also covers datasets and data representation for summarization tasks. Finally, we delve into the opportunities and challenges associated with summarization tasks and their corresponding methodologies, aiming to inspire future research efforts to advance the field further. A goal of our survey is to explain how these methods differ in their requirements as understanding them is essential for choosing a technique suited for a specific setting. This survey aims to provide a comprehensive review of existing techniques, evaluation methodologies, and practical applications of automatic text summarization. [ABSTRACT FROM AUTHOR]

Deep Neural Networks (DNNs) have been successfully applied to investigations of numerical dynamics of finite-dimensional nonlinear systems such as ODEs instead of finding numerical solutions to ODEs via the traditional Runge–Kutta method and its variants. To show the advantages of DNNs, in this paper, we demonstrate that the DNNs are more efficient in finding topological horseshoes in chaotic dynamical systems. [ABSTRACT FROM AUTHOR]

To improve the recognition ability of clinical named entity recognition (CNER) in a limited number of Chinese electronic medical records, it provides meaningful support for clinical advanced knowledge extraction. In this paper, using CCKS2019 Chinese electronic medical record as an experimental data source, a fusion model enhanced by knowledge graph (KG) is proposed, and the model is applied to specific Chinese CNER tasks. This study consists of three main parts: single-mode model construction and comparison experiment, KG enhancement experiment, and model fusion experiment. The model has achieved good performance in CNER from the results. The accuracy rate, recall rate, and F1 value are 83.825%, 84.705%, and 84.263%, respectively, which is the global optimal, which proves the effectiveness of the model. This provides a good help for further research of medical information. [ABSTRACT FROM AUTHOR]

Background: Classifying breast cancer subtypes is crucial for clinical diagnosis and treatment. However, the early symptoms of breast cancer may not be apparent. Rapid advances in high-throughput sequencing technology have led to generating large number of multi-omics biological data. Leveraging and integrating the available multi-omics data can effectively enhance the accuracy of identifying breast cancer subtypes. However, few efforts focus on identifying the associations of different omics data to predict the breast cancer subtypes. Results: In this paper, we propose a differential sparse canonical correlation analysis network (DSCCN) for classifying the breast cancer subtypes. DSCCN performs differential analysis on multi-omics expression data to identify differentially expressed (DE) genes and adopts sparse canonical correlation analysis (SCCA) to mine highly correlated features between multi-omics DE-genes. Meanwhile, DSCCN uses multi-task deep learning neural network separately to train the correlated DE-genes to predict breast cancer subtypes, which spontaneously tackle the data heterogeneity problem in integrating multi-omics data. Conclusions: The experimental results show that by mining the associations among multi-omics data, DSCCN is more capable of accurately classifying breast cancer subtypes than the existing methods. [ABSTRACT FROM AUTHOR]

The cost function gradient vanishing or exploding problem and slow convergence speed are key issues when training deep neural networks (DNNs). In this paper, we investigate the forward and backward propagation processes of DNN training and explore the properties of the activation function and derivative function (ADF) employed. The outputs' distribution of ADF with near-zero mean is proposed to reduce gradient problems. Additionally, the constant energy transfer of propagating data in the training process is also proposed to speed up convergence further. Based on wavelet frame theory, we derive a novel ADF, i.e., tight frame wavelet activation function (TFWAF) and tight frame wavelet derivative function (TFWDF) of the Mexican hat wavelet, to stabilize and accelerate DNN training. The nonlinearity of wavelet functions can strengthen the learning capacity of DNN models, while the sparse property of wavelets derived can reduce the overfitting problem and enhance the robustness of models. Experiments demonstrate that the proposed method stabilizes the DNN training process and accelerates convergence. [ABSTRACT FROM AUTHOR]

In synthetic aperture radar (SAR) imaging, intelligent object detection methods are facing significant challenges in terms of model robustness and application security, which are posed by adversarial examples. The existing adversarial example generation methods for SAR object detection can be divided into two main types: global perturbation attacks and local perturbation attacks. Due to the dynamic changes and irregular spatial distribution of SAR coherent speckle backgrounds, the attack effectiveness of global perturbation attacks is significantly reduced by coherent speckle. In contrast, by focusing on the image objects, local perturbation attacks achieve targeted and effective advantages over global perturbations by minimizing interference from the SAR coherent speckle background. However, the adaptability of conventional local perturbations is limited because they employ a fixed size without considering the diverse sizes and shapes of SAR objects under various conditions. This paper presents a framework for region-adaptive local perturbations (RaLP) specifically designed for SAR object detection tasks. The framework consists of two modules. To address the issue of coherent speckle noise interference in SAR imagery, we develop a local perturbation generator (LPG) module. By filtering the original image, this module reduces the speckle features introduced during perturbation generation. It then superimposes adversarial perturbations in the form of local perturbations on areas of the object with weaker speckles, thereby reducing the mutual interference between coherent speckles and adversarial perturbation. To address the issue of insufficient adaptability in terms of the size variation in local adversarial perturbations, we propose an adaptive perturbation optimizer (APO) module. This optimizer adapts the size of the adversarial perturbations based on the size and shape of the object, effectively solving the problem of adaptive perturbation size and enhancing the universality of the attack. The experimental results show that RaLP reduces the detection accuracy of the YOLOv3 detector by 29.0%, 29.9%, and 32.3% on the SSDD, SAR-Ship, and AIR-SARShip datasets, respectively, and the model-to-model and dataset-to-dataset transferability of RaLP attacks are verified. [ABSTRACT FROM AUTHOR]

While single-cell RNA sequencing (scRNA-seq) allows researchers to analyze gene expression in individual cells, its unique characteristics like over-dispersion, zero-inflation, high gene-gene correlation, and large data volume with many features pose challenges for most existing feature selection methods. In this paper, we present a feature selection method based on neural network (scFSNN) to solve classification problem for the scRNA-seq data. scFSNN is an embedded method that can automatically select features (genes) during model training, control the false discovery rate of selected features and adaptively determine the number of features to be eliminated. Extensive simulation and real data studies demonstrate its excellent feature selection ability and predictive performance. [ABSTRACT FROM AUTHOR]

While single-cell RNA sequencing (scRNA-seq) allows researchers to analyze gene expression in individual cells, its unique characteristics like over-dispersion, zero-inflation, high gene-gene correlation, and large data volume with many features pose challenges for most existing feature selection methods. In this paper, we present a feature selection method based on neural network (scFSNN) to solve classification problem for the scRNA-seq data. scFSNN is an embedded method that can automatically select features (genes) during model training, control the false discovery rate of selected features and adaptively determine the number of features to be eliminated. Extensive simulation and real data studies demonstrate its excellent feature selection ability and predictive performance. [ABSTRACT FROM AUTHOR]

Due to the complexity of real optical flow capture, the existing research still has not performed real optical flow capture of infrared (IR) images with the production of an optical flow based on IR images, which makes the research and application of deep learning-based optical flow computation limited to the field of RGB images only. Therefore, in this paper, we propose a method to produce an optical flow dataset of IR images. We utilize the RGB-IR cross-modal image transformation network to rationally transform existing RGB image optical flow datasets. The RGB-IR cross-modal image transformation is based on the improved Pix2Pix implementation, and in the experiments, the network is validated and evaluated using the RGB-IR aligned bimodal dataset M3FD. Then, RGB-IR cross-modal transformation is performed on the existing RGB optical flow dataset KITTI, and the optical flow computation network is trained using the IR images generated by the transformation. Finally, the computational results of the optical flow computation network before and after training are analyzed based on the RGB-IR aligned bimodal data. [ABSTRACT FROM AUTHOR]

Accurate short-term load forecasting (STLF) is essential for power grid systems to ensure reliability, security and cost efficiency. Thanks to advanced smart sensor technologies, time-series data related to power load can be captured for STLF. Recent research shows that deep neural networks (DNNs) are capable of achieving accurate STLP since they are effective in predicting nonlinear and complicated time-series data. To perform STLP, existing DNNs use time-varying dynamics of either past load consumption or past power correlated features such as weather, meteorology or date. However, the existing DNN approaches do not use the time-invariant features of users, such as building spaces, ages, isolation material, number of building floors or building purposes, to enhance STLF. In fact, those time-invariant features are correlated to user load consumption. Integrating time-invariant features enhances STLF. In this paper, a fuzzy clustering-based DNN is proposed by using both time-varying and time-invariant features to perform STLF. The fuzzy clustering first groups users with similar time-invariant behaviours. DNN models are then developed using past time-varying features. Since the time-invariant features have already been learned by the fuzzy clustering, the DNN model does not need to learn the time-invariant features; therefore, a simpler DNN model can be generated. In addition, the DNN model only learns the time-varying features of users in the same cluster; a more effective learning can be performed by the DNN and more accurate predictions can be achieved. The performance of the proposed fuzzy clustering-based DNN is evaluated by performing STLF, where both time-varying features and time-invariant features are included. Experimental results show that the proposed fuzzy clustering-based DNN outperforms the commonly used long short-term memory networks and convolution neural networks. [ABSTRACT FROM AUTHOR]

Optical remote sensing videos, as a new source of remote sensing data that has emerged in recent years, have significant potential in remote sensing applications, especially national defense. In this paper, a tracking pipeline named TDNet (tracking while detecting based on a neural network) is proposed for optical remote sensing videos based on a correlation filter and deep neural networks. The pipeline is used to simultaneously track ships and planes in videos. There are many target tracking methods for general video data, but they suffer some difficulties in remote sensing videos with low resolution and those influenced by weather conditions. The tracked targets are usually misty. Therefore, in TDNet, we propose a new multi-target tracking method called MT-KCF and a detecting-assisted tracking (i.e., DAT) module to improve tracking accuracy and precision. Meanwhile, we also design a new target recognition (i.e., NTR) module to recognise newly emerged targets. In order to verify the performance of TDNet, we compare our method with several state-of-the-art tracking methods on optical video remote sensing data sets acquired from the Jilin No. 1 satellite. The experimental results demonstrate the effectiveness and the state-of-the-art performance of the proposed method. The proposed method can achieve more than 90% performance in terms of precision for single-target tracking tasks and more than 85% performance in terms of MOTA for multi-object tracking tasks. [ABSTRACT FROM AUTHOR]

Every contact centre engages in some form of Call Quality Monitoring in order to improve agent performance and customer satisfaction. Call centres have traditionally used a manual process to sort, select, and analyse a representative sample of interactions for evaluation purposes. Unfortunately, such a process is marked by subjectivity, which in turn results in a distorted picture of agent performance. To address the challenge of identifying and removing subjectivity, empirical research is required. In this paper, we introduce an evidence-based, machine learning-driven framework for the automatic detection of subjective calls. We analyse a corpus of seven hours of recorded calls from a real-estate call centre using Deep Neural Network (DNN) for a multi-classification problem. The study establishes the first baseline for subjectivity detection, with an accuracy of 75%, which is comparable to relevant speech studies in emotional recognition and performance classification. We conclude, among other things, that in order to achieve the best performance evaluation, subjective calls should be removed from the evaluation process or subjective scores deducted from the overall results. [ABSTRACT FROM AUTHOR]

Recognition of handwritten digits has garnered significant interest among researchers in the domain of recognizing pattern. This interest stems from the recognition's relevance in various real-life applications, including reading financial checks and official documents, which has remained a persistent obstacle. To address this challenge, researchers have developed numerous algorithms focusing on recognizing handwritten digits across different human languages. This paper presents a new Kurdish Handwritten dataset, consisting of Kurdish characters, digits, texts, and symbols. The dataset consists of 1560 participants, encompassing a broad and varied group. It serves as the primary dataset for training and evaluating algorithms in Kurdish digit recognition. We used Kurdish dataset named (KurdSet) and Arabic dataset for handwritten recognition, which holds 70,000 images of Arabic digits that were written by 700 various participants. Additionally, various models are utilized in the study, including ResNet50, DenseNet121, MobileNet, and a custom CNN (convolutional neural network). Additionally, the models' effectiveness was assessed through the examination of test accuracy, which measures the percentage of correctly classified digits in the evaluation phase. ResNet50 also performs exceptionally well that achieved test accuracy 99.67%, indicating its All models exhibit good performance, DenseNet121 and the Custom CNN Model demonstrate the highest test accuracy of 99.73%, highlighting their superior performance. capabilities in capturing relevant features. Despite its accuracy, MobileNet still exhibits good recognition capability with a test accuracy 99.54%. [ABSTRACT FROM AUTHOR]

Deep learning (DL), in particular deep neural networks, by default is purely datadriven and in general does not require physics. This is the strength of DL but also one of its key limitations when applied to science and engineering problems in which underlying physical properties--such as stability, conservation, and positivity--and accuracy are required. DL methods in their original forms are often not capable of respecting the underlying mathematical models or achieving desired accuracy even in big-data regimes. On the other hand, many data-driven science and engineering problems, such as inverse problems, typically have limited experimental or observational data, and DL would overfit the data in this case. Leveraging information encoded in the underlying mathematical models, we argue, not only compensates for missing information in low data regimes but also provides opportunities to equip DL methods with the underlying physics, hence promoting better generalization. This paper develops a model-constrained DL approach and its variant TNet--a Tikhonov neural network--which are capable of learning not only information hidden in the training data but also in the underlying mathematical models to solve inverse problems governed by partial differential equations in low data regimes. We provide the constructions and some theoretical results for the proposed approaches for both linear and nonlinear inverse problems. Since TNet is designed to learn inverse solutions with Tikhonov regularization, it is interpretable: in fact it recovers Tikhonov solutions for linear cases while potentially approximating Tikhonov solutions for nonlinear inverse problems. We also prove that data randomization can enhance not only the smoothness of the networks but also their generalizations. Comprehensive numerical results confirm the theoretical findings and show that with even as little as 1 training data sample for one-dimensional (1D) deconvolution, 5 for an inverse 2D heat conductivity problem, 100 for inverse initial conditions for a time-dependent 2D Burgers's equation, and 50 for inverse initial conditions for 2D Navier--Stokes equations, TNet solutions can be as accurate as Tikhonov solutions while being several orders of magnitude faster. This is possible owing to the model-constrained term, replications, and randomization. [ABSTRACT FROM AUTHOR]

Parametric optimal control problems governed by partial differential equations (PDEs) are widely found in scientific and engineering applications. Traditional grid-based numerical methods for such problems generally require repeated solutions of PDEs with different parameter settings, which is computationally prohibitive, especially for problems with high-dimensional parameter spaces. Although recently proposed neural network methods make it possible to obtain the optimal solutions simultaneously for different parameters, challenges still remain when dealing with problems with complex constraints. In this paper, we propose the AONN, an adjoint-oriented neural network method, to overcome the limitations of existing approaches in solving parametric optimal control problems. In an AONN, the neural networks are served as parametric surrogate models for the control, adjoint, and state functions to get the optimal solutions all at once. In order to reduce the training difficulty and handle complex constraints, we introduce an iterative training framework inspired by the classical direct-adjoint looping method so that penalty terms arising from the Karush--Kuhn--Tucker system can be avoided. Once the training is done, parameter-specific optimal solutions can be quickly computed through the forward propagation of the neural networks, which may be further used for analyzing the parametric properties of the optimal solutions. The validity and efficiency of the AONN is demonstrated through a series of numerical experiments with problems involving various types of parameters. [ABSTRACT FROM AUTHOR]

Electricity load exhibits daily and weekly cyclical patterns as well as random characteristics. At present, prevailing deep learning models cannot learn electricity load cyclical and stochastic features adequately. This results in insufficient prediction accuracy and the scalability of current methods. To tackle these difficulties, this paper proposes a framework for electrical load prediction based on an Attention Mechanism Time Series Depthwise Separable Convolutional Neural Network (ELPF-ATDSCN). The framework starts by using the Maximum Information Coefficient for exogenous variable selection. It then incorporates a seasonal decomposition algorithm with manual feature engineering to extract the cyclical and stochastic features of the electrical load. Subsequently, the framework employs the ATDSCN to learn the cyclical and stochastic features of the electrical load. In addition, the Bayesian algorithm optimizes model hyperparameters for optimal model performance. Experimental results of point and interval load prediction on datasets from the US and Nordic power markets reveal that the ATDSCN model proposed in this paper enhances load prediction accuracy compared with other models. It can provide more reliable predictions for power system operation and dispatch. • A new attention mechanism is designed to improve the model memory capability. • The ATDSCN model is proposed for learning periodic and stochastic features of electricity loads. • A framework for electricity load prediction is proposed to improve the accuracy of electricity load forecasting. • The effectiveness of the proposed method is validated on the real world dataset. [ABSTRACT FROM AUTHOR]

• A novel water transparency calculation algorithm is introduced, employing deep neural network and image processing technology. This algorithm obviates the need for a water gauge, allowing the calculation of water transparency solely through the analysisof Secchi disk videos. Notably, it demonstrates heightened detection efficiency and success rates. • An enhanced Siamese tracker, termed SiamDCFF (Siamese tracker based on dual correlation feature fusion), is proposed for Secchi disk recognition. Distinguishing features of SiamDCFF include a dual correlation module and a feature fusion module. On the Secchi disk dataset, SiamDCFF achieves the highest success rate, outperforming other methods detailed in this paper. • The simplified 3D-ResNet [41] is leveraged to assess the state of the Secchi disk, transforming the critical position determination problem into an optimization challenge. The objective function involves maximizing the number of visible and invisible Secchi disk images on both sides of the critical position. This method facilitates adaptive critical position determination. • A novel Gated Recurrent Unit based on auto-regression (ARGRU) is proposed, taking into account the motion characteristics of the Secchi disk. ARGRU achieves a higher success detection rate for water transparency compared to the water gauge recognition method. Notably, ARGRU outperforms RNN [42], GRU, and LSTM [43] on the dataset presented in this paper. Water transparency is traditionally assessed through visual inspection using a lowered Secchi disk (SD) into the water, with the disappearance depth of the SD recorded as the measure of transparency. However, manual and visually-dependent frequency measurements of the SD render this process labor-intensive and time-consuming. This paper presents a comprehensive machine-vision-based algorithm designed for the automatic computation of water transparency using Secchi disk videos. The algorithm leverages multiple deep neural networks, including an enhanced Siamese tracking model, a 3D-ResNet, and an improved GRU network. Trained models directly identify the SD in the video, calculating water transparency by processing pixel information. Experimental results demonstrate the algorithm's efficacy in estimating water transparency across diverse natural environments, achieving commendable accuracy (MAE = 3.6 cm MSE = 21.5 cm RMSE = 4.6 cm), rapid processing speed (average 6.87 s), and robust stability. In comparison with the former water-gauge-based algorithm, our proposed algorithm exhibits heightened efficiency and a superior detection success rate. [ABSTRACT FROM AUTHOR]

The extraction of fault-induced repetitive transients which possess cyclo-stationarity is the key to the fault diagnosis of rotating machinery, which is of considerable significance for ensuring the safe and reliable operation of machinery equipment. Traditional deconvolution methods mainly aim to recover fault-related impulsive features from the time domain and are prone to give poor fault diagnosis results under heavy interference conditions. To solve this problem, a spectrum sparse deep deconvolution method (SSDD) with a deep neural network structure is proposed in this paper. The proposed method uses an envelope spectrum sparse criterion as the cost function to seek an optimal inverse filter through a deep neural network. Firstly, a special band-averaging strategy is designed to initialize the filters in the input layer of the neural network with a window method to provide a direction for deconvolution. Secondly, envelope spectral kurtosis that can depict the sparse feature in the envelope spectrum domain is taken as the cost function to guide the training of the deep network and lock the fault information. Then, the optimal weights are realized by the eigenvalue algorithm, and the weak sparse features are enhanced and extracted layer by layer. Finally, the most significant fault information is obtained through dimension reduction. The simulated and experimental data analysis results verified that the proposed method is superior to traditional deconvolution methods in fault diagnosis performance and robustness to random impulses and strong background noise. [ABSTRACT FROM AUTHOR]

Facial image deblurring is an active research area that aims to restore blurry face images to clear ones. However, this task requires special consideration to restore the detailed elements of facial structures, such as eyes, nose, and mouth. Additionally, facial occlusions and varying illuminance conditions in common environments can degrade the deblurring performance. Previous studies have not accounted for these conditions, necessitating the development of deblurring method that considers these factors. In this paper, we propose a novel approach, called Illuminance-robust Multi-stage DeblurNet with Channel Attention (IMDeCA), which leverages semantic mask and landmark information of the face to restore detailed facial structures. Our approach is robust to various illuminance conditions and facial occlusions. The proposed network comprises a multi-stage structure that extracts facial semantic feature maps, reconstructs clear images, and improves illuminance. We also consider facial landmark information in the loss function to ensure well-restored facial structures even in the presence of facial occlusions. Furthermore, we construct a new facial image dataset, named BIO, which includes Blurred images with various types of Illuminance conditions and facial Occlusions. Through extensive experiments on this dataset, we demonstrate the superior performance of our proposed network, outperforming the latest existing methods. [ABSTRACT FROM AUTHOR]

This paper proposes a novel hardware architecture for a Feed-Forward Neural Network (FFNN) with the objective of minimizing the number of execution clock cycles needed for the network's computation. The proposed architecture depends mainly on using two physical layers that are multiplexed and reused during the computation of the FFNN to achieve an efficient parallel design. Two physical layers are designed to handle the computation of different sizes of Neural Networks (NN). The proposed FFNN architecture hardware resources are independent of the NN's number of layers, instead, they depend only on the number of neurons in the largest layer. This versatile architecture serves as an accelerator in Deep Neural Network (DNN) computations as it exploits parallelism by making the two physical layers work in parallel through the computations. The proposed implementation was implemented with 18-bit fixed point representation reaching 200 MHz clock speed on a Spartan7 FPGA. Furthermore, the proposed architecture achieves a lower neuron computation factor compared to previous works in the literature. [Display omitted] [ABSTRACT FROM AUTHOR]

• An efficient hybrid network for supercontinuum generation prediction proposed. • Wasserstein generative adversarial network augments the original dataset. • Deep neural network with a TCN layer predicts supercontinum spectrum accurately. In this paper, a Hybrid Network is reported for predicting the output pulse shape parameters of the supercontinuum in optical fibers. The network incorporates the Wasserstein generative adversarial network (WGAN) and the deep neural network (DNN) with a TCN layer. The WGAN augments a limited set of original data, while the designed DNN accurately predicts supercontinuum. The simulation results show the spectra agreement (A gm) of the predictions by individual DNN is 89% and the root mean square error (RMSE) is 0.1092. After retraining with WGAN's extended dataset, the A gm of DNN predictions is improved to 98% and the RMSE is reduced to 0.0513. The input pump power corresponding to the appearance of the fourth soliton peak can be predicted by the retrained DNN. Numerical analysis demonstrates the effectiveness and generalization ability of a Hybrid Network in predicting pulse parameters. The network fits the propagation dynamics in the spectral domain of SC, which provides a promising tool for further research in this field. [ABSTRACT FROM AUTHOR]

The success of deep learning on image classification and recognition tasks has led to new applications in diverse contexts, including the field of medical imaging. However, two properties of deep neural networks (DNNs) may limit their future use in medical applications. The first is that DNNs require a large amount of labeled training data, and the second is that the deep learning-based models lack interpretability. In this paper, we propose and investigate a data-efficient framework for the task of general medical image segmentation. We address the two aforementioned challenges by introducing domain knowledge in the form of a strong prior into a deep learning framework. This prior is expressed by a customized dynamical system. We performed experiments on two different datasets, namely JSRT and ISIC2016 (heart and lungs segmentation on chest X-ray images and skin lesion segmentation on dermoscopy images). We have achieved competitive results using the same amount of training data compared to the state-of-the-art methods. More importantly, we demonstrate that our framework is extremely data-efficient, and it can achieve reliable results using extremely limited training data. Furthermore, the proposed method is rotationally invariant and insensitive to initialization. [Display omitted] [ABSTRACT FROM AUTHOR]

Optical Character Recognition (OCR) is a method to convert a scanned photo of"handwritten character recognition (HCR) or printed character recognition (PCR)" into a form of digital text. HCR is a version of OCR, which is remarkably modeled to identify handwritten text, while PCR aims at printed text identification. The identification of handwritten characters and digits is more complicated as compared to PCR because of the diversities in human writing styles, stokes, thickness, and curves of characters. Similarly, achievements in several computer vision tasks consider the Convolutional Neural Networks (CNN) to give an end-to-end solution for HCR with huge success. However, the process of significant feature learning for the identification of images is complicated with little data. Hence, this paper aims to develop a new handwritten character and digit recognition model with the incorporation of a deep learning strategy. Initially, the data related to Indian languages are collected from the standard benchmark datasets. Then, the collected data are given into the feature extraction phase 1, where the ResNet 151 is used for extracting the feature set 1. Similarly, the data gathered are considered in the feature extraction phase 2, where the Optimal Ensemble Pattern extraction approach is developed with Local Binary Pattern (LBP), Local Gradient Patterns (LGP), Local Tetra Pattern (LTrP), and Local Vector Pattern (LVP) for extracting the significant patterns from the language data. These extracted patterns are given into the ResNet 151 for getting the feature set 2. Here, the features from ResNet 151 get optimized with the enhanced optimization algorithm with Fitness-based Sail Fish Optimizer (F-SFO). The obtained feature set 1 and optimal feature set 2 are concatenated for performing final recognition. At last, the HCR is done with the help of developed Bi-LSTM-DNN to achieve the enhanced and accurate recognition of handwritten characters of the Indian languages. The performances of character recognition are further improved with the parameter optimization in Bi-LSTM-DNN with the same enhanced F-SFO. Overall result analysis, the accuracy of the designed method attains 95.12%. [ABSTRACT FROM AUTHOR]

The task of motion prediction is pivotal for autonomous driving systems, providing crucial data to choose a vehicle behavior strategy within its surroundings. Existing motion prediction techniques primarily focus on predicting the future trajectory of each agent in the scene individually, utilizing its past trajectory data. In this paper, we introduce an end-to-end neural network methodology designed to predict the future behaviors of all dynamic objects in the environment. This approach leverages the occupancy map and the scene's motion flow. We are investigating various alternatives for constructing a deep encoder–decoder model called OFMPNet. This model uses a sequence of bird's-eye-view road images, occupancy grid, and prior motion flow as input data. The encoder of the model can incorporate transformer, attention-based, or convolutional units. The decoder considers the use of both convolutional modules and recurrent blocks. Additionally, we propose a novel time-weighted motion flow loss, whose application has shown a substantial decrease in end-point error. Our approach has achieved state-of-the-art results on the Waymo Occupancy and Flow Prediction benchmark, with a Soft IoU of 52.1% and an AUC of 76.75% on Flow-Grounded Occupancy. [ABSTRACT FROM AUTHOR]

Alternating current optimal power flow (OPF) analysis is critical for efficient and reliable operation of power systems. For large systems or repetitive computations, the traditional methods such as the direct and gradient methods, or non-traditional methods, such as the genetic algorithm and simulating annealing, are time-consuming and unsuitable for real-time computing. The work in this paper proposes a novel framework to obtain the optimal solution of power flow in real-time using a combination of convolutional neural networks and a self-attention mechanism. All parameters of the power networks are rearranged in an image-like shape of a multi-channel image where each channel is a two-dimensional matrix. The proposed approach is adaptive with every input size of power systems as well as frequent variations of network topologies without intervention to the framework core. The encompassment of all power system contexts in which all parameters of internal elements, generation costs, and topology information are included, contributes to the higher accuracy of inference compared to other current machine-learning-based OPF-solving methods. Besides, the proposed framework established on ubiquitous platforms is effortlessly integrated into current infrastructures of power systems, and the great efficiency along with the computation speed may serve as a critical point for practical implications, such as enabling faster decision-making during real-time operations, predicting system contingencies, and remedial actions based on an offline pre-trained model. This supervised learning process is applied to the dataset of four case studies of meshed power systems: the IEEE 5-bus system (IEEE-5), the IEEE 30-bus system (IEEE-30), the IEEE 39-bus system (IEEE-39), and the IEEE 57-bus system (IEEE-57) to prove the efficacy of the proposed method. • Developing a novel approach to solve the OPF problems in real-time. • Wrap power system characteristics under a mold of an image-like shape. • Overcome the outburst of trainable parameters when training with large-scale systems. • Do not require looking up a new set of hyperparameters or re-coding a framework. • All time-consuming computations are shifted to the offline training process. [ABSTRACT FROM AUTHOR]

• A novel approach for single channel speech enhancement is presented, using the colored spectrograms. • Using deep neural network for spectrogram denoising and regression neural network for translating the colors of denoised spectrograms to short time Fouier transform (STFT) magnitudes. • Output is comparable to the state-of-the-art baseline models at a much reduced computational cost. Speech enhancement concerns the processes required to remove unwanted background sounds from the target speech to improve its quality and intelligibility. In this paper, a novel approach for single-channel speech enhancement is presented using colored spectrograms. We propose the use of a deep neural network (DNN) architecture adapted from the pix2pix generative adversarial network (GAN) and train it over colored spectrograms of speech to denoise them. After denoising, the colors of spectrograms are translated to magnitudes of short-time Fourier transform (STFT) using a shallow regression neural network. These estimated STFT magnitudes are later combined with the noisy phases to obtain an enhanced speech. The results show an improvement of almost 0.84 points in the perceptual evaluation of speech quality (PESQ) and 1 % in the short-term objective intelligibility (STOI) over the unprocessed noisy data. The gain in quality and intelligibility over the unprocessed signal is almost equal to the gain achieved by the baseline methods used for comparison with the proposed model, but at a much reduced computational cost. The proposed solution offers a comparative PESQ score at almost 10 times reduced computational cost than a similar baseline model that has generated the highest PESQ score trained on grayscaled spectrograms, while it provides only a 1 % deficit in STOI at 28 times reduced computational cost when compared to another baseline system based on convolutional neural network-GAN (CNN-GAN) that produces the most intelligible speech. [ABSTRACT FROM AUTHOR]

With the significant increase in photovoltaic (PV) electricity generation, more attention has been given to PV power forecasting. Indeed, accurate forecasting allows power grid operators to better schedule and dispatch their assets, such as energy storage systems and reserve. In this paper, a hybrid deep learning model and a convolutional neural network with memory is proposed, to provide intraday (2 h) solar irradiance forecasts using sequentially-collected images from public webcams. The performance of the proposed model is compared to those of a standard time-series forecast models, a linear regression as well as state-of-the-art neural networks. All models are trained and tested using images collected from two webcams on EPFL's campus for just over a year. The results show that the proposed method outperforms all other models and matches the state-of-the-art methodology while providing simplicity of implementation and efficient computation. • A novel method for estimating global horizontal irradiance is proposed. • The method uses images public webcams instead of all-sky cameras. • The method combines images from different cameras to improve forecasts' accuracy. • The method can be extended to include other types of data (meteorological data). • This simple method matches state-of-the-art methods. [ABSTRACT FROM AUTHOR]