Search

Artificial intelligence is more present than ever, both in our society in general and in science. At the center of this development has been the concept of deep learning, the use of artificial neural networks that are many layers deep and can often reproduce human‐like behavior much better than other machine‐learning techniques. The articles in this collection are some recent examples of its application for X‐ray photon science and crystallography that have been published in Journal of Applied Crystallography. [ABSTRACT FROM AUTHOR]

This document is a guest editorial for a special issue on computational methods and artificial intelligence applications in low-carbon energy systems. The editorial highlights the urgent need for advanced computing and artificial intelligence in the clean energy transition to improve system reliability, economics, and sustainability. The special issue includes 19 original research articles covering topics such as energy forecasting, situational awareness, multi-energy system dispatch, and power system operation. The articles present state-of-the-art methods and techniques in these areas, including wind power forecasting, demand-side flexibility, fault diagnosis of photovoltaic strings, and energy management strategies. The authors express their gratitude to the participating authors and anonymous reviewers for their contributions to the special section. [Extracted from the article]

The significant risk associated with fault reactivation often necessitates slip tendency analyses for effective risk assessment. However, such analyses are challenging, particularly in large areas with limited or absent reliable stress measurements and where the cost of extensive geomechanical analyses or simulations is prohibitive. In this paper, we propose a novel approach using a physics‐informed neural network that integrates stress orientation and satellite displacement observations in a top‐down multi‐scale framework to estimate two‐dimensional slip tendency analyses even in regions lacking comprehensive stress data. Our study demonstrates that velocities derived from a continental scale analysis, combined with reliable stress orientation averages, can effectively guide models at smaller scales to generate qualitative slip tendency maps. By offering customizable data selection and stress resolution options, this method presents a robust solution to address data scarcity issues, as exemplified through a case study of the South Australian Eyre Peninsula. Plain Language Summary: Fault reactivation poses significant risks, often requiring slip tendency analyses for thorough risk assessment. Yet, such analyses face challenges, especially in large areas lacking reliable stress measurements or where extensive geomechanical analyses are too costly. Our paper suggests a new method using a physics‐based neural network. This approach combines compressive direction and satellite displacement observations to estimate slip tendencies in two dimensions, even where stress data is lacking. Our study shows that by using displacements from a continental scale analysis and reliable averages of compressive directions, we can guide models to create smaller‐scale maps indicating where faults are more likely to reactivate. This method allows for customizable data selection and stress resolution, offering a strong solution to data scarcity issues. We demonstrate its effectiveness through a case study of South Australia's Eyre Peninsula. Key Points: Physics‐based neural networks allow two‐dimensional slip tendency analyses without prior full‐stress informationA multi‐scale approach provides required displacement constraints when inferring full stresses from global navigation satellite system (GNSS) and stress orientation dataWe present a new application for GNSS data that would welcome more stations, even in seismically stable areas [ABSTRACT FROM AUTHOR]

Eye health has become a global health concern and attracted broad attention. Over the years, researchers have proposed many state‐of‐the‐art convolutional neural networks (CNNs) to assist ophthalmologists in diagnosing ocular diseases efficiently and precisely. However, most existing methods were dedicated to constructing sophisticated CNNs, inevitably ignoring the trade‐off between performance and model complexity. To alleviate this paradox, this paper proposes a lightweight yet efficient network architecture, mixed‐decomposed convolutional network (MDNet), to recognise ocular diseases. In MDNet, we introduce a novel mixed‐decomposed depthwise convolution method, which takes advantage of depthwise convolution and depthwise dilated convolution operations to capture low‐resolution and high‐resolution patterns by using fewer computations and fewer parameters. We conduct extensive experiments on the clinical anterior segment optical coherence tomography (AS‐OCT), LAG, University of California San Diego, and CIFAR‐100 datasets. The results show our MDNet achieves a better trade‐off between the performance and model complexity than efficient CNNs including MobileNets and MixNets. Specifically, our MDNet outperforms MobileNets by 2.5% of accuracy by using 22% fewer parameters and 30% fewer computations on the AS‐OCT dataset. [ABSTRACT FROM AUTHOR]

In recent years, research on infrared and visible image fusion has mainly focused on deep learning‐based approaches, particularly deep neural networks with auto‐encoder architectures. However, these approaches suffer from problems such as insufficient feature extraction capability and inefficient fusion strategies. Therefore, this paper introduces a novel image fusion network to address the limitations of infrared and visible image fusion networks with auto‐encoder architectures. In the designed network, the encoder employs a multi‐branch cascade structure, and these convolution branches with different kernel sizes provide the encoder with an adaptive receptive field to extract multi‐scale features. In addition, the fusion layer incorporates a non‐local attention module that is inspired by the self‐attention mechanism. With its global receptive field, this module is used to build a non‐local attention fusion network, which works together with the l1${l}_1$‐norm spatial fusion strategy to extract, split, filter, and fuse global and local features. Comparative experiments on the TNO and MSRS datasets demonstrate that the proposed method outperforms other state‐of‐the‐art fusion approaches. [ABSTRACT FROM AUTHOR]

With the introduction of more recent deep learning models such as encoder‐decoder, text generation frameworks have gained a lot of popularity. In Natural Language Generation (NLG), controlling the information and style of the output produced is a crucial and challenging task. The purpose of this paper is to develop informative and controllable text using social media language by incorporating topic knowledge into a keyword‐to‐text framework. A novel Topic‐Controllable Key‐to‐Text (TC‐K2T) generator that focuses on the issues of ignoring unordered keywords and utilising subject‐controlled information from previous research is presented. TC‐K2T is built on the framework of conditional language encoders. In order to guide the model to produce an informative and controllable language, the generator first inputs unordered keywords and uses subjects to simulate prior human knowledge. Using an additional probability term, the model increases the likelihood of topic words appearing in the generated text to bias the overall distribution. The proposed TC‐K2T can produce more informative and controllable senescence, outperforming state‐of‐the‐art models, according to empirical research on automatic evaluation metrics and human annotations. [ABSTRACT FROM AUTHOR]

Coral reefs are among the most diverse ecosystems on our planet, and essential to the livelihood of hundreds of millions of people who depend on them for food security, income from tourism and coastal protection. Unfortunately, most coral reefs are existentially threatened by global climate change and local anthropogenic pressures. To better understand the dynamics underlying deterioration of reefs, monitoring at high spatial and temporal resolution is key. However, conventional monitoring methods for quantifying coral cover and species abundance are limited in scale due to the extensive manual labor required. Although computer vision tools have been employed to aid in this process, in particular structure‐from‐motion (SfM) photogrammetry for 3D mapping and deep neural networks for image segmentation, analysis of the data products creates a bottleneck, effectively limiting their scalability.This paper presents a new paradigm for mapping underwater environments from ego‐motion video, unifying 3D mapping systems that use machine learning to adapt to challenging conditions under water, combined with a modern approach for semantic segmentation of images.The method is exemplified on coral reefs in the northern Gulf of Aqaba, Red Sea, demonstrating high‐precision 3D semantic mapping at unprecedented scale with significantly reduced required labor costs: given a trained model, a 100 m video transect acquired within 5 min of diving with a cheap consumer‐grade camera can be fully automatically transformed into a semantic point cloud within 5 min. We demonstrate the spatial accuracy of our method and the semantic segmentation performance (of at least 80% total accuracy), and publish a large dataset of ego‐motion videos from the northern Gulf of Aqaba, along with a dataset of video frames annotated for dense semantic segmentation of benthic classes.Our approach significantly scales up coral reef monitoring by taking a leap towards fully automatic analysis of video transects. The method advances coral reef transects by reducing the labor, equipment, logistics, and computing cost. This can help to inform conservation policies more efficiently. The underlying computational method of learning‐based Structure‐from‐Motion has broad implications for fast low‐cost mapping of underwater environments other than coral reefs. [ABSTRACT FROM AUTHOR]

Background: Accurate segmentation of lung nodules is crucial for the early diagnosis and treatment of lung cancer in clinical practice. However, the similarity between lung nodules and surrounding tissues has made their segmentation a longstanding challenge. Purpose: Existing deep learning and active contour models each have their limitations. This paper aims to integrate the strengths of both approaches while mitigating their respective shortcomings. Methods: In this paper, we propose a few‐shot segmentation framework that combines a deep neural network with an active contour model. We introduce heat kernel convolutions and high‐order total variation into the active contour model and solve the challenging nonsmooth optimization problem using the alternating direction method of multipliers. Additionally, we use the presegmentation results obtained from training a deep neural network on a small sample set as the initial contours for our optimized active contour model, addressing the difficulty of manually setting the initial contours. Results: We compared our proposed method with state‐of‐the‐art methods for segmentation effectiveness using clinical computed tomography (CT) images acquired from two different hospitals and the publicly available LIDC dataset. The results demonstrate that our proposed method achieved outstanding segmentation performance according to both visual and quantitative indicators. Conclusion: Our approach utilizes the output of few‐shot network training as prior information, avoiding the need to select the initial contour in the active contour model. Additionally, it provides mathematical interpretability to the deep learning, reducing its dependency on the quantity of training samples. [ABSTRACT FROM AUTHOR]

Text-independent speaker verification (TI-SV) is a crucial task in speaker recognition, as it involves verifying an individual's claimed identity from speech of arbitrary content without any human intervention. The target for TI-SV is to design a discriminative network to learn deep speaker embedding for speaker idiosyncrasy. In this paper, we propose a deep speaker embedding learning approach of a hybrid deep neural network (DNN) for TI-SV in FM broadcasting. Not only acoustic features are utilized, but also phoneme features are introduced as prior knowledge to collectively learn deep speaker embedding. The hybrid DNN consists of a convolutional neural network architecture for generating acoustic features and a multilayer perceptron architecture for extracting phoneme features sequentially, which represent significant pronunciation attributes. The extracted acoustic and phoneme features are concatenated to form deep embedding descriptors for speaker identity. The hybrid DNN demonstrates not only the complementarity between acoustic and phoneme features but also the temporality of phoneme features in a sequence. Our experiments show that the hybrid DNN outperforms existing methods and delivers a remarkable performance in FM broadcasting TI-SV. [ABSTRACT FROM AUTHOR]

Producing clean energy and feeding critical loads in islanding mode are the main reasons for interest in microgrids. Different operation topologies of microgrids make traditional protection schemes inefficient. This paper proposes a multi‐agent protection scheme in which each protection agent can detect different fault events and isolate faulty phases at a fast rate. A unique algorithm is utilized for determining fault location in microgrids and system operators are informed accordingly. Microgrids have various operation modes due to the stochastic behavior of distributed generators and different topologies. Here, a significant number of operating conditions of the studied microgrid are considered. These operation conditions are simulated in the DIgSILENT Power Factory, and different parameters are stored. Raw measured parameters need to be pre‐processed by a signal processing method in MATLAB. Discrete wavelet transform is chosen for this purpose. Deep learning is used as a machine learning technique due to the various operation modes of the microgrid. Deep neural networks are constructed using Python programming language. The proposed scheme ensures high accuracy in fault detection and fault location in the microgrid, as well as fault isolation in different operation conditions. We proposed a multi‐agent algorithm for fault detection in alternating current (AC) microgrids that can diagnose the type, phase, location, and impedance of faults. This scheme performs comprehensive case studies containing possible operation conditions such as lateral branches, high fault impedance, and different types of distributed generation (DG). The proposed scheme has higher efficiency in comparison with other state‐of‐the‐art. [ABSTRACT FROM AUTHOR]

Deep learning is gaining importance due to its wide range of applications. Many researchers have utilized deep learning (DL) models for the automated diagnosis of cancer patients. This paper provides a systematic review of DL models for automated diagnosis of cancer patients. Initially, various DL models for cancer diagnosis are presented. Five major categories of cancers such as breast, lung, liver, brain and cervical cancer are considered. As these categories of cancers have a very high percentage of occurrences with high mortality rate. The comparative analysis of different types of DL models is drawn for the diagnosis of cancer at early stages by considering the latest research articles from 2016 to 2022. After comprehensive comparative analysis, it is found that most of the researchers achieved appreciable accuracy with implementation of the convolutional neural network model. These utilized the pretrained models for automated diagnosis of cancer patients. Various shortcomings with the existing DL‐based automated cancer diagnosis models are also been presented. Finally, future directions are discussed to facilitate further research for automated diagnosis of cancer patients. [ABSTRACT FROM AUTHOR]

Drought is an important meteorological event in China and can cause severe damage to both livelihoods and socio‐ecological systems, but current seasonal prediction skill for drought is far from successful. This study used convolutional neural network (CNN) to build an effective seasonal forecast model for the summer consecutive dry days (CDD) over China. The principal components (PC) of the six leading empirical orthogonal function modes of CDD anomaly were predicted by CNN, using the previous winter precipitation, 2‐m temperature and 500 hPa geopotential height as predictors. These predicted PCs were then projected onto the observed spatial patterns to obtain the final predicted summer CDD anomaly over China. In the independent hindcast period of 2007–2018, the interannual variabilities of first six PCs were significantly predicted by CNN. The spatial patterns of CDD were then skillfully predicted with anomaly correlation coefficient skills generally higher than 0.2. The interannual variability of summer CDD over the middle and lower Yangtze River valley, northwestern China and northern China were significant predicted by our CNN model three months in advance. CNN identified that the previous winter precipitation was the important predictor for PC1–PC3, whereas the previous winter 2‐m temperature and 500 hPa geopotential height were important for the prediction of PC4–PC6. This research provides a new and effective method for the seasonal prediction of summer drought. Plain Language Summary: Drought can cause serious agricultural and ecosystem disasters, so its forecast is valuable for preventing and mitigating related natural disasters and regional socioeconomic sustainability. However, current prediction skill for drought is far from successful since its extreme feature. The gradually emerging deep learning methods offer new possibilities, but how to effectively apply deep learning models in climate prediction with a small sample size remains an open question. In this paper, we build seasonal prediction convolutional neural network model for summer consecutive dry days over China using previous winter predictors. This model achieves significant prediction skill three months in advance. The empirical orthogonal function decomposition is used to reduce the dimensionality of consecutive dry days data in our model. Our research provides a new perspective for drought prediction, and it is expected that such method will be also useful for other seasonal climate prediction problems. Key Points: Convolutional neural network (CNN) skillfully predicts summer consecutive dry days (CDD) over China three months in advanceThe principal components of CDD are predicted by CNN and then projected on the observed spatial patternsPrevious winter 2‐m temperature, geopotential height at 500 hPa and precipitation are the essential predictors in CNN [ABSTRACT FROM AUTHOR]

Since the discovery of the quasicrystal, approximately 100 stable quasicrystals are identified. To date, the existence of quasicrystals is verified using transmission electron microscopy; however, this technique requires significantly more elaboration than rapid and automatic powder X‐ray diffraction. Therefore, to facilitate the search for novel quasicrystals, developing a rapid technique for phase‐identification from powder diffraction patterns is desirable. This paper reports the identification of a new Al–Si–Ru quasicrystal using deep learning technologies from multiphase powder patterns, from which it is difficult to discriminate the presence of quasicrystalline phases even for well‐trained human experts. Deep neural networks trained with artificially generated multiphase powder patterns determine the presence of quasicrystals with an accuracy >92% from actual powder patterns. Specifically, 440 powder patterns are screened using the trained classifier, from which the Al–Si–Ru quasicrystal is identified. This study demonstrates an excellent potential of deep learning to identify an unknown phase of a targeted structure from powder patterns even when existing in a multiphase sample. [ABSTRACT FROM AUTHOR]

The surface texture of asphalt pavement has a significant effect on skid resistance performance. However, its contribution to the performance of skid resistance is non‐homogeneous and subjects to local validity. There are also a few deep learning models that take into account the effective contact texture region. This paper proposes a convolutional neural network model based on the effective contact texture region, containing macro‐ and micro‐scale awareness sub‐modules. In this study, the asphalt mixture with varying gradations was designed to accurately obtain the effective contact texture region. Then, the textures were disentangled into macro‐ and micro‐texture scales by applying the fast Fourier transform and fed into the model for training. Finally, the area of effective contact texture region was calculated, and the effective contact ratio parameter was then proposed using the triangulation algorithm. The results showed that the effective contact texture area of pavement varies by the asphalt mixture type. The effective contact ratio parameter exhibited a significant positive correlation (Pearson correlation coefficient is 0.901, R2= 0.8129) with skid resistance performance and was also influenced by key sieve aggregate content from 2.36 to 4.75 mm. The data of effective contact texture region following disentanglement significantly released the model performance (the relative error dropped to 1.81%). The model exhibited improved precision and performance, which can be utilized as an efficient, non‐contact alternative method for skid resistance analysis. [ABSTRACT FROM AUTHOR]

Purpose: In MRI, motion artifacts can significantly degrade image quality. Motion artifact correction methods using deep neural networks usually required extensive training on large datasets, making them time‐consuming and resource‐intensive. In this paper, an unsupervised deep learning‐based motion artifact correction method for turbo‐spin echo MRI is proposed using the deep image prior framework. Theory and Methods: The proposed approach takes advantage of the high impedance to motion artifacts offered by the neural network parameterization to remove motion artifacts in MR images. The framework consists of parameterization of MR image, automatic spatial transformation, and motion simulation model. The proposed method synthesizes motion‐corrupted images from the motion‐corrected images generated by the convolutional neural network, where an optimization process minimizes the objective function between the synthesized images and the acquired images. Results: In the simulation study of 280 slices from 14 subjects, the proposed method showed a significant increase in the averaged structural similarity index measure by 0.2737 in individual coil images and by 0.4550 in the root‐sum‐of‐square images. In addition, the ablation study demonstrated the effectiveness of each proposed component in correcting motion artifacts compared to the corrected images produced by the baseline method. The experiments on real motion dataset has shown its clinical potential. Conclusion: The proposed method exhibited significant quantitative and qualitative improvements in correcting rigid and in‐plane motion artifacts in MR images acquired using turbo spin‐echo sequence. [ABSTRACT FROM AUTHOR]

It is important for construction personnel to observe the dress code, such as the correct wearing of safety helmets and reflective vests is conducive to protecting the workers' lives and safety of construction. A YOLO network‐based detection algorithm is proposed for the construction personnel dress code (YOLO‐CPDC). Firstly, Multi‐Head Self‐Attention (MHSA) is introduced into the backbone network to build a hybrid backbone, called Convolution MHSA Network (CMNet). The CMNet gives the model a global field of view and enhances the detection capability of the model for small and obscured targets. Secondly, an efficient and lightweight convolution module is designed. It is named Ghost Shuffle Attention‐Conv‐BN‐SiLU (GSA‐CBS) and is used in the neck network. The GSANeck network reduces the model size without affecting the performance. Finally, the SIoU is used in the loss function and Soft NMS is used for post‐processing. Experimental results on the self‐constructed dataset show that YOLO‐CPDC algorithm has higher detection accuracy than current methods. YOLO‐CPDC achieves a mAP50 of 93.6%. Compared with the YOLOv5s, the number of parameters of our model is reduced by 18% and the mAP50 is improved by 1.1%. Overall, this research effectively meets the actual demand of dress code detection in construction scenes. [ABSTRACT FROM AUTHOR]

Purpose: The macular ganglion cell layer (mGCL) is a strong potential biomarker of axonal degeneration in multiple sclerosis (MS). For this reason, this study aims to develop a computer‐aided method to facilitate diagnosis and prognosis in MS. Methods: This paper combines a cross‐sectional study of 72 MS patients and 30 healthy control subjects for diagnosis and a 10‐year longitudinal study of the same MS patients for the prediction of disability progression, during which the mGCL was measured using optical coherence tomography (OCT). Deep neural networks were used as an automatic classifier. Results: For MS diagnosis, greatest accuracy (90.3%) was achieved using 17 features as inputs. The neural network architecture comprised the input layer, two hidden layers and the output layer with softmax activation. For the prediction of disability progression 8 years later, accuracy of 81.9% was achieved with a neural network comprising two hidden layers and 400 epochs. Conclusion: We present evidence that by applying deep learning techniques to clinical and mGCL thickness data it is possible to identify MS and predict the course of the disease. This approach potentially constitutes a non‐invasive, low‐cost, easy‐to‐implement and effective method. [ABSTRACT FROM AUTHOR]

Summary: The rapid advancement and growth of technology have rendered cloud computing services indispensable to our activities. Threats and intrusions have since multiplied exponentially across a range of industries. In such a scenario, the intrusion detection system, or simply the IDS, is deployed on the network to monitor and detect any attacks. The paper proposes a feed‐forward deep neural network (FFDNN) method based on deep learning methodology using a filter‐based feature selection model. The feature selection strategy aims to determine and select the most highly relevant subset of attributes from the feature importance score for training the deep learning model. Three benchmark data sets were used to assess the experiment: CIC‐IDS 2017, UNSW‐NB15, and NSL‐KDD. In order to justify the proposed technique, a comparison was done using other learning algorithms ranging from classical machine learning to ensemble learning methods that can detect various attacks. The experiments showed that the FFDNN model with reduced feature subsets gave the highest accuracy of 99.53% and 94.45% in the NSL‐KDD and UNSW‐NB15 data sets, while the ensemble‐based XGBoost model performed better in the CIC‐IDS 2017 data set. In addition, the results show that the overall accuracy, recall, and F1 score of the deep learning algorithm are generally better for all the data sets. [ABSTRACT FROM AUTHOR]

Summary: Internet data thefts, intrusions and DDoS attacks are some of the big concerns for the network security today. Detection of these anomalies, is gaining tremendous impetus with the development of machine learning and artificial intelligence. Even now researchers are shifting the base from machine learning to the deep neural architectures with auto‐feature selection capabilities. We in this paper propose multiple deep neural network architectures which can select, co‐learn and teach the gradients of the neural network by itself with no human intervention. This is what we call as meta‐learning. The models are configured in both many to one and many to many design architectures. We combine long short‐term memory (LSTM), bi‐directional long short‐term memory (BiLSTM), convolutional neural network (CNN) layers along with attention mechanism to achieve the higher accuracy values among all the available deep learning model architectures. LSTMs overcomes the vanishing and exploding gradient problem of RNN and attention mechanism mimics the human cognitive attention that screens the network flow to obtain the key features for network traffic classification. In addition, we also add multiple convolutional layers to get the key features for network traffic classification. We get the time series analysis of the traffic done for the possibility of a DDoS attack without using any feature selection techniques and without balancing the dataset. The performance analysis is done based on confusion matrix scores, that is, accuracy, false alarm rate (FAR), sensitivity, specificity, false‐positive rate (FPR), F1 score, area under curve (AUC) analysis and loss functions on well‐known public benchmark KDD Cup'99 data set. The results of our experiments reveal that our models outperform existing techniques, showing their superiority in performance. [ABSTRACT FROM AUTHOR]

The response spectrum method has gained widespread acceptance in practical applications owing to its favorable compromise between accuracy and practical efficiency. The method predicts the peak responses of multi‐degree‐of‐freedom (MDOF) systems by combining modal responses. The Square Root of the Sum of Squares (SRSS) and Complete Quadratic Combination (CQC) rules are commonly used for modal combinations. However, it has been widely known that these rules have limitations in accurately predicting responses influenced by higher modes and cross‐modal correlations. To improve the accuracy of the response spectrum analysis method for building structures, this paper proposes a Deep learning‐based modal Combination (DC) rule by introducing modal contribution coefficients predicted by a deep neural network (DNN) model. The DC rule enhances prediction accuracy by considering the characteristics of ground motion and the dynamic properties of a structural system. The DC rule provides more accurate predictions than the conventional rules, particularly for irregular response spectra and responses affected by higher modes. The efficiency and applicability of the DC rule are demonstrated by numerical investigations of multistory shear buildings and steel frame structures with regular and irregular shapes. The source codes, data, and trained models are available for download at https://github.com/tyongkim/ERD2. [ABSTRACT FROM AUTHOR]

Skin cancer is a common type of malignant tumor that poses a serious threat to patients' lives and health, especially melanoma. It may spread to other body parts, resulting in serious complications and death. In the medical field, accurate identification of skin lesion images is crucial for diagnosing different diseases. However, due to the similarity between different skin lesions, it brings some challenges to medical diagnosis. In this paper, a novel Ensemble Learning Model (EL‐DLOA) based on deep learning and optimization algorithms is proposed, which uses four different deep neural network architectures to generate confidence levels for classes, and optimization algorithms are used to integrate these confidence levels to make the final predictions. To ensure the model's accuracy and reliability, it is first trained using three different learning rates to find the best classification performance of the model. Then, a new search method based on the grey wolf optimization algorithm is proposed to enhance the grey wolf search efficiency. The method improves the search mechanism by changing the grey wolf's individual position through random perturbation or adaptive mutation, which solves the problem that the grey wolf algorithm is easy to fall into local optimum. Finally, four different ensemble strategies are used to reduce individual model bias in the classification process. The proposed model is trained and evaluated using the publicly available dataset HAM10000. The experimental results show that the improved grey wolf optimization algorithm effectively avoids the premature convergence problem and improves the search combination efficiency. Furthermore, in the ensemble methods, the adaptive weight average ensemble strategy effectively improves the classification performance, yielding accuracy, precision, recall, and F1 scores of 0.888, 0.837, 0.897, and 0.862, respectively. These metrics show varying degrees of improvement over the best performing single model. In general, the results indicate that the proposed method achieves high accuracy and practicality in skin lesion classification. Our model shows excellent performance in comparison with other existing models, which makes it significant for research and application in dermatology diagnosis. [ABSTRACT FROM AUTHOR]

Background: Diagnosis and treatment management for head and neck squamous cell carcinoma (HNSCC) is guided by routine diagnostic head and neck computed tomography (CT) scans to identify tumor and lymph node features. The extracapsular extension (ECE) is a strong predictor of patients' survival outcomes with HNSCC. It is essential to detect the occurrence of ECE as it changes staging and treatment planning for patients. Current clinical ECE detection relies on visual identification and pathologic confirmation conducted by clinicians. However, manual annotation of the lymph node region is a required data preprocessing step in most of the current machine learning‐based ECE diagnosis studies. Purpose: In this paper, we propose a Gradient Mapping Guided Explainable Network (GMGENet) framework to perform ECE identification automatically without requiring annotated lymph node region information. Methods: The gradient‐weighted class activation mapping (Grad‐CAM) technique is applied to guide the deep learning algorithm to focus on the regions that are highly related to ECE. The proposed framework includes an extractor and a classifier. In a joint training process, informative volumes of interest (VOIs) are extracted by the extractor without labeled lymph node region information, and the classifier learns the pattern to classify the extracted VOIs into ECE positive and negative. Results: In evaluation, the proposed methods are well‐trained and tested using cross‐validation. GMGENet achieved test accuracy and area under the curve (AUC) of 92.2% and 89.3%, respectively. GMGENetV2 achieved 90.3% accuracy and 91.7% AUC in the test. The results were compared with different existing models and further confirmed and explained by generating ECE probability heatmaps via a Grad‐CAM technique. The presence or absence of ECE has been analyzed and correlated with ground truth histopathological findings. Conclusions: The proposed deep network can learn meaningful patterns to identify ECE without providing lymph node contours. The introduced ECE heatmaps will contribute to the clinical implementations of the proposed model and reveal unknown features to radiologists. The outcome of this study is expected to promote the implementation of explainable artificial intelligence‐assiste ECE detection. [ABSTRACT FROM AUTHOR]

The die shear test is a feasible and conventional method to characterize the shear strength of die‐attaching layer materials in electronic packaging. A new method for determining cohesive zone model (CZM) parameters using deep neural networks (DNN) and die shear tests is proposed, different from classical fracture framework or lap shear test‐based methods. With the sintered nano‐silver die shear test, the results show that the bilinear CZM inversion results agree well with the experimental results. It is found that the DNN model has high accuracy in predicting and identifying the maximum shear traction strength τmax, separation displacement of the interface δf, and the interface stiffness k1 of CZM parameters for sintered nano‐silver adhesive layer through die shear test load versus displacement curves. The presented DNN‐aided inverse identifying method through the die shear test in this paper could provide an alternative and convenient method for extracting CZM parameters of various kinds of adhesive materials in electronic packaging. Highlights: Die shear tests were used for the inverse identification of CZM parameters.The die shear test P–δ curves were established as the dataset.A DNN‐aided CZM inverse identification method was proposed.The DNN‐aided model can accurately identify the CZM parameters. [ABSTRACT FROM AUTHOR]

This article discusses two papers that examine different aspects of software reliability using fault injection techniques. The first paper investigates the impact of transient hardware faults on deep learning neural network inference, particularly in safety-critical applications like autonomous vehicles and healthcare systems. The authors enhance fault injection techniques to reveal the significant influence of hardware faults on these applications. The second paper addresses the challenges of fault localization in software debugging and presents a novel approach, Delta4Ms, that mitigates mutant bias and improves fault localization accuracy. These papers provide valuable insights into ensuring software reliability and resilience in different contexts. [Extracted from the article]

This document is an editorial for a special issue on "Edge computing accelerated deep learning: Technologies and applications." The traditional centralized approach to implementing machine learning (ML) and deep learning (DL) applications has limitations such as high latency, large bandwidth usage, and privacy concerns. Edge computing, which integrates mobile/wireless infrastructure and cloud datacenters, has emerged as a paradigm to address these issues. The special issue aims to promote innovative technologies and applications that accelerate DL in the distributed edge computing environment. The editorial highlights five selected papers that focus on using DL and other ML methods to address challenges in edge computing-based application scenarios and designing efficient and lightweight ML and DL models for the edge computing environment. The Lead Guest Editor expresses gratitude to the authors and reviewers for their contributions. [Extracted from the article]

Machine learning‐based approaches demonstrate a significant potential in radar quantitative precipitation estimation (QPE) applications. In contrast to conventional methods that depend on local raindrop size distributions, deep learning (DL) can establish an effective mapping from three‐dimensional radar observations to ground rain rates. However, the lack of transparency in DL models poses challenges toward understanding the underlying physical mechanisms that drive their outcomes. This study aims to develop a DL‐based QPE system and provide a physical explanation of radar precipitation estimation process. This research is designed by employing a deep neural network consisting of two modules. The first module is a quantitative precipitation estimation network that has the capability to learn precipitation patterns and spatial distribution from multidimensional polarimetric radar observations. The second module introduces a quantitative precipitation estimation shapley additive explanations method to quantify the influence of each radar observable on the model estimate across various precipitation intensities. Plain Language Summary: Ground radars can provide continuous spatial observations over large areas with high spatiotemporal resolutions, so they form the infrastructure for precipitation monitoring and observation in many countries. Recently, deep learning (DL) techniques have shown great potential for use in polarimetric radar‐based precipitation estimates. Nevertheless, the black‐box and turn‐key characteristics of DL models make it difficult for researchers to understand the model decision‐making process and cast doubt on the reliability of the model results. This study introduces a physically explainable polarization radar‐based quantitative precipitation estimation (QPE) system built on DL technology that can explain the causes of the precipitation estimates provided by deep learning models under different rainfall amounts. An experiment indicates that our model achieves better estimates than the conventional methods. Furthermore, the explainability methodology allows for visualization of the microphysical precipitation information. Being the initial attempt to apply explainability learning in the QPE domain, the explainability results may offer valuable guidance for rainfall estimation. Key Points: A polarimetric radar‐based rainfall estimation system is developed using deep neural networksThe deep learning‐based rainfall estimates generally outperform products derived from traditional parametric relationsThe proposed deep learning interpretation method can provide physical and statistical explanations of the model decision‐making process [ABSTRACT FROM AUTHOR]

Deep neural networks (DNN) have become a central method in ecology. To build and train DNNs in deep learning (DL) applications, most users rely on one of the major deep learning frameworks, in particular PyTorch or TensorFlow. Using these frameworks, however, requires substantial experience and time. Here, we present 'cito', a user‐friendly R package for DL that allows specifying DNNs in the familiar formula syntax used by many R packages. To fit the models, 'cito' takes advantage of the numerically optimized 'torch' library, including the ability to switch between training models on the CPU or the graphics processing unit (GPU) which allows the efficient training of large DNNs. Moreover, 'cito' includes many user‐friendly functions for model plotting and analysis, including explainable AI (xAI) metrics for effect sizes and variable importance. All xAI metrics as well as predictions can optionally be bootstrapped to generate confidence intervals, including p‐values. To showcase a typical analysis pipeline using 'cito', with its built‐in xAI features, we built a species distribution model of the African elephant. We hope that by providing a user‐friendly R framework to specify, deploy and interpret DNNs, 'cito' will make this interesting class of models more accessible to ecological data analysis. A stable version of 'cito' can be installed from the comprehensive R archive network (CRAN). [ABSTRACT FROM AUTHOR]

Due to hardware limitations, existing hyperspectral (HS) camera often suffer from low spatial/temporal resolution. Recently, it has been prevalent to super‐resolve a low resolution (LR) HS image into a high resolution (HR) HS image with a HR RGB (or multispectral) image guidance. Previous approaches for this guided super‐resolution task often model the intrinsic characteristic of the desired HR HS image using hand‐crafted priors. Recently, researchers pay more attention to deep learning methods with direct supervised or unsupervised learning, which exploit deep prior only from training dataset or testing data. In this article, an efficient convolutional neural network‐based method is presented to progressively super‐resolve HS image with RGB image guidance. Specifically, a progressive HS image super‐resolution network is proposed, which progressively super‐resolve the LR HS image with pixel shuffled HR RGB image guidance. Then, the super‐resolution network is progressively trained with supervised pre‐training and unsupervised adaption, where supervised pre‐training learns the general prior on training data and unsupervised adaptation generalises the general prior to specific prior for variant testing scenes. The proposed method can effectively exploit prior from training dataset and testing HS and RGB images with spectral‐spatial constraint. It has a good generalisation capability, especially for blind HS image super‐resolution. Comprehensive experimental results show that the proposed deep progressive learning method outperforms the existing state‐of‐the‐art methods for HS image super‐resolution in non‐blind and blind cases. [ABSTRACT FROM AUTHOR]

The positional information of objects is crucial to enable robots to perform grasping and pushing manipulations in clutter. To effectively perform grasping and pushing manipulations, robots need to perceive the position information of objects, including the coordinates and spatial relationship between objects (e.g., proximity, adjacency). The authors propose an end‐to‐end position‐aware deep Q‐learning framework to achieve efficient collaborative pushing and grasping in clutter. Specifically, a pair of conjugate pushing and grasping attention modules are proposed to capture the position information of objects and generate high‐quality affordance maps of operating positions with features of pushing and grasping operations. In addition, the authors propose an object isolation metric and clutter metric based on instance segmentation to measure the spatial relationships between objects in cluttered environments. To further enhance the perception capacity of position information of the objects, the authors associate the change in the object isolation metric and clutter metric in cluttered environment before and after performing the action with reward function. A series of experiments are carried out in simulation and real‐world which indicate that the method improves sample efficiency, task completion rate, grasping success rate and action efficiency compared to state‐of‐the‐art end‐to‐end methods. Noted that the authors' system can be robustly applied to real‐world use and extended to novel objects. Supplementary material is available at https://youtu.be/NhG\_k5v3NnM}{https://youtu.be/NhG\_k5v3NnM. [ABSTRACT FROM AUTHOR]

As deep learning evolves, neural network structures become increasingly sophisticated, bringing a series of new optimisation challenges. For example, deep neural networks (DNNs) are vulnerable to a variety of attacks. Training neural networks under privacy constraints is a method to alleviate privacy leakage, and one way to do this is to add noise to the gradient. However, the existing optimisers suffer from weak convergence in the presence of increased noise during training, which leads to a low robustness of the optimiser. To stabilise and improve the convergence of DNNs, the authors propose a neural dynamics (ND) optimiser, which is inspired by the zeroing neural dynamics originated from zeroing neural networks. The authors first analyse the relationship between DNNs and control systems. Then, the authors construct the ND optimiser to update network parameters. Moreover, the proposed ND optimiser alleviates the non‐convergence problem that may be suffered by adding noise to the gradient from different scenarios. Furthermore, experiments are conducted on different neural network structures, including ResNet18, ResNet34, Inception‐v3, MobileNet, and long and short‐term memory network. Comparative results using CIFAR, YouTube Faces, and R8 datasets demonstrate that the ND optimiser improves the accuracy and stability of DNNs under noise‐free and noise‐polluted conditions. The source code is publicly available at https://github.com/LongJin‐lab/ND. [ABSTRACT FROM AUTHOR]

Development of deep learning models to evaluate structural brain changes caused by cognitive impairment in MRI scans holds significant translational value. The efficacy of these models often encounters challenges due to variabilities arising from different data generation protocols, imaging equipment, radiological artifacts, and shifts in demographic distributions. Domain generalization (DG) techniques show promise in addressing these challenges by enabling the model to learn from one or more source domains and apply this knowledge to new, unseen target domains. Here we present a framework that utilizes model interpretability to enhance the generalizability of classification models across various cohorts. We used MRI scans and clinical diagnoses from four independent cohorts: Alzheimer's Disease Neuroimaging Initiative (ADNI, n = 1821), the Framingham Heart Study (FHS, n = 304), the Australian Imaging Biomarkers & Lifestyle Study of Ageing (AIBL, n = 661), and the National Alzheimer's Coordinating Center (NACC, n = 4647). With this data, we trained a deep neural network to focus on areas of the brain identified as relevant to the disease for model training. Our approach involved training a classifier to differentiate between structural neurodegeneration in individuals with normal cognition (NC), mild cognitive impairment (MCI), and dementia due to Alzheimer's disease (AD). This was achieved by aligning class‐wise attention with a unified visual saliency prior, which was computed offline for each class using all the training data. Our method not only competes with state‐of‐the‐art approaches but also shows improved correlation with postmortem histology. This alignment with the gold standard evidence is a significant step towards validating the effectiveness of DG frameworks, paving the way for their broader application in the field. [ABSTRACT FROM AUTHOR]

Background: Recommendations for cosmetics are gaining popularity, but they are not being made with consideration of the analysis of cosmetic ingredients, which customers consider important when selecting cosmetics. Aims: This article aims to propose a method for estimating the efficacy of cosmetics based on their ingredients and introduces a system that recommends personalized products for consumers, combined with AI skin analysis. Methods: We constructed a deep neural network architecture to analyze sequentially arranged cosmetic ingredients in the product and incorporated skin analysis models to get the precise skin status of users from frontal face images. Our recommendation system makes decisions based on the results optimized for the individual. Results: Our cosmetic recommendation system has shown its effectiveness through reliable evaluation metrics, and numerous examples have demonstrated its ability to make reasonable recommendations for various skin problems. Conclusion: The result shows that deep learning methods can be used to predict the effects of products based on their cosmetic ingredients and are available for use in personalized cosmetic recommendations. [ABSTRACT FROM AUTHOR]

Background and Aims: Anesthetic drugs play a vital role during surgery, however, due to individual differences and complex physiological mechanisms, the prediction of anesthetic drug dosage has always been a challenging problem. In this study, we propose a model for predicting the dosage of anesthetic drugs based on deep learning to help anesthesiologists better control their dosage during surgical procedures. Methods: We design a model based on the artificial neural network to predict the dosage of preoperative anesthetic, and use the SELU activation function and the loss function for weighted regularization to solve the problem of unbalanced sample. Moreover, we design a CNN‐based model for the prior extraction of intraoperative features by using a 7 × 1 convolution kernel to enhance the receptive field, and combine maximum pooling and average pooling to extract key features while eliminating noise. A predictive model based on the LSTM network is designed to predict the intraoperative dosage of the anesthetic, and the bidirectional propagation‐based LSTM network is used to improve the ability to learn the trend of changes in the physiological states of the patient during surgery. An attention module is added before the connection layer to appropriately attend to areas containing prominent features. Results: The results of experiments showed that the proposed method reduced values of the MAPE to 15.83% and 12.25% compared with the traditional method in predictions of the preoperative and intraoperative doses of the anesthetic, respectively, and increased the values of R2 ${R}^{2}$ to 0.887 and 0.915, respectively. Conclusion: The intelligent anesthesia prediction algorithm designed in this study can effectively predict the dosage of anesthetic drugs needed by patients, assist clinical judgment of anesthetic drug dose, and assist the anesthesiologists to ensure the smooth progress of the operation. [ABSTRACT FROM AUTHOR]

Background and Aim: Staging liver fibrosis is important, and liver biopsy is the gold standard diagnostic tool. We aim to design and evaluate an artificial neural network (ANN) method by taking advantage of the Teaching Learning‐Based Optimization (TLBO) algorithm for the prediction of liver fibrosis stage in blood donors and hepatitis C patients. Methods: We propose a method based on a selection of machine learning classification methods including multilayer perceptron (MLP) neural network, Naive Bayesian (NB), decision tree, and deep learning. Initially, the synthetic minority oversampling technique (SMOTE) is performed to address the imbalance in the dataset. Afterward, the integration of MLP and TLBO is implemented. Results: We propose a novel algorithm that reduces the number of required patient features to seven inputs. The accuracy of MLP using 12 features is 0.903, while that of the proposed MLP with TLBO is 0.891. Besides, the diagnostic accuracy of all methods, except the model designed with the Bayesian network, increases when the SMOTE balancer is applied. Conclusion: The decision tree‐based deep learning methods show the highest levels of accuracy with 12 features. Interestingly, with the use of TLBO and seven features, MLP reached an accuracy rate of 0.891, which is quite satisfactory when compared with those of similar studies. The proposed model provides high diagnostic accuracy, while reducing the required number of properties from the samples. The results of our study show that the recruited algorithm of our study is more straightforward, with a smaller number of required properties and similar accuracy. [ABSTRACT FROM AUTHOR]

Objectives: This study aimed to analyze publications on artificial intelligence (AI) for falls in older people from a bibliometric perspective. Methods: The Web of Science database was searched for titles of English‐language articles containing the words "artificial intelligence," "deep learning," "machine learning," "natural language processing,", "neural artificial network," "fall," "geriatric," "elderly," "aging," "older," and "old age." An R‐based application (Biblioshiny for bibliometrics) and VOSviewer software were used for analysis. Results: Thirty‐seven English articles published between 2018 and 2024 were included. The year 2023 is the year with the most publications with 16 articles. The most productive research field was "Engineering Electrical Electronic" with seven articles. The most productive country was the United States, followed by China. The most common words were "injuries," "people," and "risk factors." Conclusion: Publications on AI and falls in the elderly are both few in number and the number of publications has increased in recent years. Future research should include relevant analyses in scientific databases, such as Scopus and PubMed. [ABSTRACT FROM AUTHOR]

Background: Hair and scalp disorders present a significant challenge in dermatology due to their clinical diversity and overlapping symptoms, often leading to misdiagnoses. Traditional diagnostic methods rely heavily on clinical expertise and are limited by subjectivity and accessibility, necessitating more advanced and accessible diagnostic tools. Artificial intelligence (AI) and deep learning offer a promising solution for more accurate and efficient diagnosis. Methods: The research employs a modified Xception model incorporating ReLU activation, dense layers, global average pooling, regularization and dropout layers. This deep learning approach is evaluated against existing models like VGG19, Inception, ResNet, and DenseNet for its efficacy in accurately diagnosing various hair and scalp disorders. Results: The model achieved a 92% accuracy rate, significantly outperforming the comparative models, with accuracies ranging from 50% to 80%. Explainable AI techniques like Gradient‐weighted Class Activation Mapping (Grad‐CAM) and Saliency Map provided deeper insights into the model's decision‐making process. Conclusion: This study emphasizes the potential of AI in dermatology, particularly in accurately diagnosing hair and scalp disorders. The superior accuracy and interpretability of the model represents a significant advancement in dermatological diagnostics, promising more reliable and accessible diagnostic methods. [ABSTRACT FROM AUTHOR]

With the continuous development of wireless communication and artificial intelligence technology, Internet of Things (IoT) technology has made great progress. Deep learning methods are currently used in IoT technology, but deep neural networks (DNNs) are notoriously susceptible to adversarial examples, and subtle pixel changes to images can result in incorrect recognition results from DNNs. In the real‐world application, the patches generated by the recent physical attack methods are larger or less realistic and easily detectable. To address this problem, a Generative Adversarial Network based on Visual attention model and Style transfer network (GAN‐VS) is proposed, which reduces the patch area and makes the patch more natural and less noticeable. A visual attention model combined with generative adversarial network is introduced to detect the critical regions of image recognition, and only generate patches within the critical regions to reduce patch area and improve attack efficiency. For any type of seed patch, an adversarial patch can be generated with a high degree of stylistic and content similarity to the attacked image by generative adversarial network and style transfer network. Experimental evaluation shows that the proposed GAN‐VS has good camouflage and outperforms state‐of‐the‐art adversarial patch attack methods. [ABSTRACT FROM AUTHOR]

The automatic segmentation of brain tumours is a critical task in patient disease management. It can help specialists easily identify the location, size, and type of tumour to make the best decisions regarding the patients' treatment process. Recently, deep learning methods with attention mechanism helped increase the performance of segmentation models. The proposed method consists of two main parts: the first part leverages a deep neural network architecture for biggest tumour detection (BTD) and in the second part, ResNet152V2 makes it possible to segment the image with the attention block and the extraction of local and global features. The custom attention block is used to consider the most important parts in the slices, emphasizing on related information for segmentation. The results show that the proposed method achieves average Dice scores of 0.81, 0.87 and 0.91 for enhancing core, tumour core and whole tumour on BraTS2020 dataset, respectively. Compared with other segmentation approaches, this method achieves better performance on tumour core and whole tumour. Further comparisons on BraTS2018 and BraTS2017 validation datasets show that this method outperforms other models based on Dice score and Hausdorff criterion. [ABSTRACT FROM AUTHOR]

Intelligent motor fault diagnosis in industrial applications requires identifying key characteristics to differentiate various fault types effectively. Solely relying on statistical features cannot guarantee high classification accuracy, while complex feature extraction techniques can pose challenges for industry practitioners. Conversely, advanced feature extraction may not ensure that the model effectively learns these features for classification. A feature fusion approach that combines statistical and deep learning features to address these challenges is proposed. Since statistical features form the foundation for general feature extraction, statistical and deep learning features are combined using Extreme Gradient Boosting (XGBoost) algorithm with Particle Swarm Optimization (PSO). The PSO algorithm automates parameter tuning for XGBoost. A deep neural network (DNN) adaptively extracts hidden features, improving bearing fault classification precision using t‐SNE representation. Results successfully prove the DNN's ability to classify diverse motor faults using deep learning features. Thus, integrating statistical features with XGBoost further enhances DNN's performance. To ensure robustness, the proposed method has been compared with different motor fault classification methods and validated across different motor fault datasets, showcasing improved classification accuracy and robust performance, even amidst varying noise levels. This approach represents a promising advancement in intelligent fault diagnosis within industrial contexts. [ABSTRACT FROM AUTHOR]

Traditional analog and digital meters are being substantially replaced with technological advances and the Internet of Things (IoT) introduction. The smart meter is highly preferred for accessing real‐time consumption, tariff calculation, and remote system control. These smart meters also prevent the majority from bypassing theft. Despite its intelligence, it cannot be 100% secure. An error in the readings can be caused by hacking or damage to meter components making the utility companies suffer significant losses. Based on past and future energy consumption data predictions on the consumer side, various methods are used in the proposed work to identify theft or anomalies in smart meter readings. Forecast‐based detection proved to be the most effective and accurate method. The primary and secondary decision models, which employ a variety of statical analyses to identify system anomalies, serve as the foundation for the energy consumption that follows the forecasting. Past 24‐h data is needed for forecasting, which is passed through different statical calculations such as RMSE, simple moving average, and Absolute Percentage Error to conclude detecting the normal values. Long short‐term memory gives high accuracy of 97% for forecasting and detecting abnormalities. [ABSTRACT FROM AUTHOR]

The involvement of hydrogen energy systems has been recognised as a promising way to mitigate climate problems. As a kind of efficient multi‐energy complementary system, the hydropower‐photovoltaic‐hydrogen (HPH) system could be an ideal approach to combining hydrogen with an installed renewable energy system to improve the flexibility of energy management and reduce power curtailment. However, the intra‐day scheduling of HPH system brings challenges due to the time‐related nonlinear hydropower generation process, the complex energy conversion process and the uncertain natural resource supply. Faced with these challenges, an improved deep deterministic policy gradient (DDPG)‐based data‐driven scheduling algorithm is proposed. In contrast to the prevalent DDPG, two sets of actor‐critic networks are properly designed based on prior knowledge‐based deep neural networks for the considered complex uncertain system to search for near‐optimal policies and approximate actor‐value functions. In addition, customized reward functions are proposed with the consideration of interactions among different energy supplies, which helps to improve convergence speed and stability. Finally, the case study results demonstrate that the proposed system model and the optimal energy management strategy based on the improved DDPG algorithm can guide the electricity‐hydrogen system to achieve rapid response and more reasonable energy management. [ABSTRACT FROM AUTHOR]

By automatically learning the priors embedded in images with powerful modelling capabilities, deep learning‐based algorithms have recently made considerable progress in reconstructing the high‐resolution hyperspectral (HR‐HS) image. With previously collected large‐amount of external data, these methods are intuitively realised under the full supervision of the ground‐truth data. Thus, the database construction in merging the low‐resolution (LR) HS (LR‐HS) and HR multispectral (MS) or RGB image research paradigm, commonly named as HSI SR, requires collecting corresponding training triplets: HR‐MS (RGB), LR‐HS and HR‐HS image simultaneously, and often faces difficulties in reality. The learned models with the training datasets collected simultaneously under controlled conditions may significantly degrade the HSI super‐resolved performance to the real images captured under diverse environments. To handle the above‐mentioned limitations, the authors propose to leverage the deep internal and self‐supervised learning to solve the HSI SR problem. The authors advocate that it is possible to train a specific CNN model at test time, called as deep internal learning (DIL), by on‐line preparing the training triplet samples from the observed LR‐HS/HR‐MS (or RGB) images and the down‐sampled LR‐HS version. However, the number of the training triplets extracted solely from the transformed data of the observation itself is extremely few particularly for the HSI SR tasks with large spatial upscale factors, which would result in limited reconstruction performance. To solve this problem, the authors further exploit deep self‐supervised learning (DSL) by considering the observations as the unlabelled training samples. Specifically, the degradation modules inside the network were elaborated to realise the spatial and spectral down‐sampling procedures for transforming the generated HR‐HS estimation to the high‐resolution RGB/LR‐HS approximation, and then the reconstruction errors of the observations were formulated for measuring the network modelling performance. By consolidating the DIL and DSL into a unified deep framework, the authors construct a more robust HSI SR method without any prior training and have great potential of flexible adaptation to different settings per observation. To verify the effectiveness of the proposed approach, extensive experiments have been conducted on two benchmark HS datasets, including the CAVE and Harvard datasets, and demonstrate the great performance gain of the proposed method over the state‐of‐the‐art methods. [ABSTRACT FROM AUTHOR]

In recent times, an image enhancement approach, which learns the global transformation function using deep neural networks, has gained attention. However, many existing methods based on this approach have a limitation: their transformation functions are too simple to imitate complex colour transformations between low‐quality images and manually retouched high‐quality images. In order to address this limitation, a simple yet effective approach for image enhancement is proposed. The proposed algorithm based on the channel‐wise intensity transformation is designed. However, this transformation is applied to the learnt embedding space instead of specific colour spaces and then return enhanced features to colours. To this end, the authors define the continuous intensity transformation (CIT) to describe the mapping between input and output intensities on the embedding space. Then, the enhancement network is developed, which produces multi‐scale feature maps from input images, derives the set of transformation functions, and performs the CIT to obtain enhanced images. Extensive experiments on the MIT‐Adobe 5K dataset demonstrate that the authors' approach improves the performance of conventional intensity transforms on colour space metrics. Specifically, the authors achieved a 3.8% improvement in peak signal‐to‐noise ratio, a 1.8% improvement in structual similarity index measure, and a 27.5% improvement in learned perceptual image patch similarity. Also, the authors' algorithm outperforms state‐of‐the‐art alternatives on three image enhancement datasets: MIT‐Adobe 5K, Low‐Light, and Google HDR+. [ABSTRACT FROM AUTHOR]

Although previous studies have made some clear leap in learning latent dynamics from high‐dimensional representations, the performances in terms of accuracy and inference time of long‐term model prediction still need to be improved. In this study, a deep convolutional network based on the Koopman operator (CKNet) is proposed to model non‐linear systems with pixel‐level measurements for long‐term prediction. CKNet adopts an autoencoder network architecture, consisting of an encoder to generate latent states and a linear dynamical model (i.e., the Koopman operator) which evolves in the latent state space spanned by the encoder. The decoder is used to recover images from latent states. According to a multi‐step ahead prediction loss function, the system matrices for approximating the Koopman operator are trained synchronously with the autoencoder in a mini‐batch manner. In this manner, gradients can be synchronously transmitted to both the system matrices and the autoencoder to help the encoder self‐adaptively tune the latent state space in the training process, and the resulting model is time‐invariant in the latent space. Therefore, the proposed CKNet has the advantages of less inference time and high accuracy for long‐term prediction. Experiments are performed on OpenAI Gym and Mujoco environments, including two and four non‐linear forced dynamical systems with continuous action spaces. The experimental results show that CKNet has strong long‐term prediction capabilities with sufficient precision. [ABSTRACT FROM AUTHOR]

The extraction of road networks is a critical activity in contemporary transportation networks. Deep neural networks have recently demonstrated excellent performance in the field of road segmentation. However, most of the convolutional neural network (CNN) based architectures could not verify their effectiveness in remote sensing images due to a smaller ratio of the targeted pixels, simple design, and fewer layers. In this study, a practical approach is assessed for road segmentation. The investigation was begun with basic encoder–decoder based segmentation models. Different state‐of‐the‐art segmentation models like U‐Net, V‐Net, ResUNet and SegNet were used for road network detection experiments in this research. A robust model named DeepRoadNet, a more complicated alternative, is proposed by utilizing a pre‐trained EfficientNetB7 architecture in the encoder and residual blocks as the decoder which mostly resembles the U‐Net segmentation process. The proposed model has been trained, validated as well as tested using the high‐resolution aerial image datasets and yielded good segmentation results with a mean intersection over union (mIoU) of 76%, a mean dice coefficient (mDC) of 73.18%, and an accuracy of 97.64% using Massachusetts road dataset. The proposed DeepRoadNet architecture overcomes the issues of lower mIoU, lower mDC, limited flexibility and interpretability already faced by existing models in the road segmentation field. The code is available at https://github.com/Imteaz1998/DeepRoadNet. [ABSTRACT FROM AUTHOR]

Indoor positioning system (IPS) technologies have widespread applications in logistics, intelligent manufacturing, healthcare monitoring, etc. The recently released Bluetooth low‐energy (BLE) 5.1 specification enables in‐phase and quadrature‐phase (I/Q) data measurements. It allows angle of arrival estimation and becomes a natural choice for IPS implementation. Conventional BLE 5.1 IPSs use multiple anchors to provide massive redundancy to improve system robustness. It however demands effective approaches to leverage redundancy. Besides, interference due to various environmental factors can introduce severe errors to I/Q data and affect positioning accuracy. Facing these challenges, herein, a novel deep learning‐based multianchor BLE 5.1 IPS is proposed. The system aggregates measurements from multiple anchors and makes them available at regular time steps. Then, a novel attentional filtering network tailored to infer high‐quality I/Q sample data is developed and a spatial regularization loss incorporating spatial location relationships to strengthen the feature embedding discrimination is proposed. Two multianchor BLE 5.1 I/Q sample datasets are developed and released for public download. Numerical experiments are carried out to compare the proposed method with previous BLE 5.1 IPS methods and methods utilizing other radio frequency data. Results indicate that the proposed method consistently achieves submeter accuracy and significantly outperforms the state‐of‐the‐art approaches. [ABSTRACT FROM AUTHOR]

There can be numerous electronic components on a given PCB, making the task of visual inspection to detect defects very time‐consuming and prone to error, especially at scale. There has thus been significant interest in automatic PCB component detection, particularly leveraging deep learning. While deep neural networks are able to perform such detection with greater accuracy, these networks typically require high computational resources, limiting their feasibility in real‐world use cases, which often involve high‐volume and high‐throughput detection with constrained edge computing resource availability. To bridge this gap between performance and resource requirements, PCBDet, an attention condenser network design that provides state‐of‐the‐art inference throughput while achieving superior PCB component detection performance compared to other state‐of‐the‐art efficient architecture designs, is introduced. Experimental results show that PCBDet can achieve up to 2× inference speed‐up on an ARM Cortex A72 processor when compared to an EfficientNet‐based design while achieving ∼2–4% higher mAP on the FICS‐PCB benchmark dataset. [ABSTRACT FROM AUTHOR]