Search

The ionospheric foF2 prediction model GWO-ALSTM is proposed in this paper, which combines the Gray Wolf Optimization (GWO) algorithm and the attention mechanism based on the long short-term memory (LSTM) model. The datasets used in this model were obtained from the oblique ionosondes of the China Ground-based Seismo-ionospheric Monitoring Network. The data from 2010 to 2012 was used for training, and the data of 2013 was used for verification. The input parameters of the model include local time, day number, sunspot number, F10.7 solar flux, Ap index, Dst index, and foF2 at the previous moment, and the output parameter is foF2 at the current moment. The IRI2016 model, GABP model, LSTM model and GWO-ALSTM model were compared with the observed results for analysis. The results show that the GWO-ALSTM model is superior to the LSTM model, the GABP model and the IRI2016 model. Meanwhile, the comparative analyses of the diurnal variations of foF2 show that the curve of the GWO-ALSTM model fits the observed values more closely than the curve of any other model under the geomagnetic quiet and storm conditions. [ABSTRACT FROM AUTHOR]

The problem of global warming is getting more serious, and vehicle emission is the main cause. In recent years, the number of locomotives in China has been increasing at a rate of more than 20% per year, and the problem of automobile pollution is becoming more serious. The transportation industry is the main source of fossil fuel combustion and environmental pollution. Therefore, in this paper, we propose a multi-view deep neural network (MVDNN) to analyze the key factors affecting the fuel consumption of automobiles. The experiments show that the introduction of human input improves the prediction accuracy and the root mean square error (RMSE) achieves 0.993. In addition, this paper also finds that for drivers, driving habits, driving frequency, and safety awareness are the most important factors affecting the fuel consumption of vehicles by combining Lasso regression with MVDNN. Finally, by comparing the prediction accuracy of different experiments, relevant policy suggestions are made. [ABSTRACT FROM AUTHOR]

After collecting Raman spectral data of 12 compounds from commonly used Uyghur medicinal materials, the mixture data are synthesized using computer simulation methods. An accurate identification method for Uyghur medicine components based on Raman spectroscopy and multi-label deep learning is constructed, and the performance of the model Mix2Com is evaluated using several metrics such as accuracy, precision, recall value, and F1 score. [Display omitted] • Existing methods identifying Uyghur medicinal components aren't fit for large samples. • A new method based on Raman spectroscopy and multi-label learning is proposed. • Macro-parallel identification of components is achieved. • The model Mix2Com is constructed with LSTM and Attention introduced. • Mix2Com achieves 90.76% accuracy, better than XGBoost and DeepRaman. Uyghur medicine is one of the four major ethnic medicines in China and is a component of traditional Chinese medicine. The intrinsic quality of Uyghur medicinal materials will directly affect the clinical efficacy of Uyghur medicinal preparations. However, in recent years, problems such as adulteration of Uyghur medicinal materials and foreign bodies with the same name still exist, so it is necessary to strengthen the quality control of Uyghur medicines to guarantee Uyghur medicinal efficacy. Identifying the components of Uyghur medicines can clarify the types of medicinal materials used, is a crucial step to realizing the quality control of Uyghur medicines, and is also an important step in screening the effective components of Uyghur medicines. Currently, the method of identifying the components of Uyghur medicines relies on manual detection, which has the problems of high toxicity of the unfolding agent, poor stability, high cost, low efficiency, etc. Therefore, this paper proposes a method based on Raman spectroscopy and multi-label deep learning model to construct a model Mix2Com for accurate identification of Uyghur medicine components. The experiments use computer-simulated mixtures as the dataset, introduce the Long Short-Term Memory Model (LSTM) and Attention mechanism to encode the Raman spectral data, use multiple parallel networks for decoding, and ultimately realize the macro parallel prediction of medicine components. The results show that the model is trained to achieve 90.76% accuracy, 99.41% precision, 95.42% recall value and 97.37% F1 score. Compared to the traditional XGBoost model, the method proposed in the experiment improves the accuracy by 49% and the recall value by 18%; compared with the DeepRaman model, the accuracy is improved by 9% and the recall value is improved by 14%. The method proposed in this paper provides a new solution for the accurate identification of Uyghur medicinal components. It helps to improve the quality standard of Uyghur medicinal materials, advance the research on screening of effective chemical components of Uyghur medicines and their action mechanisms, and then promote the modernization and development of Uyghur medicine. [ABSTRACT FROM AUTHOR]

Intertidal mudflats are an important component of the coastal geomorphological system at the interface between ocean and land. Accurate and up-to-date mapping of intertidal topography at high spatial resolution, and tracking of its changes over time, are essential for coastal habitat protection, sustainable management and vulnerability analysis. Compared with ground-based or airborne terrain mapping, the satellite-based waterline method is more cost-effective for constructing large-scale intertidal topography. However, the accuracy of the waterline method is affected by the extraction of waterlines and the calibration of waterline height. The blurred boundary between turbid water and mudflats in the tide-dominated estuary brings enormous challenges in accurate waterline extraction, and the errors in estuarine water level simulations prevent the direct calibration of waterline heights. To address these issues, this paper developed a novel deep learning method using a parallel self-attention mechanism and boundary-focused hybrid loss to extract turbid estuarine waterlines accurately from dense Sentinel-2 time series. UAV photogrammetric surveys were employed to calibrate waterline heights rather than the simulated water levels, such that the error propagation is constrained effectively. Annual intertidal topographic maps of the Yangtze estuary in China were generated from 2020 to 2022 using the optimized waterline method. Experimental results demonstrate that the proposed deep learning method could achieve excellent performance in land and water segmentation in time-varying tidal environments, with better generalization capability compared with benchmark U-Net, U-Net++ and U-Net+++ models. The comparison between the generated topography and UAV photogrammetric observations resulted in an RMSE of 13 cm, indicating the effectiveness of the optimized waterline method in monitoring morphological changes in estuarine mudflats. The generated topographic maps successfully identified hotspots of mudflat erosion and deposition. Specifically, the mudflats connected to the land predominantly experienced deposition of 10–20 cm over the two-year period, whereas the offshore sandbars exhibited instability and significant erosion of 20–60 cm during the same period. These topographic maps serve as valuable datasets for providing scientific baseline information to support coastal management decisions. [ABSTRACT FROM AUTHOR]

Groundwater level (GWL) prediction is important for ecological protection and resource utilization; it helps in formulating policies for artificial groundwater recharge, modifying the number of extraction wells, etc., and can support sustainable human development as well as inform water resource management decisions. However, climate change, anthropogenic impacts, and the complex coupling between surface water and groundwater increase the difficulty of predicting groundwater levels. The model proposed in this paper combines external data as well as multiple models. The method leverages long and short-term memory (LSTM) and convolutional neural network (CNN) models, combined with secondary modal decomposition and slime mould algorithm (SMA), together with an adaptive weight module (AWM). The study applies this method to predict GWL for three different hydrological conditions in China, specifically for the Jinan Baotu Spring, Heihu Spring, and Zhongtianshe watershed of Taihu Lake. A comparison of metrics such as mean absolute error and Nash efficiency coefficient for single and hybrid models shows that the model in this paper is more advantageous than the single model and other hybrid models. The interpretability of the model is enhanced by SHAP values that demonstrate the degree of contribution of the input variables. This paper uses SHAP analyses to identify the key drivers affecting groundwater levels. These factors must be detected in order to develop groundwater resource protection measures. [Display omitted] • Multivariate fusion data including hydrology and meteorology are used as model input. • A secondary modal decomposition module for historical groundwater level data was utilized. • The neural network hyperparameters are optimized using the slime mould algorithm. • Aggregate subsets of prediction with adaptive modules rather than linear summation. • The interpretable SHAP model measures the degree of influence of external variables. [ABSTRACT FROM AUTHOR]

In this paper, we propose a new deformable convolutional network with attention mechanism to deal with the Chinese character of various dynasties. These ancient Chinese characters are hieroglyph with special spatial structure and gradually evolving into modern Chinese characters. However, because of the intrinsic limitations of traditional convolutional networks in the model, the geometric structure of the convolutional module is fixed. Therefore, we build a new architecture called attention-based deformable networks for this task. Besides, we construct a new dataset called the Chinese characters from Various Dynasties Dataset (CCDD), which includes the evolution of Chinese characters between major dynasties. Deformable convolution based on attention mechanism can obtain those more important offsets, resulting in better performance than the original deformable convolution. In the experimental stage, extensive experiments validate the performance of our approach. Firstly, the effectiveness of this network was verified on two text datasets, Mnist(99.38%) and notMnist(98.22%) both comparing with several modules. Finally, the best performance was achieved on the CCDD dataset(91.50%) compared to other classical deep learning networks. This is the first paper to apply deep learning methods to recognize and classify the characters of major dynasties in China. And archaeological workers need to be able to recognize Chinese characters from various periods in China, and our algorithm can assist them in completing this task to a certain extent. It can also help cultural relic enthusiasts roughly determine the age of cultural relics. [ABSTRACT FROM AUTHOR]

The complex and varied urban road conditions have always been a difficult and a pivotal component in the study of driverless technology especially at intersections. In China, it is necessary for driverless cars to understand the gestures of traffic police. To identify the traffic police gesture at the intersection, the key step is to detect the traffic police at the intersection. At present, the research on traffic police detection is still in its infancy, there exists common problems such as slow detection speed and other real time problems in this method, and there is not a standardized traffic police data set ether. For the real time problems, this paper introduces the attention mechanism, and proposes a new real-time detection method of traffic police based on attention mechanism. The method proposed in this paper has strong robustness and can quickly complete the target detection task. For the data set problem, this paper analyzes and discloses similar data sets published through research. In the meanwhile, this paper collects 24,530 video data on the actual road, and it extracts and saves 12,000 pictures containing traffic police at frame rate of 30FPS as traffic police detection training and validated data set. [ABSTRACT FROM AUTHOR]

Mammography is an important screening criterion for breast cancer, one of the major diseases causing numerous deaths among female patients. Meanwhile, manual diagnosis of mammography is a time-consuming and labor-consuming job. Mammogram classification based on deep learning plays a vital role in computer-aided diagnosis (CAD) systems to mitigate the pressure on physicians. This paper proposes a learning-based multi-view mammogram classification model that captures long-distance dependence and extracts features of multiple receptive fields. Our model considers global and local features of mammography images using Transformer for global features and the proposed multiplex convolutions module for local features. We evaluate our proposed method on a dataset of mammography images obtained from a hospital in China. The proposed method achieves 90.57% accuracy and 94.86% AUC in benign or malignant classification tasks and outperforms other advanced methods for mammogram classification. It is worth noting that the proposed method only requires image-level labels and acts on the whole raw mammogram, which has clinical significance. [ABSTRACT FROM AUTHOR]

The Aerial Building Machine (ABM) is a complex construction equipment used in high-rise building construction, facing challenges due to environmental uncertainties. This paper introduces a multi-condition generative design framework to improve ABM's visualization, operability, and intelligence. It seamlessly integrates real-time data between geometric and physical models and employs an ensemble deep learning model for objective value prediction, using a snapshot strategy. Combining structural reliability concepts with Latin hypercube sampling-based stochastic optimization, an optimal design scheme is obtained for uncertain loads. An ABM case study in China illustrates the approach's feasibility, showing it meets reliability requirements across different conditions and achieves significant improvements (16.59% under normal conditions and 16.91% under extreme wind conditions). Additionally, ensemble deep learning outperforms existing methods for ABM structural reliability estimation. Identifying optimal designs and evaluation options, this paper contributes a multi-condition optimization approach for enhanced structural reliability and establishes an efficient generative design workflow and system for exploring a vast solution space. • A multi-condition generative design framework for ABM is proposed. • The ensemble deep learning approach is employed to predict the structural response precisely. • The load uncertainties are considered according to the developed LHS-based SO approach. • A real aerial building machine in China is taken as a case study for illustration. • The proposed approach shows great performance in the structural optimization design. [ABSTRACT FROM AUTHOR]

• The training time of Long Short-Term Memory is reduced by changing network structure. • A new hybrid model is proposed for wind speed forecasting. • Four experiments are designed to verify the performance and reliability of the model. Obtaining accurate wind speed forecast result plays a decisive role in ensuring the reliable operation of the power system integrated with large-scale wind power. Deep learning methods are increasingly being used to predict wind speed, which have relatively high prediction accuracy but more time-consuming training processes. The purposes of this study are further to improve the prediction accuracy of the wind speed and reduce the training time of the deep learning method. In this paper, a novel hybrid model, New Cell Update Long Short-Term Memory combined with Empirical Wavelet Transform, is proposed to increase the prediction accuracy in shorter training time. At first, the original wind speed sequence is preprocessed into a series of sub-sequence by the empirical wavelet decomposition. Then each sub-sequence is trained by New Cell Update Long Short-Term Memory which is proposed by this paper respectively and the sum of each sub-sequence is treated as the final prediction results. In order to verify the performance of the proposed model, different decomposition methods and different prediction methods are compared on the four actual wind speed prediction cases in the Inner Mongolia, China from prediction accuracy and training time. The results demonstrate that: (1) New Cell Update Long Short-Term Memory network has slightly higher prediction accuracy and shorter training time than Long Short-Term Memory network. (2) The prediction accuracy of the model is significantly improved after empirical wavelet decomposition. Therefore, New Cell Update Long Short-Term Memory network combined with empirical wavelet decomposition is a competitive wind speed prediction method compared to the existing state-of-the-art approach. [ABSTRACT FROM AUTHOR]

Accurately predicting wind and photovoltaic power is one of the keys to improving the economy of wind-solar complementary power generation system, reducing scheduling costs and no-load losses, and ensuring grid stability. However, the natural properties of energy result in complex fluctuations in their corresponding power sequences, making accurate predictions difficult. Therefore, this paper proposes an intelligent prediction system that combines decomposition algorithms and deep learning for ultra-short-term prediction of wind and photovoltaic power. First, an improved decomposition algorithm is proposed, based on fuzzy entropy's property that its value increases with the increase of sequence uncertainty, particle swarm optimization (PSO) is used to search for the optimal parameter combinations of variational modal decomposition (VMD), so that it can automatically adjust the parameters for energy data with different characteristics to reduce the human error. Then, a convolutional neural network (CNN) architecture that balances operational efficiency and prediction performance is constructed, and the hyperparameters of the CNN are optimized using the wild horse optimization algorithm (WHO) to improve the stability and accuracy of the prediction model. In this paper, real data from wind power plants and photovoltaic power plants in China are used as experimental objects, and experiments are carried out in three aspects, namely, benchmark model selection, decomposition algorithm comparison and combined model comparison. The results show that selecting CNN as the benchmark model is a good choice; the improved VMD has better decomposition performance than other state-of-the-art decomposition algorithms. The system proposed in this paper is highly generalizable and adaptive, and its prediction performance and accuracy greatly outperform that of other comparative models, with prediction accuracies improved by 72% and 79%, respectively, compared to a single CNN model. • The hyperparameters of VMD are determined by using PSO based on fuzzy entropy. • Optimize convolutional neural network using the wild horse optimization algorithm. • The intelligent prediction system can accurately predict wind power and PV power. • Experiments based on power data from actual wind farms and PV plants. • A deep learning prediction method applied to wind and solar complementary systems. [ABSTRACT FROM AUTHOR]

Deep Learning is an effective tool to invert the underground electrical conductivity structure. In this study, we build a new 8-layer residual neural network (ResNet1D-8) for audio magnetotelluric (AMT) data inversion based on the deep learning theory. In terms of the network structure, the degradation of the model with the increase of depth is effectively avoided by adding shortcut connections. Meanwhile, adding a batch normalization layer greatly improves the model training speed and network generalization ability. This study uses the parallel technology to quickly generate millions of samples, which effectively reduces the computational time and provides a large number of high-quality samples for deep learning model training. Compared with the simulated annealing algorithm, the neural network model in this paper has the advantages of high reliability, short inversion time, and strong model generalization ability. Moreover, we add Gaussian noise to the data of testing samples, and inversion results show that the model has good robustness. The inversion test is carried out on the field measured AMT data set collected in the Dachaidan area of Qinghai Province, China. The results show that the ResNet1D-8 residual neural network model established in this paper can effectively invert the underground electrical structure. • A new residual neural network model (ResNet1D-8) for 1-D MT inversion is constructed. • A uniform stochastic model and parallel techniques are used to rapidly generate samples. • The generalizability and robustness of ResNet1D-8 are verified through numerical experiments. [ABSTRACT FROM AUTHOR]

To ensure the operational reliability of power systems, it is important for wind speed signal forecasting systems of wind turbines to be efficient, accurate and stable. This paper proposes a two-phase deep learning structure with network augmentation and pruning. By introducing the cross-correlation and quasi-convex optimization, a fractional quadratic programming problem and related convex optimization models are constructed to generate the augmented data for this proposed internal network; by pruning weakly correlated convolution channels, the redundant features of its external network are reduced. Furthermore, the closed-form solution of the convex optimization model is derived, which reduces the computational complexity considerably from O(n *log(2 N)) to O(n). The proposed approach has been extensively validated using the real data of the wind farm in China. The results of the numerical experiments demonstrate that the proposed method achieves the superior performance in the training flexibility, model accuracy, stability, and interpretability. [ABSTRACT FROM AUTHOR]

The use of robots is being widely considered to ensure the inheritance of calligraphy, which is a cultural art peculiar to China and Japan. Many researchers have attempted to generate motion command values to allow robots to inherit calligrapher skills. In this paper, we constructed a model that generates motion command values considering time-series positions and forces from human motion data and character images. However, since it is difficult to acquire a large amount of motion data due to physical restrictions, we compared the accuracy based on the number of training data. The performance of the model was evaluated with motion data generated using the image data for evaluation. We also showed that calligraphy is a delicate task and that using force data can improve results. • This research proposes a deep-learning-based skill transfer of a calligrapher. • The proposed model generates motion command from human motion data and images. • The accuracy based on the number of training data is compared and analyzed. • The performance was evaluated with motion data using the image data for evaluation. [ABSTRACT FROM AUTHOR]

The increase in air pollution has posed numerous new challenges for human society, making the exploration of an effective method for predicting air pollutant concentrations highly significant. The current research faces several primary challenges: the neglect of non-Euclidean characteristics of site distribution on data and the strong spatiotemporal dependencies in the dispersion process of pollutants. To address these issues, this paper constructs a spatiotemporal hybrid prediction model – the MGAtt-LSTM method – for predicting PM 2.5 concentrations, which employs the dynamic multi-graph attention module (MGAtt) to tackle spatial dependencies and Long Short-Term Memory networks (LSTM) to address temporal dependencies. Additionally, extensive experiments are conducted by using historical air pollutant monitoring data and meteorological data from the Beijing-Tianjin-Hebei region. The results demonstrate that the proposed MGAtt-LSTM model achieved superior performance in concentration prediction compared to existing benchmark models. • The neural network consists of multi-graph attention network and LSTM network. • Air pollution data in North China and meteorological data are used to forecast. • The uneven distribution of pollutant sites is considered to predict PM 2.5 concentrations precisely. • The use of multi-graph attention networks addresses the issue of traditional GCN methods relying on fixed graph structures. [ABSTRACT FROM AUTHOR]

Accurate PM 2.5 forecasting provides a possibility for establishing an early warning system to notify the public and take precautionary measures to prevent negative effects on ambient air quality and public health. Considering strong seasonal variation in meteorological conditions, in this paper, a seasonal stacked autoencoder model combining seasonal analysis and deep feature learning is proposed for forecasting the hourly PM 2.5 concentration, named DL-SSAE model. The original data are firstly decomposed into four seasonal subseries according to the Chinese calendar, and then the Kendall correlation coefficient method is employed to search inherent relationships between PM 2.5 concentrations and meteorological parameters within 1-h ahead for each seasonal time series. The inherent relationships of each seasonal subseries are finally extracted, learned, and modeled by different deep neural networks (stacked autoencoders for regression), and the hourly PM 2.5 forecasts are yielded. The addressed model is tested by the dataset collected from three environmental monitoring stations in Beijing, China. The results demonstrate that the proposed model outperforms all other considered models with/without seasonality consideration in this paper. [ABSTRACT FROM AUTHOR]

The purpose of video process detection is to identify key processes of interest and associated semantic categories, which is an important computer vision task for localizing key processes in space science experiments. To detect experimental processes in space electrostatic suspension material experiment on China's Space Station, we propose a Slow-and-Fast Dual-Stream Network. It extracts multi-scale temporal semantic information by designing a Slow-and-Fast Temporal Stream structure and effectively detects the experimental processes with various temporal scales. In addition, the Two Path Temporal Pyramid Network (TP2N) enhances the bi-direction flow among multi-scale features on the temporal dimension to alleviate boundary sensitivity. The validation experiments on our established dataset demonstrate the effectiveness of our proposed method with 62.43% mAP, outperforming the related state-of-the-art temporal action localization methods with nearly 7% mAP gain. This paper is the first application of deep learning in space electrostatic suspension material experiments on China's Space Station and further provides important technical support for the identification of key processes in space science experiments. [ABSTRACT FROM AUTHOR]

Accurate photovoltaic (PV) power forecast is crucial for carbon neutrality. Current researches on PV power forecast mainly focus on using temporal information from single PV station, and the spatial information in multiple PV power stations are often neglected. To address this problem, this paper introduces Moran index to verify the spatial autocorrelation of PV power for the first time, uses Granger causality test and transfer entropy to reveal the spatial information gain in PV power forecast for the first time, and proposes a novel spatial–temporal probabilistic PV forecast method using deep Graph Gated Recurrent Unit (GraphGRU) network-based spatial–temporal multi-task learning and Kernel Density Estimation (KDE). Deep GraphGRU combines the advantages of Graph Convolutional Network (GCN) in spatial feature extraction and the advantages of Gated Recurrent Unit (GRU) network in temporal feature extraction, and thus has strong ability to extract spatial–temporal information in historical data of multiple different PV power stations. Through GraphGRU, temporal dependency information extracted from historical data of multiple PV stations can promote each other to improve the forecast accuracy of each PV stations. KDE is used for estimating the joint probabilistic density function and giving the spatial–temporal probabilistic confidence interval of PV power. Experiments were performed in the five-year actual PV power data from 11 provinces of Belgium and the three-year solar irradiation data from 12 places in China to verify the superiorities of the proposed method. Comparison with conventional spatial–temporal and temporal forecast methods show that the proposed GraphGRU-based spatial–temporal forecast method can extract the spatial–temporal information from multiple PV power stations well and significantly outperform conventional temporal forecast methods. [ABSTRACT FROM AUTHOR]

Wind speed accurate forecasting plays an important role in preserving the stability of offshore wind power. Most of current predictions are based on a single wind node. It is difficult for this methods to capture wind high dimensional features and latent spatio-temporal dependencies. In this paper, we propose a general graph optimization neural network specifically for multi-node offshore wind speed prediction named spatio-temporal correlation graph neural network. The proposed model firstly employs graph convolution, which performs Laplace transforms instead of the traditional convolution, to capture the potential spatial dependencies from the nodes' relationship and historical time series better. The channel-wise attention makes the original concentrated weights within a region constructed by adjacency matrix disperse to all input nodes, which distinguishes the nodes' contributions and generates high dimensional spatial features. Long short-term memory is applied to extract temporal correlation from the high dimensional spatial features. The model has ability to excavate the full potential of spatial-temporal dependencies from multiple wind nodes to the utmost. Experiments select 120 wind nodes in the China sea for prediction. The results show that the proposed model can be very competitive with state-of-the-art methods and holds great performance on multi-node and multi-step wind speed forecasting. • A graph optimization model is proposed for multi-node offshore wind speed prediction. • The GCN module can capture multiple wind nodes relationships' spatial correlation. • Channel-wise attention distinguishes the nodes' contributions to spatial dimensions. • The parallel LSTMs is specifically adopted to further integrate time dependencies. • The proposed STGN can track actual wind speed tendencies and maintain stability. [ABSTRACT FROM AUTHOR]

This paper studied the carbon peak through the cross-analysis of low-carbon economics and deep learning. The STIRPAT model and ridge regression was used to distinguish and rank the importance of influencing factors to carbon emissions. In addition, an innovative GA-LSTM model was constructed for prediction. It combined the scenario analysis to explore the path of China's low-carbon development. The results showed that China's carbon emissions have been showing a growing trend, and among the many influencing factors, only the technological level had an inhibitory effect on carbon emissions. China's carbon peak will be reached around 2030 under all three scenarios of benchmark, steady growth, and green development, with peak values of 11.82, 11.94, and 11.64 billion tons, respectively. Meanwhile, there was a big difference in the rate of change in China's carbon emissions before and after the carbon peak. The rising-rate was faster before the carbon emission peak, while the decline rate was slower after the peak. This paper argued that China should start from the energy consumption structure and industrial structure to promote the development of emission reduction work and, at the same time, vigorously promote the effect of the technological level of emission reduction. • The cross-analysis of the low carbon economy and deep learning is used for research. • A GA-LSTM model was built to predict the carbon peak in China. • Scenario analysis is applied to explore China's scientific emission reduction path. • China's 2030 carbon peak target is likely to be met. [ABSTRACT FROM AUTHOR]

Greenhouses have revolutionized farming all over the world. To estimate vegetable yields, greenhouse mapping using high spatial resolution (HSR) remote sensing imagery is very important. Although automatic greenhouse mapping methods have been proposed, they are often applied in limited small-scale areas (i.e. a parcel, a city, or a province). Large-scale greenhouse mapping (i.e. national-scale) faces the diversity of greenhouses in different areas, the difficulty of the simultaneous extraction of the number and area of greenhouses, and the dense spatial distribution of greenhouses. In this paper, to solve the problem of large-scale greenhouse mapping, a novel data-driven deep learning framework is proposed, which we refer to as the dense object dual-task deep learning (DELTA) framework. The dual-task learning module simultaneously extracts the number and area of greenhouses by adopting a greenhouse area extraction branch and a greenhouse number extraction branch. A high-density-biased sampler module is proposed to select more samples in areas with a dense distribution, so that the trained model is more effective at dense greenhouse extraction. Six regions in China were selected for evaluation, which obtained a performance increment of 1.8% in mean average precision (mAP) when compared with Faster R-CNN. Finally, the whole of China was taken as the research area, and remote sensing image tiles at a 1-m spatial resolution from all over China were obtained. All the images were captured by different sensors and downloaded from open-source sites or purchased. The experimental results indicate that more than 13 million greenhouses were extracted in China. [ABSTRACT FROM AUTHOR]

• We construct a fine-grained classification dataset of wet Age-related macular degeneration (AMD). • A knowledge-driven deep learning model, KFWC, is proposed to improve accuracy in fine-grained disease classification with insufficient data. • KFWC can detect 10 signs of lesions on medical images and can diagnose two subtypes of wet-AMD. • Multiple experiments demonstrate the effectiveness and interpretability of the proposed KFWC. Automated diagnosis using deep neural networks can help ophthalmologists detect the blinding eye disease wet Age-related Macular Degeneration (AMD). Wet-AMD has two similar subtypes, Neovascular AMD and Polypoidal Choroidal Vasculopathy (PCV). However, due to the difficulty in data collection and the similarity between images, most studies have only achieved the coarse-grained classification of wet-AMD rather than a fine-grained one of wet-AMD subtypes. Therefore, designing and building a deep learning model to diagnose neovascular AMD and PCV is a great challenge. To solve this issue, in this paper, we propose a Knowledge-driven Fine-grained Wet-AMD Classification Model (KFWC) to enhance the model's accuracy in the fine-grained disease classification with insufficient data. We innovatively introduced a two-stage method. In the first stage, we present prior knowledge of 10 lesion signs through pre-training; in the second stage, the model implements the classification task with the help of human knowledge. With the pre-training of priori knowledge of 10 lesion signs from input images, KFWC locates the powerful image features in the fine-grained disease classification task and therefore achieves better classification. To demonstrate the effectiveness of KFWC, we conduct a series of experiments on a clinical dataset collected in cooperation with a Grade III Level A ophthalmology hospital in China. The AUC score of KFWC reaches 99.71%, with 6.69% over the best baseline and 4.14% over ophthalmologists. KFWC can also provide good interpretability and effectively alleviate the pressure of data collection and annotation in the field of fine-grained disease classification for wet-AMD. The model proposed in this paper effectively solves the difficulties of small data volume and high image similarity in the wet-AMD fine-grained classification task through a knowledge-driven approach. Besides, this method effectively relieves the pressure of data collection and annotation in the field of fine-grained classification. In the diagnosis of wet-AMD, KFWC is superior to previous work and human ophthalmologists. [ABSTRACT FROM AUTHOR]

• A defencing algorithm based on deep learning was proposed. • The algorithm based on U-Net and Pix2pixHD. • Chicken contours occlude by the wire mesh can be recovered. • The accuracy of caged chicken detection is improved after defencing. Cage farming is the mainstream farming mode in China. Accurate individual identification and behavioral detection of caged chickens can provide managers with a better understanding of chicken status. However, for image detection of caged chickens, the cage may affect the accuracy of the detection algorithm. For this reason, CCD (caged chicken defencing), a defencing algorithm based on U-Net and pix2pixHD, was proposed to improve caged chickens' detection accuracy. The proposed defencing algorithm can accurately identify the cage wire mesh and recover the chicken contours completely. In the test set, the detection accuracy of the cage wire mesh was 94.71%, while a structural similarity (SSIM) of 90.04% and a peak signal-to-noise ratio (PSNR) of 25.24 dB were obtained in the image recovery. To verify the practicality of the method proposed in this paper, we analyzed the performance of the object detection algorithm before and after defencing from the perspective of the most basic individual detection in the caged chicken detection task. We validated the defencing algorithm with different YOLOv5 detection algorithms, including YOLOv5s, YOLOv5m, YOLOv5l, and YOLOv5x. The experimental results showed that the defencing algorithm improved the detection precision of caged chickens by 16.1%, 12.1%, 7.3%, and 5.4%, respectively, compared with before defencing. The recall improvement was 29.1%, 16.4%, 8.5%, and 6.8%. To our knowledge, this is the first time that a deep learning-based defencing algorithm has been applied to caged chickens, and the detection accuracy can be significantly improved. The method proposed in this paper can remove cage wire mesh greatly and provide a technical reference for subsequent poultry researchers. [ABSTRACT FROM AUTHOR]

With the acceleration of economic globalization and trade liberalization, environmental issues have attracted more and more attention from the world and become one of the major issues facing mankind. A series of global environmental problems that have emerged under the background of economic globalization have become barriers restricting the sustainable development of the economy and society. As one of the key issues in network research, BP neural network provides an effective tool for studying the multi-objective and multi-dimensional interaction between networks. During training, the BP neural network can adaptively memorize the learning content in the weights of the network through learning. That is, the BP neural network has a high self-learning and self-adaptive ability.This paper fully contrasts and studies several neural networks and deep learning models at this stage,This paper fully contrasts and studies several neural networks and deep learning models at this stage, and forms a new type of network model. In this paper, through the construction of BP neural network training model for experimental simulation, the experimental results show that our city's urban expansion phenomenon in 2019 is still showing a growth momentum. The urbanization and eco-environment system of our city are developing in a good level with fast and stable development rate, which indicates that the development under the existing development conditions can realize the coordinated development of urbanization and environment. In order to protect the damaged ecological environment, it is necessary to improve the path of ecological environment damage compensation responsibility and diversified ecological environment damage compensation relief procedures. Academic and legal practitioners can distinguish clearly from the elements of tort construction and environmental tort, improve the system of environmental tort liability, and strengthen China's environmental legislation. • The neural network toolbox is used to implement a multi-layer feedforward network. • Solves the problems of overlapping information in the selection of indicators and the one-sided problems of screened indicators. • The researchers study from three aspects: the causes, the damage assessment methods and the proposed measures. • The research on environmental damage compensation can learn from the environmental damage compensation in the marine field to study other fields. [ABSTRACT FROM AUTHOR]

Various satellite sensors have provided a huge amount of observations of Earth's environment at variable spatial and temporal resolutions. Many global coarse-resolution land products have been generated from single-satellite data, but global temporally regular land products at fine spatial resolutions (say 10-30 m) are scarce because of infrequent observations. An ideal inversion framework can estimate multiple global continuous land variables at different spatial resolutions by combining all sources of satellite data. This paper proposes a new framework that can estimate five land variables simultaneously from the top-of-atmosphere (TOA) reflectance acquired by seven satellite sensors based on a multi-scale and multi-depth convolutional neural network (MSDCNN). This framework enables us to estimate temporally regular land variables at as fine as 10 m spatial resolution by transforming information from satellite data at coarser spatial resolutions. The estimated land variables include Leaf Area Index (LAI), Fraction of Absorbed Photosynthetically Active Radiation (FAPAR), shortwave albedo, visible albedo, and spectral reflectance. Seven sensors that acquire data at different spatial resolutions, include Visible Infrared Imaging Radiometer Suite (VIIRS) (750 m), Moderate Resolution Imaging Spectroradiometer (MODIS) (500 m), Fengyun-3 (FY-3) Medium Resolution Spectral Imager (MERSI) (250 m), China-Brazil Earth Resources Satellite 04 (CBERS-04) Wide Field Imager (WFI) (73 m), Landsat 8 Operational Land Imager (OLI) (30 m), Gaofen-1 (GF-1) Wide-Field-of-View (16 m) and Sentinel 2 A/B Multispectral Imager (MSI) (10 m). This framework is mainly composed of four steps. First, a Shuffled Complex Evolution (SCE) optimization method is adopted to estimate these variables from VIIRS TOA reference. Second, a joint-output random forest regression (RF) method is used to link the satellite observations and the estimated values from step 1. Third, the six-type multi-resolution satellite observations are used to downscale the VIIRS TOA reflectance to six different spatial resolutions by using the MSDCNN. Finally, the downscaled six-resolution VIIRS TOA reflectance is fed into the multiple-variable RF model to estimate land variables. The framework was validated, and the results based on high-resolution reference maps from ImagineS network and time series shortwave albedo field values from Surface Radiation (SURFRAD) and Integrated Carbon Observation System (ICOS) network show that the retrieved variables had high validation accuracy, with root mean square error (RMSE) ranges of 0.361–0.489 (LAI), 0.023–0.120 (FAPAR), 0.013–0.026 (snow-free shortwave albedo), respectively. Comparison results of the retrieved multi-scale variables with the Sentinel 2 A/B (10 m), Landsat 8 (30 m), Global LAnd Surface Satellite (GLASS, 250 m, 500 m), MODIS (500 m), and VIIRS (750 m) products show that their values were close, with RMSE ranges of 0.107–0.273 (LAI), 0.015–0.027 (FAPAR), 0.003–0.007 (shortwave albedo), 0.001–0.007 (visible albedo), and 0.003–0.025 (spectral reflectance), respectively. The results of the direct validation as well as the product intercomparison show that this novel framework has the potential to be used in estimating global land variables at various spatial scales using a variety of satellite data sources. • Exploring a deep learning method for seven-resolution land surface variable estimation. • The finest resolution of estimated variable is 10 m and the temporal continuity is improved. • Estimated variables at seven resolution have consistent spatial distribution. • The estimated multi-resolution variables have high accuracy based on extensive validation. [ABSTRACT FROM AUTHOR]

Accurate prediction of photovoltaic (PV) power plays a pivotal role in ensuring safe and stable grid operation, enhancing power supply quality, and facilitating proactive grid management to mitigate power fluctuations. This paper introduces a novel hybrid model named KS-CEEMDAN-SE-LSTM, which combines K-Shape (KS) clustering, Complete Ensemble Empirical Mode Decomposition with Adaptive Noise (CEEMDAN), Sample entropy (SE), and Long Short-Term Memory (LSTM) techniques. To improve the accuracy of PV power prediction, three advanced preprocessing techniques, KS clustering, CEEMDAN and SE, are employed for feature extraction before prediction. In the proposed approach, KS clustering is employed to capture the similarity features of PV power. Subsequently, CEEMDAN decomposes the PV power into multiple intrinsic mode functions (IMFs) to reduce the influence of non-stationary and noisy features in power prediction. To further enhance prediction accuracy and streamline computational complexity, SE methods and K-means clustering are utilized to reconstruct the decomposed IMFs into new sub-sequences with refined granularity features. Finally, the LSTM model is applied to predict each reconstructed sub-sequence, and the ultimate prediction results are obtained by aggregating the predictions from all sub-sequences. The effectiveness of this method is demonstrated using real-world PV power data collected from China. Comprehensive experiments substantiate that the proposed KS-CEEMDAN-SE-LSTM model outperforms benchmark methods in short-term PV power prediction. • A novel hybrid model proposed for PV power short-term prediction. • Three enhanced preprocessing techniques are used to extract PV power refinement features. • The deep learning based LSTM is used to predict PV power. • The proposed method achieves state-of-the-art performance and outperforms other baselines. [ABSTRACT FROM AUTHOR]

Probabilistic load forecasting (PLF) are of critical importance to reliable and economical power system operations. To further improve the PLF performance, this paper extends the load forecasting model with high accuracy to PLF by using Monte Carlo (MC) Dropout and soft thresholding (ST) for uncertainty modeling and noise reduction, respectively. Specifically, the method first builds a Dual-stage Attention-based Long and Short Time Pattern Network (DA-LSTPNet) by weighting the temporal information extracted from Gate Recurrent Unit (GRU) and convolutional neural networks (CNN) through the time attention mechanism(TAM). Then, an adaptive ST structure is used to reduce the impact of meteorological data noise on model prediction performance. The term ST-DA-LSTPNet refers to DA-LSTPNet with the addition of the ST structure. Finally, an appropriate Dropout setting is inserted in the neural network layers during the model's testing phase to achieve the model parameters' stochastic process. After multiple sampling and calculations, the final PLF result is obtained. To demonstrate the effectiveness of the proposed method, this work construct a case study using the actual data from a prefecture-level city in southern China. Compared with several benchmark methods, the proposed method significantly improves the performance of PLF and has good anti-noise generalization ability. [Display omitted] • The high-accuracy load forecasting model extends to probabilistic load forecasting. • The adaptive soft thresholding can mitigate noise impact on prediction. • The model performs well on actual grid data, with over 92% prediction coverage. • This work provides effective support for economic scheduling and management. [ABSTRACT FROM AUTHOR]

Aiming to automatically segment multi-class objects on the tunnel point cloud, a deep learning network named dual attention-based point cloud network (DAPCNet) is developed in this paper to act on point clouds for segmentation. In the developed model, data normalization and feature aggregation are first processed to eliminate data discrepancies and enhance local features, after which the processed data are input into the built network layers based on the encoder-decoder architecture coupled with an improved 3D dual attention module to extract and learn features. Furthermore, a custom loss function called Facal Cross-Entropy ("FacalCE") is designed to enhance the model's ability to extract and learn features while addressing imbalanced data distribution. To validate the effectiveness and feasibility of the developed model, a dataset of tunnel point clouds collected from a real engineering project in China is employed. The experimental results indicate that (1) the developed model has excellent performance with Mean Intersection over Union (MIoU) of 0.8597, (2) the improved 3D dual attention module and "FacalCE" contribute to the model performance, respectively, and (3) the developed model is superior to other state-of-the-art methods, such as PointNet and DGCNN. In summary, the DAPCNet model exhibits exceptional performance, offering effective and accurate results for segmenting multi-class objects within tunnel point clouds. • A deep learning method named DAPCNet is developed for 3D point cloud segmentation. • An improved 3D dual attention module is introduced to enhance model performance. • A custom loss "FacalCE" is designed to strengthen feature learning with handling data imbalance. • Conducting comparisons to examine the model performance for segmentation. • The developed method demonstrates outstanding performance, achieving an MIoU score of 0.8597 when applied to tunnel point clouds. [ABSTRACT FROM AUTHOR]

• A transfer and contrastive learning architecture identifying multi-label loads. • Contrastive learning architecture to extract deep features. • Transfer learning design solves the sparsity of multi-label appliances. • Gramian angular field encoding enhances the feature extraction efficiency. • Verifications on both public datasets and real-world measurements from China. To achieve the goal of peaking carbon emissions globally and carbon neutrality, smart energy management is a promising way to boost energy conservation and estimate the residential potential for regional demand response, among which the non-intrusive load monitoring technique is highlighted due to its effectiveness on the residential side. However, the identification of multi-label appliance switching operations is still a challenge in this field, which may critically affect the total identification results due to the few-shot learning problem and the complicated overlap of features belonging to different appliances. Therefore, this paper proposed a transfer and contrastive learning architecture to identify multi-label appliances effectively. In the first stage, Gramian angular field encoding is implemented to visualize power sequences to highlight the correlation between timestamps and enhance the feature extraction efficiency. Secondly, a contrastive learning architecture is implemented to learn the general distinguishing features between samples of different labels, and density-based spatial clustering of applications with noise clustering is utilized to detect multi-label samples. Thirdly, transfer learning is utilized to enhance the multi-label identification capacity of contrastive learning structures based on the existing trained model. Finally, the effectiveness of the proposed algorithm is verified through two low-frequency non-intrusive load monitoring public datasets and real-world measurements from a pilot project in China. The results show that the proposed architecture can achieve the efficacy of deep features extraction and few-shot learning in identifying multi-label appliance switching operations. [ABSTRACT FROM AUTHOR]

The air pollutant diffusion trend prediction plays an important role in the environment protection. The approaches in the existing studies rarely consider spatial and temporal information sequentially or cannot use spatiotemporal structure information effectively. In this paper, we propose a new deep learning-based air pollutant diffusion trend prediction model, called GATBL-Learning. It is an end-to-end neural network model consisting of Graph Attention Network (GAT) and Encoder-Decoder model. Since each city in the region will set as a node, GAT can extract spatial information from them by aggregation operation. Then Encoder-Decoder model, consisting of Bi-LSTM as Encoder and LSTM as Decoder, mines time series features and outputs the data containing spatiotemporal features. In addition, we also observe that the changes of air pollutants show seasonality discipline and training the seasonal model can improve the model adaptability. The results demonstrate the proposed prediction model is effective and superior. [ABSTRACT FROM AUTHOR]

[Display omitted] • In the case of a small number of known deposits and mineral occurrences, the prospecting prediction model based on the Siamese network can extract effective features without relying on prior knowledge to realize supervised prospecting prediction. • On the basis of sliding window data enhancement, this paper generates paired sample data sets by randomly combining deposit samples and non-deposit samples, which can further enhance the data and improve the generalization ability and prediction accuracy of the model to a certain extent. • Compared with the prediction results of the weight of evidence method and the convolutional neural network, the prediction results of the Siamese network are better, and 7 new gold prospecting areas have been discovered in Zhangjiakou, Hebei Province, China. Supervised neural networks constitute an important research area for the intelligent prediction of locations for prospecting mineral deposits. Accurate training of a supervised neural network requires many training samples, something that is often difficult to obtain. This paper reports the use of geological, geochemical, and geophysical data by a Siamese network to overcome the problem of insufficient training samples, and implements a supervised deep learning prospecting prediction method based on the Siamese network. Intelligent prediction for gold deposits prospecting is carried out for the Chongli mineral concentrate area in Zhangjiakou, Hebei Province, China, and compared with the weight of evidence method and convolutional neural network model. The results show that:(a) the performance of the Siamese network model is no less than that of the convolutional neural network (CNN) model and better than that of the weight of evidence (WOE) method; (b) the gold prospective areas differentiated by the established models are strongly consistent with geological and metallogenic characteristics in the study area. This study suggests Siamese network model as an effective mineral prospectivity modeling tool. This method is also suitable for prospecting prediction using geoscience data in other areas. [ABSTRACT FROM AUTHOR]

[Display omitted] • This paper shows how the problem of CM&FD of OOR wheels can be transformed into an ML problem. • A vehicle-track MBS model considering track irregularities and worn wheel profiles is built. • A DL model, OORNet, is proposed to realize CM&FD of our-of-round wheels. • The feasibility and superiority of OORNet are demonstrated by three other DL models. The problem of train wheel out-of-roundness (OOR) negatively affects both humans and the vehicle-track system, incl. reduced passenger comfort, rapid aging of vehicle/track components, increase in derailment risk, etc. It is therefore of interest to develop an on-board condition monitoring and fault diagnosis (CM&FD) technique for wheel OOR, which contributes not only to the maintenance decision-making of wheelsets but also to clarifying its triggering and evolution mechanisms. This paper first shows how to express the problem of CM&FD of our-of-round wheels as a machine learning problem. A deep learning model, OORNet, is then developed for CM&FD of out-of-round wheels. A vehicle-track multi-body dynamics model of a China railway high-speed (CRH) trailer is meanwhile built to produce a database consisting of vertical axlebox vibration accelerations caused by 2000 different wheel OOR curves. The simulated database is finally used to test the performance of OORNet, and its feasibility and superiority are verified. [ABSTRACT FROM AUTHOR]

In recent years, due to the rapid development of Chinese textile industry, the domestic demand for cotton increases sharply. Conversely, the cotton plantation area increasingly dwindled, resulting in the constant rise of cotton imports. China, as a great cotton importer, has classified manually the cotton grades for a long time, which not only results in a consumption of labor and financial resources, but also leads to some mistakes generated b the labor s subjective evaluation. This paper presents a method for automatic cotton classification for different appearance grades. Based on a comprehensive comparison, our method performs better in the classification of cotton appearance grades. PCANet feature recognition with basic impurity identification achieves the best performance. [ABSTRACT FROM AUTHOR]

• Deep reinforcement learning (DRL) based models are proposed for HVAC system energy consumption prediction. • Autoencoder (AE) algorithm is used to extract the high-level features of state space of DDPG. • The DDPG based models are applied for energy consumption forecasting over short-term time horizon. • Improvement in prediction performance is observed when combine AE with DDPG. • The DDPG based prediction models, outperform common supervised models like BP neural network and support vector machine. Short-term energy consumption prediction has fundamental importance in many HVAC system management tasks, such as demand-side management, short-term maintenance, etc. Currently, the prevailing data-driven techniques, especially supervised machine learning methods, are widely applied for short-term energy consumption prediction. Deep reinforcement learning (DRL), as the state-of-the-art machine learning techniques, have been applied for HVAC system control, but rarely for energy consumption prediction. In this paper, a DRL algorithm, namely Deep Deterministic Policy Gradient (DDPG), is firstly introduced for short-term HVAC system energy consumption prediction. Moreover, Autoencoder (AE), which is powerful in processing data in their raw form, is incorporated into DDPG method to extract the high-level features of state space and optimize the prediction model. The operation data of the ground source heat pump (GSHP) system of an office building in Henan province, China is used to train and assess the proposed models. The results demonstrate that the proposed DDPG based models can achieve better prediction performance than common supervised models like BP Neural Network and Support Vector Machine. This study is an enlightening work which may inspire other researchers to tap the potential of DRL algorithms in this field. [ABSTRACT FROM AUTHOR]

Accurate and reliable runoff prediction is of great significance to water resources management, disaster monitoring and rational development and utilization of water resources. In this paper, a metaheuristic evolutionary deep learning model based on Temporal Convolutional Network (TCN), Improved Aquila Optimizer (IAO) and Random Forest (RF) is proposed for rainfall-runoff simulation and multi-step runoff prediction. In this study, the influence of various input variables on the prediction accuracy is discussed. First of all, in order to avoid the dimensional disaster problems and reduce the calculation time, RF is used to calculate the correlation between the input variables and the prediction object, and the data with high correlation is selected as the final inputs. Then, the filtered data are sent to the TCN model, and the parameters of the TCN model are optimized using the IAO algorithm, and the final prediction results are obtained. In this study, the rainfall and runoff data of five stations in the middle reaches of Jinsha River, China were selected, and the runoff of Panzhihua station was simulated and predicted by establishing multiple models. By analyzing and comparing the predictive results of several models, it shows that the models and improvements proposed in this study are effective. [ABSTRACT FROM AUTHOR]

• The effect of time delay phenomenon in complex groundwater systems is stressed when spring discharge prediction methods based on deep learning are applied. • A delay-aware spring discharge prediction (DSDP) model is developed. • DSDP model outperforms baselines on a real karst spring and can predict the general trend of spring discharge. As an essential component of the hydrological cycle at both regional and global scales, karst aquifers store large amounts of water, serving as the main source of fresh water supply in many areas. Karst springs provide a natural drainage pathway for karst groundwater systems to participate in the hydrological cycle. Accurate prediction of karst spring discharge is of great value to the long-term planning and management of karst water resources. The mainstream methods proposed have achieved high prediction accuracy by introducing deep learning technology. However, these methods ignore the delay time of spring water discharge to water supply or drainage, so there is still room for improvement in prediction accuracy. To address this issue, this paper proposes a delay-aware spring discharge prediction (DSDP) framework that introduces the delay time to predict spring discharge. Specifically, the framework consists of three modules, including a time calibration module, a variation estimation module and a discharge prediction module. The time calibration module identifies the delay time of spring discharge to each element to calibrate the observations. And the variation estimation module utilizes a neural network to infer spring discharge variations. Finally, the discharge prediction module predicts spring discharge with derived discharge variations. This study validates the performance of the proposed method by using spring discharge data from Jinci Spring, a typical karst spring in China. The experimental results show that the proposed method outperforms baselines, with significant decreases in both MAE and RSM. The study will provide evidence for the precise management of Jinci Spring, and provide a methodological reference for the discharge prediction of other springs. [ABSTRACT FROM AUTHOR]

Accurate estimation of ride-hailing travel mobility is significant for demand management, and transportation planning. Although existing intervening opportunities models based on individual destination selection behavior can estimate travel mobility patterns (e.g., commuter flow, and migration flow), they usually ignore the substitutability of candidate destinations. In the context of ride-hailing travel, people tend to have strong destination preferences, and candidate destinations should be related to individual travel needs. Meanwhile, artificial intelligence offers powerful tools to extract complex nonlinear dependencies of mobility data, which are difficult to capture by traditional intervention opportunities models. This paper proposes a novel ensemble model with conditional intervening opportunities to improve the accuracy of ride-hailing travel mobility estimation by considering the substitutability of candidate destinations, that is, only the location related to people's trip purpose will likely affect people's travel behavior. The proposed ensemble model employs a stacking strategy to integrate six advanced machine learning and deep learning algorithms to extract complex nonlinear dependencies from ride-hailing travel mobility data, and achieve accurate mobility estimation. Furthermore, datasets from two major cities in China with more than 25 million ride-hailing trips are used for model training and experimental comparison. The results indicate that the proposed model outperforms other baseline models in ride-hailing travel mobility estimation tasks. It accurately predicts trip flows and the trip distance distribution, and can capture mobility patterns with strong interpretability. The proposed model can be applied to analyze the travel behavior of ride-hailing passengers, as well as the mobility patterns between different regions. • Propose a novel ensemble model for ride-hailing travel mobility estimation. • Develop conditional intervening opportunities model with substitutability of candidate locations. • Extract nonlinear dependencies from ride-hailing data using artificial intelligence algorithm. • Verify the proposed prediction model using over 25-million ride-hailing trips. • Interpret ride-hailing travel mobility patterns using Shapley additive explanations method. [ABSTRACT FROM AUTHOR]

Accurate forecasting is required for the effective risk management of drought disasters. Many machine learning- and deep learning-based models have been proposed for drought forecasting, however, they cannot handle the temporal and/or spatial dependencies in the input data, causing unexpected forecasting results. In order to solve the challenging issue, in this paper we proposed the Heterogeneous Spatio-Temporal Graph (HetSPGraph), for drought forecasting. It includes three major layers: spatial aggregations including inter and intra aggregations, temporal aggregation, and a forecasting network. The main function of HetSPGraph is to learn the dynamic spatiotemporal correlations between the regions and to further predict the drought in different regions, based on which accurate drought forecasting can be achieved. Experimental forecasting results of the Standardized Precipitation Evapotranspiration Index (SPEI) in China indicated that the HetSPGraph model outperformed the traditional baseline methods including the Long Short-Term Memory model (LSTM), Convolutional Neural Network-LSTM (CNN-LSTM), Gated Recurrent Unit (GRU), Spatio-Temporal Graph Convolutional Networks (STGCN) and Geographic-Semantic-Temporal Hypergraph Convolutional Network (GST-HCN). Even for long-term forecasting (12 months), more accurate forecasting results, with the coefficient of determination R 2 higher than 0.89, can also be obtained by HetSPGraph compared to the other three models. The proposed HetSPGraph model has the potential for wider use in forecasting drought and other natural disasters. • A novel HetSPGraph approach with three layers is proposed for drought forecasting. • Dynamic spatiotemporal correlation of droughts at multi-time scales can be described by HetSPGraph. • The HetSPGraph model outperformed conventional Deep Learning models for drought forecasting. • The HetSPGraph model can keep its stable accuracy at long-term forecasting period (12 months). [ABSTRACT FROM AUTHOR]

• Introduce the state-of-art temporal fusion transformers (TFT) into economic system forecasting based on the high compatibility between deep learning and the nonlinear characteristics of socio-economic systems. • Propose a scientific economic system forecasting framework on the foundation of the extended gross final product (GFP). • Build a four-dimensional evaluation system including numerical error, direction accuracy, timing error and turning point accuracy. • Empirically forecast the monthly output indicators of the Chinese macroeconomic system. • TFT results have obvious advantages over benchmark models including ARIMA, BP, LSTM, TCN and N-BEATS through comparison analysis. Although helpful in reducing the uncertainty associated with economic activities, economic forecasting often suffers from low accuracy. Recognizing the high compatibility between deep learning and the nonlinear characteristics of socioeconomic systems, in this paper, we introduce state-of-the-art temporal fusion transformers (TFTs) into the field of economic system forecasting and predict the performance of the Chinese macroeconomic system. Based on an extended analysis of gross final product (GFP) and the intertemporal dynamic relationship between demand-side indicators and output indicators, we establish a scientific economic forecasting framework. To summarize the forecasting characteristics of the TFT algorithm, we compare its one-step and three-step modeling effects in forecasting output indicators with a series of representative benchmark models. According to our proposed four-dimensional evaluation system, the forecasts for China's macroeconomic system provided by the TFT model have obvious advantages in terms of overall stability, forecasting efficiency, reduction of numerical and timing errors, direction accuracy, and turning point accuracy. The forecast results show that China's economy faces a risk of slowing growth in the post-pandemic period. [ABSTRACT FROM AUTHOR]

Quality control (QC) of marine data is a critical aspect in ensuring the usability of oceanic data. In this paper, we propose a novel QC method for time-series ocean wave data, which combines deep learning prediction and statistical analysis. Our method first realizes multi-element LSTM prediction of the time-series ocean wave observation data, capturing both temporal consistency and physical relationships between the multi-element inputs. Then, it applies peak detection on the regional mean difference ratio derived from the predicted and true values of the ocean wave data, and finally labels the anomalous data points based on peak detection results. We conducted experiments on the time-series wave data of four sites from the National Marine Science Center, China, and compared our proposed method with traditional QC method that is currently used for operational marine data quality control, as well as classic anomaly detection models including Isolation Forest and VAE-LSTM. The results show a significant improvement in the Precision, Recall, and F1 Score for erroneous samples using our proposed method. Our proposed method takes advantages of both deep learning and statistical analysis and considers physical correlation of multiple elements of marine data, effectively addressing the problem of erroneous discrimination of abnormal sea conditions in the traditional method, and providing valuable insights for the study of marine time-series observational data. • The results show a significant improvement in the Precision, Recall, and F1 Score for negative samples using our proposed approach. • By integrating deep learning and statistical methods, our proposed method achieves unsupervised quality control, effectively addressing the problem of erroneous discrimination of abnormal sea conditions in traditional methods, and providing valuable insights for the study of marine time-series observational data. [ABSTRACT FROM AUTHOR]

• A deep learning-based prediction framework of temperature control time is proposed. • Designing a historical rolling data processing method based on mean coding. • Designing a feature selection strategy based on maximum information coefficient. • Designing adaptive genetic algorithms to improve the deep neural network. • Experiments demonstrate the proposed prediction framework has the best performance. Accurate prediction of temperature control time is essential to improving the stability of wide-thick slab hot rolling production. However, the historical rolling data have high-dimensional and nonlinear characteristics, and there are complex relationships between the features affecting the temperature control time, which significantly impede modeling performance. To solve these problems and enhance prediction accuracy, this study proposes a prediction framework of temperature control time (PF-TCT). The proposed PF-TCT framework contains data preprocessing, feature selection, and model training. In the data preprocessing stage, considering the uneven distribution of categorical features taking values, a uniform random sampling method is designed to divide the dataset, and mean encoding is used to convert the categorical features into numerical features. In the feature selection stage, considering the interpretability of relationships between static parameters in historical rolling data and the accuracy of identifying redundant features, a selection strategy is designed based on the maximum information coefficient and approximate Markov blanket. In the model training stage, the deep neural network (DNN), which has a strong fitting ability and high flexibility, is used to build the prediction model. To avoid the dependencies between features and uneven sample distribution affecting the performance of the DNN, an adaptive genetic algorithm (AGA) is designed to optimize the DNN. In the AGA, population fitness and genetic iteration effect are considered to endow the selection, crossover, and mutation operators with adaptive learning ability, which can effectively provide good initial settings for the DNN parameters, thereby significantly improving the prediction performance of the PF-TCT. The proposed PF-TCT is verified by numerical experiments on actual data from a large steel group in China. The experimental results show that the data preprocessing and feature selection process in the PF-TCT can effectively improve the performance of the prediction model. Compared with the classical models, the AGA-based DNN proposed in this paper has lower prediction error and better stability and generalization performances. Finally, considering the maximum information coefficient matrix and approximate Markov blanket of features, the feature selection process is further analyzed, and the process knowledge for improving the hot rolling production of wide-thick slabs is discussed. [ABSTRACT FROM AUTHOR]

The accurate spatial-temporal prediction of photovoltaic (PV) power generation helps the power system dispatching department to make reasonable dispatching plans. In this paper, a robust spatial-temporal prediction model for PV power generation based on a denoising autoencoder (DAE) combining Gramian angular summation fields (GASF) and convolutional neural network (CNN) is proposed. First, the downscaling and modeling of power data use Pearson correlation to reduce inaccurate and inefficient predictions due to large-scale sparse data. Second, the model training data are encoded with noise reduction using a denoising autoencoder. Finally, the performance of the proposed model is experimentally verified. The results show that the model still performs well when the data exist in different degrees of contamination, with an average mean square error (MSE) of 71.58. Experiments based on China and the USA PV power generation datasets with mean absolute error (MAE) of 4.26 and 3.96, respectively. [ABSTRACT FROM AUTHOR]

• A PV/wind power prediction model based on Time2Vec, WDCNN and BiLSTM is proposed. • WDCNN model is adopted to expand receptive field and improve feature extraction. • Case study validates the effectiveness and superiority of the proposed method. • The proposed model shows significance in processing data with strong volatility. Accurate prediction of photovoltaic (PV)/wind power is an effective solution for the grid stability, reasonable dispatching and power supply reliability. Nowadays, various deep learning prediction methods are developed, most of which are recurrent neural network (RNN)-based and convolutional neural network (CNN)-based. To further improve the prediction performance, a novel hybrid method based on deep CNN with wide first-layer kernels (WDCNN) and bidirection long short-term (BiLSTM) is presented in this paper, in which WDCNN is introduced for large receptive field and useful information extraction, and stacked BiLSTM layers is incorporated to extract temporal correlations of past and future datasets. Besides, Time2Vec is adopted for better feature extraction through decomposing the time series data into non-periodic and periodic components. Several ablation and comparison experiments are carried out with a case study in Yongxing Island, China, and the performance metrics including normalized mean absolute error (NMAE), normalized mean square error (NMSE), normalized root mean square error (NRMSE), and mean absolute scaled error (MASE) confirm the effectiveness and superiority of the proposed model. Compared with individual WDCNN and BiLSTM, the performance of the combined WDCNN-BiLSTM are improved by 5.79%, 6.61%, 3.36%, 5.79% and 2.03%, 3.54%, 1.78%, 2.03% for PV prediction, and 20.38%, 29.37%, 15.96%, 20.38% and 39.12%, 59.68%, 36.50%, 39.12% for wind prediction, respectively. The adoption of Time2Vec further improves the prediction performance by 5.96%, 8.10%, 4.14%, 5.96% for PV prediction, and 4.86%, 9.19%, 4.70%, 4.86% for wind prediction, respectively. The proposed model yields most accurate prediction compared with other competing models. The replacement of CNN with WDCNN improves the prediction accuracy by 6.64%, 8.47%, 4.33%, 6.64% for PV prediction, and 30.59%, 42.78%, 24.35%, 30.59% for wind prediction, respectively. Moreover, the proposed model significantly outperforms compared models in prediction with more random fluctuations, which demonstrates the superiority of the proposed model in mining complex relationships. [ABSTRACT FROM AUTHOR]

Accurate forecasting of significant wave height (SWH) is crucial for ensuring the safety of marine navigation. To achieve good accuracy, deep learning models need to be trained separately using SWH data from all forecast locations. However, improving the generalizability of SWH prediction models is vital for practical engineering purposes. This paper proposes a training strategy that utilizes multi-point data fusion by using wave data obtained from different locations to construct a training dataset. To verify the feasibility of training strategy, the artificial neural network (ANN), long short-term memory (LSTM), and temporal convolutional network (TCN) models are used to test in China offshore. The RMSE and MAPE of different models with 2h advance forecasts are almost 0.05 and 4%, indicating that the models have good forecasting performance. Additionally, the models exhibit consistency with the model trained solely by single position data, which demonstrates that the training strategy applies to various models. In addition, in sea areas without wave training data, two locations are randomly selected and used in the pre-trained model to predict SWH. The various models demonstrate favorable forecasting results, further demonstrating that the training strategy proposed in this study can enhance the generalizability of models. • The training method based on multi-point data fusion can improve the spatial generalization of the SWH forecasting model. • Using ANN, LSTM, and TCN models for significant wave height forecasting at different locations. • The prediction accuracy of the multi-point data fusion training method and single location training method is the same. • The training method based on multi-point data fusion has good applicability to different deep learning models. [ABSTRACT FROM AUTHOR]

• Combined with the error reconstruction rate, an adaptive VMD technology is developed and used for data preprocessing. • Two mainstream deep learning methods (LSTM, DBN) are used as sub-series predictors, and the classic PSO algorithm is utilized to determine the number of neurons in the DBN network. • A novel hybrid prediction model based on the PSO-DBN nonlinear combination mechanism is proposed. Wind power forecasting plays a vital role in enhancing the efficiency of power grid operation and increasing the competitiveness of power market. In this paper, a novel hybrid forecasting model is developed by using the decomposition strategy, nonlinear weighted combination, and two deep learning models to overcome the drawbacks of the linear weighted combination and further enhance wind power forecasting accuracy and stability. Firstly, the variational mode decomposition (VMD) technique is employed to decompose the original wind power series to extract local features. Then, the long short-term memory neural networks (LSTM) and deep belief networks based on particle swarm optimization (PSO-DBN) are utilized to construct sub-series prediction models. Finally, the multiple sub-series forecasting models are integrated by a nonlinear weighted combination method based on PSO-DBN to construct a hybrid model for short-term wind power forecasting. To verify the performance of the developed forecasting model, wind speed data of 10 min from a wind farm in Shaanxi Dingbian, China are selected as case studies. The results show that the proposed method in this paper is more effective than other existing methods. [ABSTRACT FROM AUTHOR]

• Plug-in hybrids with dual-mode engines greatly reduce fuel consumption and emissions. • A hierarchical energy management strategy is proposed for close-to-optimal control. • Compared with mainstream control methods under standard and real-world drive cycles. • Generalization and robustness to random initial and terminal conditions are proved. The dedicated hybrid engines (DHEs) with dual-mode combustion technology can drastically reduce the fuel consumption and emissions while guarantee the power density. This paper aims to investigate the optimal control of such DHE-based plug-in hybrid electric vehicles (PHEVs) under real driving conditions, with minimum fuel penalties caused by transient engine dynamics. For this purpose, the benefits brought by artificial intelligent control and traffic preview in terms of energy efficiency can be combined with the advantages of advanced combustion engine. This paper presents a hierarchical energy management strategy (HEMS) to realize the synergy of global and instantaneous optimization. At the cloud level of HEMS, dynamic programming is applied to obtain optimal combustion mode and state of charge reference trajectories in a receding horizon. At the powertrain level, deep reinforcement learning with a ranking-prioritized experience replay algorithm is used to output optimal engine power and combustion mode for the energy management. To evaluate the proposed strategy, a dual-mode engine with homogeneous charge compression ignition and spark ignition systems is tested and mapped, with which the PHEV is modeled in GT-Suite and Matlab/Simulink. Comprehensive experiments are carried out to verify the optimality, generalization and robustness based on a standard driving cycle and a real-world driving cycle in China with GPS data recorded. The results show that the HEMS avoids frequent switching of combustion modes and outperforms the conventional methods by more than 4% and 10% in terms of fuel economy and NOx emissions, respectively, with random initial and terminal conditions. [ABSTRACT FROM AUTHOR]

• A deep learning algorithm is used to build a comprehensive drought monitoring model. • Multi-source remote sensing data is processed as input data. • The comprehensive drought model has good applicability in the monitoring of meteorological drought and agricultural drought. • The model successfully monitored a drought process in Henan Province, China. • This study provides a new method for the comprehensive assessment of regional drought. Drought is a popular scientific issue in global climate change research. Accurate monitoring of drought has important implications for the sustainable development of regional agriculture in the context of increasingly complex global climate change. Deep learning is a widely used technique in the field of artificial intelligence. However, ongoing on drought monitoring using deep learning is relatively scarce. In this paper, the various hazard factors in drought development were comprehensively considered based on satellite data including Moderate Resolution Imaging Spectroradiometer (MODIS) and tropical rainfall measuring mission (TRMM) as multi-source remote sensing data. By using the deep learning technique, a comprehensive drought monitoring model was constructed and tested in Henan Province of China as an example. The results showed that the comprehensive drought model has good applicability in the monitoring of meteorological drought and agricultural drought. There was a significant positive correlation between the drought indicators of the model output and the comprehensive meteorological drought index (CI) measured at the site scale. The consistency rate of the drought grade of the two models was 85.6% and 79.8% for the training set and the test set, respectively. The correlation coefficient between the drought index of the model and the standard precipitation evapotranspiration index (SPEI) was between 0.772 and 0.910 (P < 0.01), which indicated a strong level of significance. The correlation coefficient between the drought index of the model and the soil relative moisture at a 10 cm depth was greater than 0.550 (P < 0.01), and there was a good correlation between them. This study provides a new method for the comprehensive assessment of regional drought. [ABSTRACT FROM AUTHOR]

One of the key and difficult points in seismic data processing is seismic data denoising. Under the influence of the acquisition environment, the collected seismic data usually have low signal-to-noise ratio (SNR) and low resolution. In desert region of western China, different from other regions, the random noise there has complex characteristics of non-Gaussian, non-stationary and non-linearity. Its main frequency is quite low and its spectral overlap with those of effective signals. Moreover, large amount of data and intelligent requirements make traditional denoising methods encounter difficulties. In addition, the existing deep learning denoising methods also reveal two problems: first, they can only effectively suppress simple seismic noise such as Gaussian noise. But for the case of desert noise, they usually mistakenly judge noise as effective signals and destroy signal characteristic structure. Second, what is more important, in supervised-based methods paired noisy and pure data are usually used as training set. However, there is no noise-free data in actual desert seismic data, which severely limits the processing performance of supervised-based networks. Thus, a new parameter-shared variational auto-encoding adversarial network (PS-VAAN) is proposed for desert seismic data denoising in this paper. This new method includes two encoders, two generators and two discriminators. It can realize domain conversion from noisy data domain to pure data domain. We construct relatively complete unpaired pure and noisy data training sets for training, and design reconstruction loss function and adversarial loss function to optimize network parameters. Moreover, cycle-consistency loss is introduced to make sure two domains mapping into the same space. Compared with the state-of-art denoising methods on synthetic and actual seismic records, the proposed method not only has superior denoising ability, but also has strong feature extraction and fitting ability to recover effective signals with almost no energy loss. • Desert seismic noise has the characteristics of low frequency and non-stationary. • This paper designs a parameter-shared variational auto-encoding adversarial network. • The network use unpaired noisy and pure training sets to optimize parameters. • The proposed method has a strong ability to recover effective signals from noise. [ABSTRACT FROM AUTHOR]

The research of traffic revitalization index can provide support for the formulation and adjustment of policies related to urban management, epidemic prevention and resumption of work and production. This paper proposes a deep model for the prediction of urban Traffic Revitalization Index (DeepTRI). The DeepTRI builds model for the data of COVID-19 epidemic and traffic revitalization index for major cities in China. The location information of 29 cities forms the topological structure of graph. The Spatial Convolution Layer proposed in this paper captures the spatial correlation features of the graph structure. The special Graph Data Fusion module distributes and fuses the two kinds of data according to different proportions to increase the trend of spatial correlation of the data. In order to reduce the complexity of the computational process, the Temporal Convolution Layer replaces the gated recursive mechanism of the traditional recurrent neural network with a multi-level residual structure. It uses the dilated convolution whose dilation factor changes according to convex function to control the dynamic change of the receptive field and uses causal convolution to fully mine the historical information of the data to optimize the ability of long-term prediction. The comparative experiments among DeepTRI and three baselines (traditional recurrent neural network, ordinary spatial–temporal model and graph spatial–temporal model) show the advantages of DeepTRI in the evaluation index and resolving two under-fitting problems (under-fitting of edge values and under-fitting of local peaks). [ABSTRACT FROM AUTHOR]