Traffic flow prediction is an important problem in the field of intelligent transportation systems. Most existing traffic-flow prediction methods have made good progress, which however still face the following two key challenges. (1) The underlying pattern of traffic flow depends on not only the historical information along the timeline, but also the information of spatially adjacent areas, making it a challenging problem on how to balance the two temporal-spatial patterns; (2) Due to fact that time information has the characteristic of multiple granularity, such as hour, day and week, how to capture the multi-grained temporal patterns is another challenge problem. In this paper, we design a multi-grained spatio-temporal deep regression model (MGSTDR) to address the above challenges. By extending the typical autoregressive integrated moving average model (ARIMA) on the basis of multi-grained spatio-temporal traffic flow information, the proposed model can effectively use historical information along the timeline as well as the information of adjacent regions, such that the prediction of multi-grained spatio-temporal traffic flow can be performed. Experimental results on several datasets demonstrate that the proposed model outperforms existing benchmark methods on the task of multi-granularity, and particularly obtains an approximately 5.66% improvement in the hourly traffic flow prediction.

Aiming at the problem that the existing fault diagnosis method has too high requirements for systems, the weak diagnosability of fuzzy discrete event system (FDES) is studied, and a weak fuzzy diagnosability method is proposed, which extends the weak diagnosability method of classical discrete event system (DES) to the fuzzy system. Firstly, the notion of weak fuzzy diagnosability of FDES is formalized; In order to verify the weak fuzzy diagnosability of fuzzy systems, a verifier automaton is constructed, and the necessary and sufficient conditions for the weak fuzzy diagnosability of FDES are obtained, in which the weak fuzzy fault diagnosis of fuzzy systems is realized. This method is suitable for weak fault diagnosis of both FDES and classical DES.

Professor Cai Wen, a nationally renowned expert with outstanding contributions from Guangdong University of Technology, published a paper on Extension Set and non-compatible Problem in the Journal of Science Exploration in 1983, marking the birth of an original Chinese discipline called Extenics. Extenics is a kind of science that uses formal models to study the laws and methods of extension and transformation of things, and is used to achieve innovation and handle contradictory problems. The year 2023 marks the 40th anniversary of the establishment of Extenics. Through the joint efforts of numerous Extenics researchers, the theory and methodology of Extenics has been gradually improved, and has been widely applied in fields such as engineering technology, information science and intelligent science, management and economy, education and teaching, innovation and entrepreneurship, and has shown its important application value and effectiveness. This article summarizes the development process, research overview, internationalization and socialization of Extenics, and briefly introduces the scientific significance and academic evaluation of Extenics followed with the prospects for the future trends in the development of Extenics.

To address the problems of missing detection and low detection accuracy of the existing helmet wearing detection algorithms for small and crowded targets detection, this paper proposes a helmet wearing detection method based on improved YOLOv5 and transfer learning. First, different from the default priori frame that is not suitable for the task, we use the K-means algorithm to cluster the suitable priori frame size for the detection task. Then, in the back of the feature extraction network, we introduce a spatial channel mixed attention module to strengthen the learning of relevant weights and suppress the weights of irrelevant backgrounds, respectively. Further, we improve the judgment metric of the non-maximum-suppression (NMS) algorithm in the post-processing stage of YOLOv5 to reduce the phenomenon of false deletion and missing of prediction boxes. After that, the proposed network is trained based on the strategy of transfer learning, which can overcome the scarcity of limited existing data sets and improve the generalization ability of the model. Finally, we build a cascade judgment framework for helmet wearing deployed in visual sensor networks. The experimental results show that our proposed method improves the average accuracy (IOU=0.5) to 93.6%, which is 5% higher than the original model in the helmet wearing data set. The proposed model also outperforms other state-of-the-art algorithms by obviously improving the accuracy of helmet wearing detection in the construction scenarios.

In order to realize the wide application and improve the performance of third-generation semiconductor devices, the development of new and high-performance packaging interconnect materials has become a key initiation. Among them, micro/nano copper material can be sintered into block structures with high electrical conductivity, high thermal conductivity, high stability and electromigration resistance under low temperature conditions due to its surface effect and small size effect, which has become a research hotspot for the development of new packaging interconnect materials. However, the problems of easy oxidation, agglomeration and low yield of micro/nano copper material limit its application in third-generation semiconductor devices, and the oxidation resistance improvement of micro/nano copper material has become a key problem to solve its application. In this research, benzimidazole was used to treat micro/nano copper material, and the coating of micro/nano copper particles by benzimidazole was confirmed by Scanning Electron Microscope (SEM), infrared spectroscopy and other characterization methods. The copper paste was subjected to X-ray diffraction (XRD) and X-ray photoelectron spectroscopy (XPS) to prove that the paste could be left in air for 120 days without oxidation. The strength of the interconnect joint prepared at 300 °C reached 62.3 MPa, and the resistivity of the sintered layer was as low as 6.18×10 ⁻⁸ Ω·m. The results show that the method of treating micro/nano copper material by benzimidazole can help it to achieve super oxidation resistance and good interconnection performance, which is of profound significance for the research and development of third-generation semiconductor packaging interconnect materials.

Users' long-term preferences and the relationship between historical information and current situation are usually ignored in existing user preference modeling and user record unified modeling. To address this, in this paper, we propose a music recommendation model based on users' long-term and short-term preferences and music emotional attention. Firstly, we divide the user's listening record into multiple historical and current sequences, and learn the features by multiple long and short-term memory networks, respectively, to obtain user's long and short-term preferences. For the historical music sequences, we propose the concept of sequence period is proposed to obtain the long-term preferences by calculating the weights of the sequence period. For the current sequence, we use the average pooling to extract the music features of the current scene to obtain the short-term preference. Secondly, we learn the emotional characteristics of music from the acoustic signals, and use the attention mechanism to calculate the emotional factors of music. Finally, we integrate the music emotion factor into the user's long-term and short-term preference to generate a music recommendation list. The experimental results on the last.fm real data set show that the proposed model achieves 0.5435 in term of the NDCG@10, which is better than the existing methods. The ablation experiment and characteristic contribution analysis further demonstrate the effectiveness of the model.

In order to avoid the problems of life shortening or safety accidents caused by the surface defects of lithium batteries, it is necessary to study an efficient and accurate methods for lithium battery electrode plate defect detection. In this research, the simple Unet semantic segmentation network is used to detect defects of lithium battery. In order to improve the segmentation accuracy, first the coding structure in the original network is replaced with VGG16, which is similar to the Unet coding structure, to obtain the pre-training weights had been trained. Then, feature fusion module of the simply fusion pyramid network (SFPN) is added to the skip connection of the Unet network to avoid large information differences between feature maps. Finally, label smoothing is applied to optimize the loss function to prevent the network from overfitting. Through experimental verification, the accuracy of the semantic segmentation network optimized by the proposed method is improved to 93.70%, and the probability of false segmentation and segmentation discontinuity is significantly reduced. This optimization process has certain practical value.

Currently, the amount of both information and knowledge is growing rapidly, and many problems faced by enterprises, organizations and individuals are characterized by non-linearity, high uncertainty and dynamism. Problems with relatively definite goals and domains and uncertain or insufficient boundary conditions leading to unclear problem expression and difficult goal achievement are called ill-defined problems. In this research, based on the method of constructing the extension model and solving the contradictory problem in Extenics, the method of constructing the initial extension model of the ill-defined problems and the method of solving the single-goal ill-defined problems are proposed by using the extensible analysis method and the extension transformation method. Finally, the feasibility of the method is demonstrated by a case study. Since this study applies a combination of formal and quantitative methods, it also lays the foundation for the intelligent solution of single-goal ill-defined problems and the study of multi-goal ill-defined problems.

Falling behavior will bring serious injury to the elderly, especially to the elderly living alone. How to correctly identify the falling behavior and issue a warning is an important factor for reducing the injury. A fall detection algorithm is proposed, which is based on TSSI (tree structure skeleton image) and learnable STB-CNN (spatio-temporal block convolutional neural network). Firstly, human joint point is extracted by the three-dimensional pose estimation algorithm, and the corresponding skeleton sequence can be obtained. Secondly, the skeleton sequence is encoded into TSSI by the algorithm based DFS (depth first search) method. Finally, a learnable STB-CNN is proposed to classify TSSI and detect the fall behavior, which consists of spatio-temporal difference module, learnable spatio-temporal framework and spatio-temporal multi-branch convolution module,. Experiments are carried out on the public datasets UR FALL Detection Datasets and the simulation datasets. Experimental results are shown that our fall detection algorithm is more accurate than other related algorithms, especially the accuracy to 98.6% and 98.3% respectively.

In order to ensure the safe production of glyphosate production enterprises and reduce or avoid the occurrence of fire and explosion accidents in the production process, the improved Dow Chemical Fire and explosion risk index evaluation method (Dow method) is proposed to carry out safety evaluation. In view of the inaccurate calculation of safety measure coefficient in the current general Dow method, the personnel safety compensation coefficient is added, and the safety compensation is divided into event prevention type and consequence mitigation type, and then the hazardous process units are quantitatively analyzed and calculated, so as to improve the accuracy of the safety evaluation results of Dow method. The improved Dow method is applied to the safety evaluation of a glyphosate production enterprise, and it is found that compared with the traditional Dow method, the improved fire and explosion risk index (F&EI) and exposure area are more sensitive and can more accurately reflect the fire and explosion safety evaluation results of glyphosate enterprises.

There are many factors involved in the evaluation of coal mine dust concentration, and the output evaluation results lack comprehensiveness when the single type of dust concentration is taken as the evaluation standard. In this research, a comprehensive evaluation method of coal mine dust concentration is established by using the superiority evaluation method in the extension innovation method, and the corresponding extension intelligent evaluation software is developed on the C++platform. First, many kinds of dust in the environment is analyzed, the data queried to determine the measurement indicators and the harmfulness of various kinds of dust, and the weight coefficient set, then an appropriate correlation function selected to calculate the degree of dust concentration in the environment meeting the requirements. Finally, a relatively comprehensive conclusion is obtained through the calculation and analysis of the comprehensive correlation function, by which the hazard of the coal mine dust concentration in a certain environment is comprehensively evaluated. The calculation and analysis process adopts the self-designed extension intelligent evaluation software of the coal mine dust concentration, which can intelligently output coping strategies according to the calculation results.

For a class of complex dynamic networks composed of nonidentical nodes, a decentralized control strategy is proposed to achieve the finite-time partial state components synchronization. The finite-time partial state components synchronization means that only some, not all, state components of each node in the network can achieve synchronization in finite time. Firstly, for the convenience of the theoretical analyses and derivation, a special diagonal matrix is introduced, which can formulate the desired state components of each node. Secondly, based on this special diagonal matrix, the finite-time partial state components synchronization is defined. Compared to the finite-time synchronization, the finite-time partial state components synchronization is more popular. Thirdly, according to both the control theory and the finite-time stability theorem, a decentralized control strategy is proposed, so that our networks can achieve the finite-time partial state components synchronization. Finally, a simulation example is shown to verify the effectiveness and correctness of the proposed control strategy in this paper.

Deepfake detection aims to authenticate facial images and videos, which can offer operational and technical support to safeguard personal portrait rights, prevent fake news, and curb online deceit. Early detection technologies are usually based on convolutional neural networks (CNNs) and have achieved promising detection performance. However, there exists a prevalent issue of mediocre generalisation performance. To enhance the overall generality of Deepfake detection, recent research has focused on a deep neural network Transformer by utilizing the self-attention mechanisms. The Transformer can better model long-distance dependency and global receptive fields to capture the image context association and video timing relationships, such that the representation ability of the detectors can be improved. This survey first provides an overview of the research background in this field, followed by an explanation of the common techniques used to generate Deepfake. Then, the existing Transformer-based detection methods are summarized and comparatively evaluated. Finally, the challenges and future research directions of Deepfake detection are discussed.

Owing to advantages of high theoretical energy density, low cost and environmental friendliness, lithium-sulfur (Li-S) battery is considered as one of the most promising next-generation high-energy-density batteries. The "shuttle effect" of polysulfides is the key issue hindering the commercialization of Li-S batteries. Adoption of "catalytic" strategy to enhance the sulfur redox kinetics has been demonstrated to be an effective way to alleviate the "shuttle effect". Single-atom catalysts (SACs) have received much attention in the field of catalysis due to their uniform metal active centers, unique electronic properties, and theoretically 100% metal atom utilization. In recent years, SACs have been introduced into Li-S systems and studied to achieve fast sulfur conversion kinetics. In this research, the latest progress in the application of SACs in Li-S batteries was reviewed, with special emphasis on the discussion of key factors affecting the catalytic activity of SACs. The prospects of SACs for Li-S batteries were pointed out and highlighted. Important guidance is provided for future design and fabrication of high-performance SACs for Li-S battery application.

With the rapid development of mobile communication technology, the contradiction between the limited spectrum utilization resources and the demand of a lot of spectrum communication is increasingly aggravated. New intelligent methods are needed to improve the utilization rate of spectrum. A multi-user dynamic spectrum access method based on distributed priority experience pool and double deep Q network is proposed. This method can help the secondary users to continuously learn according to their perceived environment information in the dynamic environment, and choose the idle channel to complete the spectrum access task for improving the spectrum utilization rate. In this method, a distributed reinforcement learning framework is adopted, and each secondary user is regarded as an agent. Each agent learns by using standard single-agent reinforcement learning method to reduce the underlying computing overhead. In addition, the method adds priority sampling on the basis of neural network training, and then optimizes the training efficiency of neural network to help sub-users choose the optimal strategy. The simulation results show that this method can improve the success rate, reduce the collision rate and improve the communication rate.

With the advantages of high flexibility and lightweight, unmanned aerial vehicles (UAVs) have been widely used in data collection of wireless sensor networks. For a multi-UAV-assisted wireless sensor network with randomly distributed and moved users, how to plan the flight paths of the UAVs to effectively collect data from the users remains a challenging problem. This paper aims to maximize the average throughput of data collection in a dynamic environment where the user's location cannot be predicted by optimizing the flight path of the UAVs, which is subject to the shortest flight time and range constraints of UAVs, the constraints of UAV start and end points, the communication distance constraints, the user communication constraints, and the UAV collision avoidance constraints. The resultant problem can be solved by using existing optimization methods with high complexity, which however is difficult to obtain the globally optimal solution. To address this problem efficiently, this paper proposes a deep reinforcement learning algorithm based on Dueling Double DQN (Dueling-DDQN). The proposed algorithm adopts the Dueling network architecture, which enhances the learning ability of the algorithm and improves the robustness and convergence speed of tracked in suboptimal solutions due to the over-estimation on the $ Q $ value. Simulation results show that the proposed algorithm can efficiently obtain the flight paths of multiple UAVs under all constraints. In particular, our proposed algorithm has encouraging convergence and stability performance in comparison with the existing benchmark algorithms.

In this paper, we propose a channel-split-based dual-branch block (CDBlock) based on the PointMLP method, which splits input features into two groups along the channel dimension and feeds them into different branches of the dual-branch network module. Specifically, our proposed CDBlock consists of a lightweight branch and a heavyweight branch. The lightweight branch, which uses the structure of the residual MLP, is designed to extract coarse semantic features. The heavyweight branch, which uses the bottleneck network as backbone, is responsible for extracting more deeper and distinguishable information. By doing this, our proposed CDBlock effectively improves the network's feature extraction and learning capabilities. Experimental results show that our proposed method outperforms the existing PointMLP by achieving an overall accuracy of 86.2% and a class average accuracy of 84.97% on ScanObjectNN dataset. In particular, our approach achieves better results with using fewer parameters and computational cost than the PointMLP. Additionally, our approach also achieves encouraging performance on the ShapeNetPart dataset.

To address the scheduling problem of multi-department cooperative treatment of patients under the condition of limited medical resources in a hospital, this paper proposes a collaborative treatment scheduling algorithm based on intelligent optimization. The proposed algorithm regards the cooperative treatment scheduling of doctors, nurses and patients in different scenarios as a multi-role cooperative control problem. In order to optimize the role's access behavior, we propose a decision-making model to guide the role to make the optimal access behavior, and introduce an intelligent optimization algorithm to optimize the decision-making model. For the case scenarios of collaborative treatment of patients, doctor ward rounds, and patient physical examinations, we conduct experiment to compare four scheduling strategies, includingthe random, shortest distance, maximum free space, and decision-making models, and comparatively analyze the performance of the genetic algorithms, particle swarm optimization, simulated annealing, and differential evolution in optimizing the decision-making models. The experimental results demonstrates that the decision-making model based on the differential evolution algorithm performs the best, and the optimized decision-making model can find feasible solutions in the case scenarios and also obtain the optimal scheduling results.

Graph convolutional networks (GCN), with natural advantages for skeleton-based action recognition, has attracted more and more attention. The key lies in how to obtain richer feature information and the design of the skeleton topology. In this research, the feature fusion method of joint and semantics (joint type and frame index) is improved, and integrated into a Semantics Coding Module (SCM), which is more applicable for complex multi-layer networks. Guided by the SCM, the network can obtain more feature information of skeleton. Secondly, a skeleton Topology Inference Network (TIN) is proposed, which adaptively learns different adjacency matrices according to the context information of different samples with the efficient feature learning ability of CNN, so that the network can get rid of the limitation of fixed topology. By applying the SCM and TIN to 2s-AGCN, we propose a semantics-guided multi-stream adaptive topology inference graph convolutional network for skeleton-based action recognition. Extensive experiments on datasets, NTU RGB+D and NTU RGB+D 120, demonstrate that our methods obviously improve the accuracy of network and our model has achieved the state-of-the-art performance.

This paper proposes a new distance metric (NewDM) of categorical data based on the graph structure of ordinal and nominal attributes to address the poor effect of most existing measurement methods for categorical data. In NewDM, it first summarizes the basic framework formula of distance definition of categorical data and analyzes the challenges of measuring categorical data. Then, the graph structures of ordinal attributes and nominal ones are utilized to define the distance between two probability distribution columns. Finally, a new distance metric of categorical data is obtained through simultaneous weight. Experimental results on six public datasets show that the proposed NewDM is superior to the state-of-the-art approaches.

As a pivotal part of the task-oriented dialogue system, dialogue policy can be trained by using the discriminative deep Dyna-Q framework. However, the framework uses vanilla deep Q-network method in the direct reinforcement learning phase and adopts MLPs as the basic network of world model, which limits the efficiency and stability of the dialogue policy learning. In this paper, we purpose an improved discriminative deep Dyna-Q method for task-oriented dialogue policy learning. In the improved direct RL phase, we first employ a NoisyNet to improve the exploration method, and then combine the dual-stream architecture of Dueling Network, Double-Q Network and n-step bootstrapping to optimize the calculation of the Q values. Moreover, we design a soft-attention-based model to replace the MLPs in the world model. The experimental results show that our proposed method achieves better results than other baseline models in terms of task success rate, average dialog turns and average reward. We further validate the effectiveness of proposed method by conducting both ablation and robustness analysis.

As a key component of nuclear fuel, uranium resources are crucial to ensuring the construction and sustainable development of our country's nuclear industry. However, conventional terrestrial uranium resources are facing the challenges of relatively poor and low-grade mineable uranium ore. In contrast, seawater uranium resource reserves are nearly a thousand times that of land, and China owns vast sea areas and is rich in seawater resources. Therefore, how to efficiently extract uranium resources from seawater to meet the development needs of nuclear industry in China is an important scientific issue that needs to be explored and solved urgently. The development status of seawater uranium extraction technology is systematically introduced, the research progress of the uranium extraction materials especially the most promising adsorption materials summarized, taking material design and engineering application as the key points to clarify the main challenges and future development prospects of seawater uranium extraction technology.

The method of transforming bridge is an effective method in Extenics used to solve antithetical problems. The study addresses the problem in the application of the transforming bridge, refines and improves the existing transforming bridge, optimizes the general steps, and applies it to the planning and design of a golf course to obtain a variety of effective design strategies. The further refinement of the transforming bridge will provide more specific methods for the solution of antithetical problems in various fields, and has very important application value for the formal and quantitative study of antithetical problems.

Due to the fact that features are not category-deterministic and the equipment resources are usually limit to support the category structure learning of samples, existing knowledge distillation methods possibly ignore the category knowledge distillation of samples. Therefore, this paper proposes a distillation method based on incremental class activation knowledge (ICAKD). First, this paper uses the class activation gradient map to extract class-discriminative sample features and proposes a class-activation constraint loss. Then, an incremental memory bank is built to store class-deterministic features, and multiple training batch samples are saved and updated iteratively. Finally, our proposed method calculates the quasi-quality center of the samples in the memory bank to construct the category structure relationship, and further performs the category knowledge distillation according to the class-activation constraint and the category structure relationship. Experimental results on the Cifar10, Cifar100, Tiny-ImageNet, and ImageNet datasets show that the proposed method achieves a 0.4%～1.21% improvement in term of accuracy when compared with the Category Structure Knowledge Distillation(CSKD) methods, demonstrating the promising effectiveness of the characteristics and increment of category judgment for category knowledge distillation.

Cardiac multi-class segmentation is of great significance in medical imaging, which can provide accurate cardiac structure information and assist clinical diagnosis. However, in the training of multi-class semantic segmentation models with high-resolution cardiac images, the loss of deep features due to multiple downsampling operations leads to the problems oforgan discontinuity and incorrect edge segmentation in the segmented cardiac. To address this, this paper proposes a 3DCSNet based on self-attention and 3D convolution for cardiac multi-class segmentation. Specifically, our proposed network introduces the 3D feature fusion module and a 3D spatial perception module into the segmentation network. The former 3D feature fusion module integrates self-attention and 3D convolution for parallel feature extraction, which is able to efficiently allocate the attentions weights within and between channels under the same dimension of the feature map. The latter 3D spatial perception module captures the positional correlation information between different dimensions by integrating the self-attention mechanism, avoiding the loss of important information in downsampling and further retaining the deep key features. Experimental results show that the proposed 3DCSNet outperforms several existing models on a publicly available 3D computed tomography image dataset (ImageCHD).

Machine learning tasks are usually single-task learning. However, in practical applications, the learning tasks are often related. As a result, the correlation between tasks is often ignored, and the complexity of samples is sometimes not considered in single-task learning. To address this, this paper proposes a new inductive transfer learning model based on the self-paced learning strategyby constructing a prediction model to learn the shared parameters of multiple related source tasks. First, the proposed model uses the strategy of self-paced learning to jointly learns the multiple related source tasks, and weights the learning samples according to the loss and difficulty degree of the samples in the source tasks. The parameters of the self-paced learning are iteratively updated to select the optimal samples with the less loss. Then, the knowledge learned from multiple related source tasks guide the learning of target tasks to construct multiple models for related transfer learning target tasks, and transfer these models to related target tasks to improve their generalization ability. Finally, we optimize the target model by using the Lagrange function to improve the performance of the classifiers. Experimental results show that the proposed model is superior to the existing transfer learning model under the same experimental conditions.

In this paper, we investigate the problem of consensus in leader-following multi-agent system, where the information of the leader is only accessed by a subset of the following agents. For the part of the follower who cannot obtain the leader’s information, the state of the leader need to be estimated. In addition, considering the discontinuity of obtaining the output information of agents under certain conditions, an impulsive observer is introduced to reduce the sampling times among multiple agents. To achieve this, this paper aims to design a controller based on impulsive observer to achieve the consensus of leader-following multi-agent systems. Firstly, an impulsive full-order observer and a consensus protocol are designed for each follower, so that the follower can use the observer to estimate the leader. Secondly, the dynamic equation of the error system is derived, and the appropriately Lyapunov function is constructed by using the error variables. Finally, the stability of error systems is studied by using the Lyapunov stability theory combining with the linear matrix inequalities, so that the sufficient conditions for the leader-following consensus problem of multi-agent systems can be obtained. Numerical simulation results clearly show the effectiveness of the proposed controller.

With the continuous expansion of differential privacy and its applications, its application in the privacy protection field of trajectory data release has received extensive attention. However， most existing research methods use the Kmeans to cluster the trajectory, , which cannot guarantee the final convergence due to the fact that the trajectory datasets under differential privacy constraints are usually disturbed by noise. To addrss this, this paper proposes an orientation control-based differential privacy-preserving trajectory data publishing method. Firstly, a trajectory generalization algorithm based on SKmeans|| clustering is proposed, which updates the centroid via a direction control mechanism during iterative process of clustering, and designs a scoring function in the index mechanism to control the convergence of the centroid, such that the quality of high dimensional data clustering can be improved. Secondly, a trajectory data publishing algorithm based on bounded noise mechanism is designed, which improves the availability of trajectory data after publishing. Meanwhile, the bounded noise mechanism ensures the true count of the hidden trajectory. Finally, the effectiveness of the method proposed in this paper is evaluated by experiments.

Chronic obstructive pulmonary disease (COPD) is a common respiratory disease of the world, and the doctors need a lot of time to read the abdominal CT images for COPD pre-evaluation. To improve the pre-evaluation efficiency, a deep network based on slice correlation information was proposed for COPD auxiliary diagnosis. First, by using a grouping approach, the architecture of the deep network is divided into several network branches, each of which aims to extract the local intra-slice association information of the CT images. Then, the outputs from multiple network branches are integrated via a BiLSTM to extract the global inter-slice association information between the adjacent CT slices. To further improve the ability of local feature extraction for each network branch, the enhanced multi-headed convolutional attention is designed by embedding the ConvNeXt into the existing multi-headed convolutional attention. Experimental results show that the proposed deep network achieves promising effectiveness for CT image classification on auiliarily diagnose of COPD, and the accuracy, sensitivity, specificity and AUC of the proposed network reach to approximately 92.15%, 94.17%, 91.17% and 95.33%, respectively.

Na metal batteries are considered to be one of the most promising large-scale energy storage batteries due to their high theoretical specific capacity and low cost. However, the high reactivity of sodium metal can easily lead to problems such as instability of the solid electrolyte interface (SEI) film, uneven deposition of sodium, and dendrite growth. Here, an Al ₂O ₃ in-situ modified Al foil current collector (Al@Al ₂O ₃) was fabricated by a facile one-step calcination method to promote uniform Na deposition/stripping. During the discharge process, Al ₂O ₃ is sodiumified to form a Na-Al-O film with high ion conductivity, which not only stabilizes the electrode/electrolyte interface, but also regulates the nucleation behavior on the current collector surface, reducing the formation of nuclear energy barrier, improving ion mass transfer kinetics, and achieving uniform deposition of dendrite-free sodium and long cycle life. The results show that Al@Al ₂O ₃ can stably deposit/strip sodium for 50 times with an average Coulombic efficiency of 99.6% under 3 mA·cm ^-2/3 mAh·cm ^-2; and that the Na-Al@Al ₂O ₃‖Na-Al@Al ₂O ₃ symmetric battery can be cycled stably for 1000 h at 1 mA·cm ^-2 and 1 mAh·cm ^-2. Even at a high current density of 10 C, the NVP‖Na-Al@Al ₂O ₃ full cell can be cycled stably for 250 cycles with a high capacity retention of 94%.

Hepatobiliary stone disease is a common liver disease and has become the main cause of death from non-neoplastic biliary diseases in China, and it is important to achieve interpolated segmentation reconstruction between slices of hepatobiliary ducts. In this research, an end-to-end framework for super-resolution abdominal CT image processing is proposed based on deep correlation mechanism. The framework cascades an inter-slice interpolation network and a segmentation network, in which the ConvLSTM is introduced to enhance the extraction of high-dimensional feature information of hepatobiliary ducts between slices. A novel loss is designed by combining the loss of the interpolation network and the loss of the segmentation network. Experimental results show that the proposed framework is superior to the existing deep learning methods for the segmentation of hepatobiliary ducts, which is beneficial for the 3D reconstruction of hepatobiliary ducts.

In order to improve the performance and efficiency of image segmentation with multilevel threshold of two-dimensional entropy for practical industrial applications, this paper proposes a fast image segmentation method with multilevel threshold of two-dimensional entropy based on ISSA and integral graph. Firstly, we introduce and analyze the sparrow search algorithm (SSA). To address the shortcomings of SSA, such as poor global search ability and easy to fall into local optimal solution, we propose an improved sparrow search algorithm (ISSA) based on Gaussian perturbation strategy with linear decreasing variance and moving strategy with random step size. Then, we further introduce the integral graph method to reduce the calculation amount of the entropy, use the entropy as the fitness function of ISSA to search the optimal threshold, and propose a fast algorithm for image segmentation with multilevel threshold of two-dimensional entropy based on ISSA and integral graph. Finally, we compare the proposed method with the existing segmentation algorithms, and the experimental results show that the proposed method improves the segmentation efficiency of image segmentation with multilevel threshold of two-dimensional entropy in industrial application scenarios.

It is of great significance to develop an inexpensive and efficient catalyst to enhance the catalytic activity of oxidation coupling of alcohols with amines to imine. In this research, a series of manganese-iron bimetallic oxide catalysts have been prepared by a simple co-precipitation method, which were applied to the catalytic reaction of oxidation coupling of benzyl methanol with aniline to N-benzylideneaniline. The effects of different Fe/Mn feed ratios on the catalytic activity of the products have been explored. The Mn _0.5Fe _0.5O _x sample with a Fe/Mn feed ratio of 1:1 demonstrated the best catalytic activity, giving an aniline conversion rate of 74.7%, 99.9% selectivity and 74.6% yield of N-benzylidene, respectively. Through a variety of characterizations, Mn _0.5Fe _0.5O _x exhibited rich mesoporous structure and surface adsorbed oxygen species, as well as excellent oxidation ability. The high-efficiency catalyst synthesized in this work has great application potential in the preparation of imines by oxidation coupling of alcohols with amines.

Climate change caused by greenhouse gas emissions and aquatic eutrophication formed by nitrogen and phosphorus enrichment are the key issues that need to be solved urgently in the field of ecological environment in China. The agricultural sector is one of the main sources of carbon, nitrogen, and phosphorus emissions, and analyzing their characteristics and driving forces is of great significance to the implementation of China's pollution reduction and carbon reduction strategy. Based on the cross-sectional data of the agricultural sector in Guangdong province, the carbon emission model, nitrogen and phosphorus runoff model, and the LMDI driver model of agricultural sources in Guangdong province are constructed to analyze the evolution characteristics and driving factors of carbon, nitrogen and phosphorus emissions in each agricultural source sector at the provincial and county levels. The results showed that: (1) the carbon, nitrogen and phosphorus emissions from county agricultural sources in Guangdong province had significant spatial differences, and had the characteristics of concentration, correlation and homology in space; (2) From 1990 to 2021, the carbon and nitrogen emissions from agricultural sources showed heterogeneous changes, carbon emissions increased, nitrogen and phosphorus emissions decreased, and the nitrogen to phosphorus ratio showed an upward trend; (3) Per capita primary industry added value and unit primary industry added value emissions were the largest driving forces affecting the rise and decrease of agricultural carbon, nitrogen and phosphorus emissions in Guangdong province, respectively, and the ranking was different among different elements, indicating that there were differences in the driving factors controlling the change of agricultural source carbon and nitrogen emissions. The results of this study will provide decision-making support for the identification and collaborative regulation of key source areas of agricultural source carbon, nitrogen and phosphorus emissions in Guangdong province.

The medical named entity recognition aims to automatically identify and classify medical entities in electronic medical records, which plays a very important role in downstream tasks such as information retrieval and knowledge graph. Existing methods usually ignore the dependencies between entities. To address this, this paper proposes a gated attention unit-based model for Chinese medical named entity recognition. First, the proposed model uses the pre-training model MC-BERT to capture contextual information. Then, it uses the cross-attention and gated attention unit to enhance the interaction between entity query and contextual semantics, and further extract the dependency and correlation between entities. Finally, the proposed model uses the matching algorithm of bipartite graph to calculate the loss. This paper conducted experiments on three datasets, including the CMeEE, CMQNN, and MSRA. The experimental results show that the F ₁ values of the proposed model on the three datasets are 70.74%, 96.92%, and 95.53%, respectively, which outperforms other related models, demonstrating the effectiveness of the proposed model in Chinese medical named entity recognition task.

Due to the fact that most existing process similarity metrics focus on a single dimension of the process and lack comprehensive consideration of process information, the accuracy of process retrieval needs to be improved and the application focuses on a single scenario. In this paper, an efficient and multidimensional similarity measure of string sequences of weighted Petri nets is proposed based on the structural and behavioral information. First, the proposed method weights the event log information to Petri nets. Then, the proposed method converts the weighted Petri net model into a string sequence using breadth-first traversal, and further divides the sequence into a set of immediately adjacent variation pairs with weights and a structural sequence and calculates the similarity value separately. Finally, the similarity value between processes is obtained by weighting. The experimental results show that the metric has a high accuracy rate of metric is 99.51% with a low time complexity.

The energy-efficient caching strategy and computation offloading design of mobile edge computing (MEC) faces the double randomness challenges, which require to adapt to the time-varying wireless channel state and the dynamic task arrivals. This paper investigates a cache-enabled mobile edge computing system with dynamic tasks arriving at multiple wireless devices. By minimizing the system weighted sum energy over multiple time slots, we optimize the AP's task caching decision and MEC execution, the wireless devices’ local computing and tasks offloading under the caching capacity and computation causality, and the computation deadline constraints. The branch-and-bound (BnB) method is first presented to obtain the globally optimal solution to define the lower bound for practical schemes. Then, a relaxation-based scheme is proposed to efficiently achieve a near-optimal solution. Numerical results show that the proposed relaxation-based scheme achieves a closer performance to the optimal BnB scheme when compared to the benchmark schemes.

In order to deeply understand the physical properties of strained heterojunction quantum well structures and improve the design of wide-bandgap nitride semiconductor devices, in this paper, based on the six-band stress-dependent k·p Hamiltonian and the self-consistent Schrodinger-Poisson equation, the valence subband model of wurtzite GaN/AlN quantum well with polar (0001) , semi-polar ($10\bar 12 $) and non-polar ($10\bar 10 $) orientations under field confinement was established. The subband energy dispersion relations between GaN/AlN quantum well with different crystal orientations under biaxial and uniaxial stresses were also given. According to the influence of stress on the valence band structure of quantum well, the microcosmic physical relationship between stress and hole effective mass was studied comprehensively. The results show that the valence band structure heavily depends on the modification in crystal orientation. The biaxial stress has little effect on the improvement of effective mass. However, uniaxial compressive stress can obtain more holes in the region of low effective mass by reducing the energy in the vertical channel direction, such that the hole effective mass can be effectively reduced. And it is reduced by about 90% in the structure of different crystal orientations.

Currently, straight lines or arcs are commonly used to approximate the profile of the formed grinding wheel for gears, resulting in discontinuity and fluctuations in the profile curve of the grinding wheel and even changing the concavity and convexity of the original profile curve, limiting the precision of the gears processed by the dressed formed grinding wheel. Besides, the dressing program is cumbersome and the amount of data is large. Aiming at this problem, a method of using B-spline curve to fit the profile of the formed grinding wheel is proposed, which is convenient for the numerical control system of the gear grinding machine to use the spline interpolation function to dress the formed grinding wheel. The method first calculates the profile of the involute helical gear formed grinding wheel. Feature points are then extracted from the data points of the grinding wheel profile and fitted with a B-spline curve. The fitting errors of the non-feature points are calculated using differential evolution algorithm. The data point with the maximum fitting error is added to the feature points. This process is iteratively repeated until the generated B-spline curve satisfies the fitting error requirement, fitting the profile of formed grinding wheel with fewer control points while meeting the specified error requirements. Simulation results show that the method can effectively fit the profile of the formed grinding wheel and the fitting error can meet the specified requirements.

Emerging contaminants (ECs) are characterized by stable structures and low concentrations, making them difficult to remove completely using traditional wastewater treatment processes. ECs are posing potential risks to aquatic ecosystems and human health. Advanced oxidation processes (AOPs) can rapidly and effectively degrade persistent pollutants. However, for trace refractory ECs in real water matrices, AOPs require excessive oxidants or consume more energy, resulting in low cost-effectiveness of water treatment and even secondary pollution. Therefore, developing efficient and low-energy selective oxidation processes for treating trace ECs in water has practical significance. Targeted adsorption-transformation technology can effectively enhance the utilization of free radicals and efficiently remove trace ECs. The concept of advanced water purification processes is elaborated based on selective oxidation, with a focus on the technical characteristics and recent development of selective electrochemical adsorption-transformation technology to remove per- and polyfluoroalkyl substances (PFAS) from water. Finally, an outlook is provided on the future research directions and trends.