Traditional large-scale multi-objective optimization algorithms(LSMOEAs)encounter difficulties when dealing with sparse large-scale multi-objective optimization problems(SLM-OPs)where most decision variables are zero....Traditional large-scale multi-objective optimization algorithms(LSMOEAs)encounter difficulties when dealing with sparse large-scale multi-objective optimization problems(SLM-OPs)where most decision variables are zero.As a result,many algorithms use a two-layer encoding approach to optimize binary variable Mask and real variable Dec separately.Nevertheless,existing optimizers often focus on locating non-zero variable posi-tions to optimize the binary variables Mask.However,approxi-mating the sparse distribution of real Pareto optimal solutions does not necessarily mean that the objective function is optimized.In data mining,it is common to mine frequent itemsets appear-ing together in a dataset to reveal the correlation between data.Inspired by this,we propose a novel two-layer encoding learning swarm optimizer based on frequent itemsets(TELSO)to address these SLMOPs.TELSO mined the frequent terms of multiple particles with better target values to find mask combinations that can obtain better objective values for fast convergence.Experi-mental results on five real-world problems and eight benchmark sets demonstrate that TELSO outperforms existing state-of-the-art sparse large-scale multi-objective evolutionary algorithms(SLMOEAs)in terms of performance and convergence speed.展开更多
Retinal blood vessel segmentation is crucial for diagnosing ocular and cardiovascular diseases.Although the introduction of U-Net in 2015 by Olaf Ronneberger significantly advanced this field,yet issues like limited t...Retinal blood vessel segmentation is crucial for diagnosing ocular and cardiovascular diseases.Although the introduction of U-Net in 2015 by Olaf Ronneberger significantly advanced this field,yet issues like limited training data,imbalance data distribution,and inadequate feature extraction persist,hindering both the segmentation performance and optimal model generalization.Addressing these critical issues,the DEFFA-Unet is proposed featuring an additional encoder to process domain-invariant pre-processed inputs,thereby improving both richer feature encoding and enhanced model generalization.A feature filtering fusion module is developed to ensure the precise feature filtering and robust hybrid feature fusion.In response to the task-specific need for higher precision where false positives are very costly,traditional skip connections are replaced with the attention-guided feature reconstructing fusion module.Additionally,innovative data augmentation and balancing methods are proposed to counter data scarcity and distribution imbalance,further boosting the robustness and generalization of the model.With a comprehensive suite of evaluation metrics,extensive validations on four benchmark datasets(DRIVE,CHASEDB1,STARE,and HRF)and an SLO dataset(IOSTAR),demonstrate the proposed method’s superiority over both baseline and state-of-the-art models.Particularly the proposed method significantly outperforms the compared methods in cross-validation model generalization.展开更多
Despite its remarkable performance on natural images,the segment anything model(SAM)lacks domain-specific information in medical imaging.and faces the challenge of losing local multi-scale information in the encoding ...Despite its remarkable performance on natural images,the segment anything model(SAM)lacks domain-specific information in medical imaging.and faces the challenge of losing local multi-scale information in the encoding phase.This paper presents a medical image segmentation model based on SAM with a local multi-scale feature encoder(LMSFE-SAM)to address the issues above.Firstly,based on the SAM,a local multi-scale feature encoder is introduced to improve the representation of features within local receptive field,thereby supplying the Vision Transformer(ViT)branch in SAM with enriched local multi-scale contextual information.At the same time,a multiaxial Hadamard product module(MHPM)is incorporated into the local multi-scale feature encoder in a lightweight manner to reduce the quadratic complexity and noise interference.Subsequently,a cross-branch balancing adapter is designed to balance the local and global information between the local multi-scale feature encoder and the ViT encoder in SAM.Finally,to obtain smaller input image size and to mitigate overlapping in patch embeddings,the size of the input image is reduced from 1024×1024 pixels to 256×256 pixels,and a multidimensional information adaptation component is developed,which includes feature adapters,position adapters,and channel-spatial adapters.This component effectively integrates the information from small-sized medical images into SAM,enhancing its suitability for clinical deployment.The proposed model demonstrates an average enhancement ranging from 0.0387 to 0.3191 across six objective evaluation metrics on BUSI,DDTI,and TN3K datasets compared to eight other representative image segmentation models.This significantly enhances the performance of the SAM on medical images,providing clinicians with a powerful tool in clinical diagnosis.展开更多
Based on the synchronous joint gravity and magnetic inversion of single interface by Pilkington and the need of revealing Cenozoic and crystalline basement thickness in the new round of oil-gas exploration, we propose...Based on the synchronous joint gravity and magnetic inversion of single interface by Pilkington and the need of revealing Cenozoic and crystalline basement thickness in the new round of oil-gas exploration, we propose a joint gravity and magnetic inversion methodfor two-layer models by concentrating on the relationship between the change of thicknessI and position of the middle layer and anomaly and discuss the effects of the key parameters. Model tests and application to field data show the validity of this method.展开更多
A weak nonlinear model of a two-layer barotropic ocean with Rayleigh dissipation is built.The analytic asymptotic solution is derived in the mid-latitude stationary wind field,and the physical meaning of the correspon...A weak nonlinear model of a two-layer barotropic ocean with Rayleigh dissipation is built.The analytic asymptotic solution is derived in the mid-latitude stationary wind field,and the physical meaning of the corresponding problem is discussed.展开更多
Brain encoding and decoding via functional magnetic resonance imaging(fMRI)are two important aspects of visual perception neuroscience.Although previous researchers have made significant advances in brain encoding and...Brain encoding and decoding via functional magnetic resonance imaging(fMRI)are two important aspects of visual perception neuroscience.Although previous researchers have made significant advances in brain encoding and decoding models,existing methods still require improvement using advanced machine learning techniques.For example,traditional methods usually build the encoding and decoding models separately,and are prone to overfitting on a small dataset.In fact,effectively unifying the encoding and decoding procedures may allow for more accurate predictions.In this paper,we first review the existing encoding and decoding methods and discuss the potential advantages of a“bidirectional”modeling strategy.Next,we show that there are correspondences between deep neural networks and human visual streams in terms of the architecture and computational rules.Furthermore,deep generative models(e.g.,variational autoencoders(VAEs)and generative adversarial networks(GANs))have produced promising results in studies on brain encoding and decoding.Finally,we propose that the dual learning method,which was originally designed for machine translation tasks,could help to improve the performance of encoding and decoding models by leveraging large-scale unpaired data.展开更多
The coefficients embodied in a Boussinesq-type model are very important since they are determined to optimize the linear and nonlinear properties.In most conventional Boussinesq-type models,these coefficients are assi...The coefficients embodied in a Boussinesq-type model are very important since they are determined to optimize the linear and nonlinear properties.In most conventional Boussinesq-type models,these coefficients are assigned the specific values.As for the multi-layer Boussinesq-type models with the inclusion of the vertical velocity,however,the effect of the different values of these coefficients on linear and nonlinear performances has never been investigated yet.The present study focuses on a two-layer Boussinesq-type model with the highest spatial derivatives being 2 and theoretically and numerically examines the effect of the coefficient on model performance.Theoretical analysis show that different values for(0.13≤α≤0.25)do not have great effects on the high accuracy of the linear shoaling,linear phase celerity and even third-order nonlinearity for water depth range of 0<kh≤10(k is wave number and h is water depth).The corresponding errors using different values are restricted within 0.1%,0.1%and 1%for the linear shoaling amplitude,dispersion and nonlinear harmonics,respectively.Numerical tests including regular wave shoaling over mildly varying slope from deep to shallow water,regular wave propagation over submerged bar,bichromatic wave group and focusing wave propagation over deep water are conducted.The comparison between numerical results using different values of,experimental data and analytical solutions confirm the theoretical analysis.The flexibility and consistency of the two-layer Boussinesq-type model is therefore demonstrated theoretically and numerically.展开更多
When pycnocline thickness of ocean density is relatively small, density stratification can be well represented as a two-layer system. In this article, a depth integrated model of the two-layer fluid with constant dens...When pycnocline thickness of ocean density is relatively small, density stratification can be well represented as a two-layer system. In this article, a depth integrated model of the two-layer fluid with constant density is considered,and a variant of the edge-based non-hydrostatic numerical scheme is formulated. The resulting scheme is very efficient since it resolves the vertical fluid depth only in two layers. Despite using just two layers, the numerical dispersion is shown to agree with the analytical dispersion curves over a wide range of kd, where k is the wave number and d the water depth. The scheme was tested by simulating an interfacial solitary wave propagating over a flat bottom, as well as over a bottom step. On a laboratory scale, the formation of an interfacial wave is simulated,which also shows the interaction of wave with a triangular bathymetry. Then, a case study using the Lombok Strait topography is discussed, and the results show the development of an interfacial wave due to a strong current passing through a sill.展开更多
In Information Centric Networking(ICN)where content is the object of exchange,in-network caching is a unique functional feature with the ability to handle data storage and distribution in remote sensing satellite netw...In Information Centric Networking(ICN)where content is the object of exchange,in-network caching is a unique functional feature with the ability to handle data storage and distribution in remote sensing satellite networks.Setting up cache space at any node enables users to access data nearby,thus relieving the processing pressure on the servers.However,the existing caching strategies still suffer from the lack of global planning of cache contents and low utilization of cache resources due to the lack of fine-grained division of cache contents.To address the issues mentioned,a cooperative caching strategy(CSTL)for remote sensing satellite networks based on a two-layer caching model is proposed.The two-layer caching model is constructed by setting up separate cache spaces in the satellite network and the ground station.Probabilistic caching of popular contents in the region at the ground station to reduce the access delay of users.A content classification method based on hierarchical division is proposed in the satellite network,and differential probabilistic caching is employed for different levels of content.The cached content is also dynamically adjusted by analyzing the subsequent changes in the popularity of the cached content.In the two-layer caching model,ground stations and satellite networks collaboratively cache to achieve global planning of cache contents,rationalize the utilization of cache resources,and reduce the propagation delay of remote sensing data.Simulation results show that the CSTL strategy not only has a high cache hit ratio compared with other caching strategies but also effectively reduces user request delay and server load,which satisfies the timeliness requirement of remote sensing data transmission.展开更多
We investigated the relationship between chromophore concentrations in two-layered scattering media and the apparent chromophore concentrations measured with broadband optical spectroscopy in conjunction with commonly...We investigated the relationship between chromophore concentrations in two-layered scattering media and the apparent chromophore concentrations measured with broadband optical spectroscopy in conjunction with commonly used homogeneous medium inverse models.We used diffusion theory to generate optical data from a two-layered distribution of relevant tissue absorbers,namely,oxyhemoglobin,deoxyhemoglobin,water,and lipids,with a top-layer thickness in the range 1–15 mm.The generated data consisted of broadband continuous-wave(CW)diffuse reflectance in the wavelength range 650–1024 nm,and frequency-domain(FD)diffuse reflectance at 690 and 830 nm;two source-detector distances of 25 and 35mm were used to simulate a dual-slope technique.The data were inverted using diffusion theory for a semi-infinite homogeneous medium to generate reduced scattering coeffcients at 690 and 830nm(from FD data)and effective absorption spectra in the range 650–1024nm(from CW data).The absorption spectra were then converted into effective total concentration and oxygen saturation of hemoglobin,as well as water and lipid concentrations.For absolute values,it was found that the effective hemoglobin parameters are typically representative of the bottom layer,whereas water and lipid represent some average of the respective concentrations in the two layers.For concentration changes,lipid showed a significant cross-talk with other absorber concentrations,thus indicating that lipid dynamics obtained in these conditions may not be reliable.These systematic simulations of broadband spectroscopy of two-layered media provide guidance on how to interpret effective optical properties measured with similar instrumental setups under the assumption of medium homogeneity.展开更多
The two-layered nonwoven geotextile, which consists of a layer constructed with fine fibers for providing optimal filtration characteristics and another layer constructed with coarse fibers for providing the required ...The two-layered nonwoven geotextile, which consists of a layer constructed with fine fibers for providing optimal filtration characteristics and another layer constructed with coarse fibers for providing the required mechanical properties, is desirable for drainage and filtration system. Based on Darcy’s law and drag force theory, a mathematical model on vertical permeability coefficient of two-layered nonwoven geotextile is estabilished. Comparison with experimental results shows that the present model possesses 83.6% accuracy for needle-punched two-layered nonwoven geotextiles. And experimental results also show that with the increasing of needle density the vertical permeability coefficient of two-layered nonwoven geotextiless firstly decreases and then increases, reaching the smallest value at 470 p/cm2.展开更多
Digital twin(DT)modelling is a prerequisite for the successful application of DT technology in the power industry.However,traditional scene modelling methods are costly,time-consuming,focus on overall features and lac...Digital twin(DT)modelling is a prerequisite for the successful application of DT technology in the power industry.However,traditional scene modelling methods are costly,time-consuming,focus on overall features and lack real-time updates,hindering the interaction between DT models and physical power equipment scenes.Therefore,a scene DT modelling technique focusing on local features in risk areas and real-time updates is urgently needed.Herein,real-time modelling of the±800 kV converter transformer is achieved by improving the neural radiation field based on a hybrid attention mechanism and multiresolution hash encoding.Compared to traditional methods,modelling time is reduced from hours to 1 min without professional equipment or manual intervention.The model quality is more concerned with local features of risk areas in transformers while ensuring the overall scene,and the accuracy is improved by about 6%,realising the real-time modelling of transformers and the DT of scenes.展开更多
Damage detection in structures is performed via vibra-tion based structural identification. Modal information, such as fre-quencies and mode shapes, are widely used for structural dama-ge detection to indicate the hea...Damage detection in structures is performed via vibra-tion based structural identification. Modal information, such as fre-quencies and mode shapes, are widely used for structural dama-ge detection to indicate the health conditions of civil structures.The deep learning algorithm that works on a multiple layer neuralnetwork model termed as deep autoencoder is proposed to learnthe relationship between the modal information and structural stiff-ness parameters. This is achieved via dimension reduction of themodal information feature and a non-linear regression against thestructural stiffness parameters. Numerical tests on a symmetri-cal steel frame model are conducted to generate the data for thetraining and validation, and to demonstrate the efficiency of theproposed approach for vibration based structural damage detec-tion.展开更多
Research into the moisture transport processes in porous materials is primarily important for theoretical modelling and industrial applications in the design of energy saving buildings and living environments, etc. Ba...Research into the moisture transport processes in porous materials is primarily important for theoretical modelling and industrial applications in the design of energy saving buildings and living environments, etc. Based on experimental investigation, we propose new models which describe one-dimensional transport through one-layered uniform materials and dissimilar two-layered composites. Diffusivity as a function of moisture content is obtained through a Boltzman transformation, master curves, and combined numerical and regression techniques. Transport processes in one and two-layered composites are simulated on the basis of extended unsaturated Darcy’s Law using the finite element method (FEM). Simulation results show significantly different transport patterns of moisture profile when moisture migrates in different directions, and high agreement with experimental moisture profiles. Keywords Porous materials - moisture transport - two-layered composites - modelling and simulation Qingguo Wang graduated from Hebei Normal University, China, in 1985. He received the M.Sc. degree from Beijing Petroleum University in 1988 and the Ph.D. degree from the University of Luton, UK, in 2005. He is currently a Research Associate in the Department of Electrical Engineering and Electronics at the University of Liverpool, UK and an Associate Professor of Shijiazhuang Mechanical Engineering College, China. His research interests include measurement and control, mass and heat transportation, EMC, etc.Kemal Ahmet graduated in physics from the University of Leeds. Following the completion of his masters degree, he completed his Ph.D. at the University of London in the area of nuclear instrumentation in 1992. Until recently, he was a Principal Lecturer at the University of Luton, leading a research group in moisture instrumentation, measurement and monitoring. In 2004 he joined Medtronic, world leader in medical technology, and is currently working in the Neurologic Technologies division as a specialist in powered surgical instrumentation.Young Yue is a Principal Lecturer at the University of Luton, UK. He holds a B.Sc. in mechanical engineering from the Northeastern University, China, and a Ph.D. from Heriot-Watt University, UK. He is a chartered engineer and a member of the Institution of Mechanical Engineers, UK. Dr. Yue has been working in academia for 15 years following his 8 years of industrial experience. His main research interests are CAD/CAM, geometric modelling, virtual reality, and pattern recognition. He has over 70 publications in refereed books, journals and conferences.展开更多
The Earth's rotational normal modes depend on Earth model used, including the layer structures,principal inertia moments of different layers and the compliances. This study focuses on providing numerical solution ...The Earth's rotational normal modes depend on Earth model used, including the layer structures,principal inertia moments of different layers and the compliances. This study focuses on providing numerical solution of the rotational normal modes of a triaxial two-layered anelastic Earth model without external forces but with considering the complex forms of compliances and the electromagnetic coupling between the core and mantle. Based on the present knowledge of the Chandler wobble(CW) and Free Core Nutation(FCN), we provide a set of complete compliances which could be used for reference in further investigations. There are eight rotational normal mode solutions, four of which might exist in nature. However, in reality only two of these four solutions correspond to the present motion status of the prograde CW and the retrograde FCN. On one hand, our numerical calculations show that the periods and quality factors(Qs) of CW and FCN are respectively 434.90 and 429.86 mean solar days(d) and 76.56 and 23988.47 under frequency-dependent assumption, and the triaxiality prolongs CW about 0.01 d and has hardly effect on FCN. On the other hand, we analyze the sensibility of compliances and electromagnetic coupling parameter on the periods and Qs of CW and FCN and find the sensitive parameters with respect to them.展开更多
For realizing of long text information hiding and covert communication, a binary watermark sequence was obtained firstly from a text file and encoded by a redundant encoding method. Then, two neighboring blocks were s...For realizing of long text information hiding and covert communication, a binary watermark sequence was obtained firstly from a text file and encoded by a redundant encoding method. Then, two neighboring blocks were selected at each time from the Hilbert scanning sequence of carrier image blocks, and transformed by 1-level discrete wavelet transformation (DWT). And then the double block based JNDs (just noticeable difference) were calculated with a visual model. According to the different codes of each two watermark bits, the average values of two corresponding detail sub-bands were modified by using one of JNDs to hide information into carrier image. The experimental results show that the hidden information is invisible to human eyes, and the algorithm is robust to some common image processing operations. The conclusion is that the algorithm is effective and practical.展开更多
Sentence classification is the process of categorizing a sentence based on the context of the sentence.Sentence categorization requires more semantic highlights than other tasks,such as dependence parsing,which requir...Sentence classification is the process of categorizing a sentence based on the context of the sentence.Sentence categorization requires more semantic highlights than other tasks,such as dependence parsing,which requires more syntactic elements.Most existing strategies focus on the general semantics of a conversation without involving the context of the sentence,recognizing the progress and comparing impacts.An ensemble pre-trained language model was taken up here to classify the conversation sentences from the conversation corpus.The conversational sentences are classified into four categories:information,question,directive,and commission.These classification label sequences are for analyzing the conversation progress and predicting the pecking order of the conversation.Ensemble of Bidirectional Encoder for Representation of Transformer(BERT),Robustly Optimized BERT pretraining Approach(RoBERTa),Generative Pre-Trained Transformer(GPT),DistilBERT and Generalized Autoregressive Pretraining for Language Understanding(XLNet)models are trained on conversation corpus with hyperparameters.Hyperparameter tuning approach is carried out for better performance on sentence classification.This Ensemble of Pre-trained Language Models with a Hyperparameter Tuning(EPLM-HT)system is trained on an annotated conversation dataset.The proposed approach outperformed compared to the base BERT,GPT,DistilBERT and XLNet transformer models.The proposed ensemble model with the fine-tuned parameters achieved an F1_score of 0.88.展开更多
基金supported by the Scientific Research Project of Xiang Jiang Lab(22XJ02003)the University Fundamental Research Fund(23-ZZCX-JDZ-28)+5 种基金the National Science Fund for Outstanding Young Scholars(62122093)the National Natural Science Foundation of China(72071205)the Hunan Graduate Research Innovation Project(ZC23112101-10)the Hunan Natural Science Foundation Regional Joint Project(2023JJ50490)the Science and Technology Project for Young and Middle-aged Talents of Hunan(2023TJ-Z03)the Science and Technology Innovation Program of Humnan Province(2023RC1002)。
文摘Traditional large-scale multi-objective optimization algorithms(LSMOEAs)encounter difficulties when dealing with sparse large-scale multi-objective optimization problems(SLM-OPs)where most decision variables are zero.As a result,many algorithms use a two-layer encoding approach to optimize binary variable Mask and real variable Dec separately.Nevertheless,existing optimizers often focus on locating non-zero variable posi-tions to optimize the binary variables Mask.However,approxi-mating the sparse distribution of real Pareto optimal solutions does not necessarily mean that the objective function is optimized.In data mining,it is common to mine frequent itemsets appear-ing together in a dataset to reveal the correlation between data.Inspired by this,we propose a novel two-layer encoding learning swarm optimizer based on frequent itemsets(TELSO)to address these SLMOPs.TELSO mined the frequent terms of multiple particles with better target values to find mask combinations that can obtain better objective values for fast convergence.Experi-mental results on five real-world problems and eight benchmark sets demonstrate that TELSO outperforms existing state-of-the-art sparse large-scale multi-objective evolutionary algorithms(SLMOEAs)in terms of performance and convergence speed.
文摘Retinal blood vessel segmentation is crucial for diagnosing ocular and cardiovascular diseases.Although the introduction of U-Net in 2015 by Olaf Ronneberger significantly advanced this field,yet issues like limited training data,imbalance data distribution,and inadequate feature extraction persist,hindering both the segmentation performance and optimal model generalization.Addressing these critical issues,the DEFFA-Unet is proposed featuring an additional encoder to process domain-invariant pre-processed inputs,thereby improving both richer feature encoding and enhanced model generalization.A feature filtering fusion module is developed to ensure the precise feature filtering and robust hybrid feature fusion.In response to the task-specific need for higher precision where false positives are very costly,traditional skip connections are replaced with the attention-guided feature reconstructing fusion module.Additionally,innovative data augmentation and balancing methods are proposed to counter data scarcity and distribution imbalance,further boosting the robustness and generalization of the model.With a comprehensive suite of evaluation metrics,extensive validations on four benchmark datasets(DRIVE,CHASEDB1,STARE,and HRF)and an SLO dataset(IOSTAR),demonstrate the proposed method’s superiority over both baseline and state-of-the-art models.Particularly the proposed method significantly outperforms the compared methods in cross-validation model generalization.
基金supported by Natural Science Foundation Programme of Gansu Province(No.24JRRA231)National Natural Science Foundation of China(No.62061023)Gansu Provincial Science and Technology Plan Key Research and Development Program Project(No.24YFFA024).
文摘Despite its remarkable performance on natural images,the segment anything model(SAM)lacks domain-specific information in medical imaging.and faces the challenge of losing local multi-scale information in the encoding phase.This paper presents a medical image segmentation model based on SAM with a local multi-scale feature encoder(LMSFE-SAM)to address the issues above.Firstly,based on the SAM,a local multi-scale feature encoder is introduced to improve the representation of features within local receptive field,thereby supplying the Vision Transformer(ViT)branch in SAM with enriched local multi-scale contextual information.At the same time,a multiaxial Hadamard product module(MHPM)is incorporated into the local multi-scale feature encoder in a lightweight manner to reduce the quadratic complexity and noise interference.Subsequently,a cross-branch balancing adapter is designed to balance the local and global information between the local multi-scale feature encoder and the ViT encoder in SAM.Finally,to obtain smaller input image size and to mitigate overlapping in patch embeddings,the size of the input image is reduced from 1024×1024 pixels to 256×256 pixels,and a multidimensional information adaptation component is developed,which includes feature adapters,position adapters,and channel-spatial adapters.This component effectively integrates the information from small-sized medical images into SAM,enhancing its suitability for clinical deployment.The proposed model demonstrates an average enhancement ranging from 0.0387 to 0.3191 across six objective evaluation metrics on BUSI,DDTI,and TN3K datasets compared to eight other representative image segmentation models.This significantly enhances the performance of the SAM on medical images,providing clinicians with a powerful tool in clinical diagnosis.
基金Supported by the National Natural Science Foundation of China(Grant No.40674063)National Hi-tech Research and Development Program of China(863Program)(Grant No.2006AA09Z311)
文摘Based on the synchronous joint gravity and magnetic inversion of single interface by Pilkington and the need of revealing Cenozoic and crystalline basement thickness in the new round of oil-gas exploration, we propose a joint gravity and magnetic inversion methodfor two-layer models by concentrating on the relationship between the change of thicknessI and position of the middle layer and anomaly and discuss the effects of the key parameters. Model tests and application to field data show the validity of this method.
基金Project supported by the National Basic Research Program of China (Grant No. 2011CB403501)the National Natural Science Foundation of China (GrantNos. 41175058,41275062,and 11202106)
文摘A weak nonlinear model of a two-layer barotropic ocean with Rayleigh dissipation is built.The analytic asymptotic solution is derived in the mid-latitude stationary wind field,and the physical meaning of the corresponding problem is discussed.
基金This work was supported by the National Key Research and Development Program of China(2018YFC2001302)National Natural Science Foundation of China(91520202)+2 种基金Chinese Academy of Sciences Scientific Equipment Development Project(YJKYYQ20170050)Beijing Municipal Science and Technology Commission(Z181100008918010)Youth Innovation Promotion Association of Chinese Academy of Sciences,and Strategic Priority Research Program of Chinese Academy of Sciences(XDB32040200).
文摘Brain encoding and decoding via functional magnetic resonance imaging(fMRI)are two important aspects of visual perception neuroscience.Although previous researchers have made significant advances in brain encoding and decoding models,existing methods still require improvement using advanced machine learning techniques.For example,traditional methods usually build the encoding and decoding models separately,and are prone to overfitting on a small dataset.In fact,effectively unifying the encoding and decoding procedures may allow for more accurate predictions.In this paper,we first review the existing encoding and decoding methods and discuss the potential advantages of a“bidirectional”modeling strategy.Next,we show that there are correspondences between deep neural networks and human visual streams in terms of the architecture and computational rules.Furthermore,deep generative models(e.g.,variational autoencoders(VAEs)and generative adversarial networks(GANs))have produced promising results in studies on brain encoding and decoding.Finally,we propose that the dual learning method,which was originally designed for machine translation tasks,could help to improve the performance of encoding and decoding models by leveraging large-scale unpaired data.
基金supported by the National Natural Science Foundation of China(Grant Nos.51779022,51809053,and 51579034)the Innovation Team Project of Estuary and Coast Protection and Management(Grant No.Y220013)the Open Project Fund of State Key Laboratory of Coastal and Offshore Engineering,Dalian University of Technology(Grant No.LP19015).
文摘The coefficients embodied in a Boussinesq-type model are very important since they are determined to optimize the linear and nonlinear properties.In most conventional Boussinesq-type models,these coefficients are assigned the specific values.As for the multi-layer Boussinesq-type models with the inclusion of the vertical velocity,however,the effect of the different values of these coefficients on linear and nonlinear performances has never been investigated yet.The present study focuses on a two-layer Boussinesq-type model with the highest spatial derivatives being 2 and theoretically and numerically examines the effect of the coefficient on model performance.Theoretical analysis show that different values for(0.13≤α≤0.25)do not have great effects on the high accuracy of the linear shoaling,linear phase celerity and even third-order nonlinearity for water depth range of 0<kh≤10(k is wave number and h is water depth).The corresponding errors using different values are restricted within 0.1%,0.1%and 1%for the linear shoaling amplitude,dispersion and nonlinear harmonics,respectively.Numerical tests including regular wave shoaling over mildly varying slope from deep to shallow water,regular wave propagation over submerged bar,bichromatic wave group and focusing wave propagation over deep water are conducted.The comparison between numerical results using different values of,experimental data and analytical solutions confirm the theoretical analysis.The flexibility and consistency of the two-layer Boussinesq-type model is therefore demonstrated theoretically and numerically.
基金financially supported by the Institut Teknologi Bandung Research(Grant No.107a/I1.C01/PL/2017)
文摘When pycnocline thickness of ocean density is relatively small, density stratification can be well represented as a two-layer system. In this article, a depth integrated model of the two-layer fluid with constant density is considered,and a variant of the edge-based non-hydrostatic numerical scheme is formulated. The resulting scheme is very efficient since it resolves the vertical fluid depth only in two layers. Despite using just two layers, the numerical dispersion is shown to agree with the analytical dispersion curves over a wide range of kd, where k is the wave number and d the water depth. The scheme was tested by simulating an interfacial solitary wave propagating over a flat bottom, as well as over a bottom step. On a laboratory scale, the formation of an interfacial wave is simulated,which also shows the interaction of wave with a triangular bathymetry. Then, a case study using the Lombok Strait topography is discussed, and the results show the development of an interfacial wave due to a strong current passing through a sill.
基金This research was funded by the National Natural Science Foundation of China(No.U21A20451)the Science and Technology Planning Project of Jilin Province(No.20200401105GX)the China University Industry University Research Innovation Fund(No.2021FNA01003).
文摘In Information Centric Networking(ICN)where content is the object of exchange,in-network caching is a unique functional feature with the ability to handle data storage and distribution in remote sensing satellite networks.Setting up cache space at any node enables users to access data nearby,thus relieving the processing pressure on the servers.However,the existing caching strategies still suffer from the lack of global planning of cache contents and low utilization of cache resources due to the lack of fine-grained division of cache contents.To address the issues mentioned,a cooperative caching strategy(CSTL)for remote sensing satellite networks based on a two-layer caching model is proposed.The two-layer caching model is constructed by setting up separate cache spaces in the satellite network and the ground station.Probabilistic caching of popular contents in the region at the ground station to reduce the access delay of users.A content classification method based on hierarchical division is proposed in the satellite network,and differential probabilistic caching is employed for different levels of content.The cached content is also dynamically adjusted by analyzing the subsequent changes in the popularity of the cached content.In the two-layer caching model,ground stations and satellite networks collaboratively cache to achieve global planning of cache contents,rationalize the utilization of cache resources,and reduce the propagation delay of remote sensing data.Simulation results show that the CSTL strategy not only has a high cache hit ratio compared with other caching strategies but also effectively reduces user request delay and server load,which satisfies the timeliness requirement of remote sensing data transmission.
基金supported by National Institutes of Health(Nos.R01 NS095334,R01 EB029414).
文摘We investigated the relationship between chromophore concentrations in two-layered scattering media and the apparent chromophore concentrations measured with broadband optical spectroscopy in conjunction with commonly used homogeneous medium inverse models.We used diffusion theory to generate optical data from a two-layered distribution of relevant tissue absorbers,namely,oxyhemoglobin,deoxyhemoglobin,water,and lipids,with a top-layer thickness in the range 1–15 mm.The generated data consisted of broadband continuous-wave(CW)diffuse reflectance in the wavelength range 650–1024 nm,and frequency-domain(FD)diffuse reflectance at 690 and 830 nm;two source-detector distances of 25 and 35mm were used to simulate a dual-slope technique.The data were inverted using diffusion theory for a semi-infinite homogeneous medium to generate reduced scattering coeffcients at 690 and 830nm(from FD data)and effective absorption spectra in the range 650–1024nm(from CW data).The absorption spectra were then converted into effective total concentration and oxygen saturation of hemoglobin,as well as water and lipid concentrations.For absolute values,it was found that the effective hemoglobin parameters are typically representative of the bottom layer,whereas water and lipid represent some average of the respective concentrations in the two layers.For concentration changes,lipid showed a significant cross-talk with other absorber concentrations,thus indicating that lipid dynamics obtained in these conditions may not be reliable.These systematic simulations of broadband spectroscopy of two-layered media provide guidance on how to interpret effective optical properties measured with similar instrumental setups under the assumption of medium homogeneity.
文摘The two-layered nonwoven geotextile, which consists of a layer constructed with fine fibers for providing optimal filtration characteristics and another layer constructed with coarse fibers for providing the required mechanical properties, is desirable for drainage and filtration system. Based on Darcy’s law and drag force theory, a mathematical model on vertical permeability coefficient of two-layered nonwoven geotextile is estabilished. Comparison with experimental results shows that the present model possesses 83.6% accuracy for needle-punched two-layered nonwoven geotextiles. And experimental results also show that with the increasing of needle density the vertical permeability coefficient of two-layered nonwoven geotextiless firstly decreases and then increases, reaching the smallest value at 470 p/cm2.
基金National Key Research and Development Program of China,Grant/Award Number:2021YFB2401700。
文摘Digital twin(DT)modelling is a prerequisite for the successful application of DT technology in the power industry.However,traditional scene modelling methods are costly,time-consuming,focus on overall features and lack real-time updates,hindering the interaction between DT models and physical power equipment scenes.Therefore,a scene DT modelling technique focusing on local features in risk areas and real-time updates is urgently needed.Herein,real-time modelling of the±800 kV converter transformer is achieved by improving the neural radiation field based on a hybrid attention mechanism and multiresolution hash encoding.Compared to traditional methods,modelling time is reduced from hours to 1 min without professional equipment or manual intervention.The model quality is more concerned with local features of risk areas in transformers while ensuring the overall scene,and the accuracy is improved by about 6%,realising the real-time modelling of transformers and the DT of scenes.
文摘Damage detection in structures is performed via vibra-tion based structural identification. Modal information, such as fre-quencies and mode shapes, are widely used for structural dama-ge detection to indicate the health conditions of civil structures.The deep learning algorithm that works on a multiple layer neuralnetwork model termed as deep autoencoder is proposed to learnthe relationship between the modal information and structural stiff-ness parameters. This is achieved via dimension reduction of themodal information feature and a non-linear regression against thestructural stiffness parameters. Numerical tests on a symmetri-cal steel frame model are conducted to generate the data for thetraining and validation, and to demonstrate the efficiency of theproposed approach for vibration based structural damage detec-tion.
文摘Research into the moisture transport processes in porous materials is primarily important for theoretical modelling and industrial applications in the design of energy saving buildings and living environments, etc. Based on experimental investigation, we propose new models which describe one-dimensional transport through one-layered uniform materials and dissimilar two-layered composites. Diffusivity as a function of moisture content is obtained through a Boltzman transformation, master curves, and combined numerical and regression techniques. Transport processes in one and two-layered composites are simulated on the basis of extended unsaturated Darcy’s Law using the finite element method (FEM). Simulation results show significantly different transport patterns of moisture profile when moisture migrates in different directions, and high agreement with experimental moisture profiles. Keywords Porous materials - moisture transport - two-layered composites - modelling and simulation Qingguo Wang graduated from Hebei Normal University, China, in 1985. He received the M.Sc. degree from Beijing Petroleum University in 1988 and the Ph.D. degree from the University of Luton, UK, in 2005. He is currently a Research Associate in the Department of Electrical Engineering and Electronics at the University of Liverpool, UK and an Associate Professor of Shijiazhuang Mechanical Engineering College, China. His research interests include measurement and control, mass and heat transportation, EMC, etc.Kemal Ahmet graduated in physics from the University of Leeds. Following the completion of his masters degree, he completed his Ph.D. at the University of London in the area of nuclear instrumentation in 1992. Until recently, he was a Principal Lecturer at the University of Luton, leading a research group in moisture instrumentation, measurement and monitoring. In 2004 he joined Medtronic, world leader in medical technology, and is currently working in the Neurologic Technologies division as a specialist in powered surgical instrumentation.Young Yue is a Principal Lecturer at the University of Luton, UK. He holds a B.Sc. in mechanical engineering from the Northeastern University, China, and a Ph.D. from Heriot-Watt University, UK. He is a chartered engineer and a member of the Institution of Mechanical Engineers, UK. Dr. Yue has been working in academia for 15 years following his 8 years of industrial experience. His main research interests are CAD/CAM, geometric modelling, virtual reality, and pattern recognition. He has over 70 publications in refereed books, journals and conferences.
基金supported by the NSFC (grant Nos. 41631072, 41721003, 41874023, 41574007, and 41429401)the Discipline Innovative Engineering Plan of Modern Geodesy and Geodynamics (grant No. B17033)the DAAD Thematic Network Project (grant No. 57173947)
文摘The Earth's rotational normal modes depend on Earth model used, including the layer structures,principal inertia moments of different layers and the compliances. This study focuses on providing numerical solution of the rotational normal modes of a triaxial two-layered anelastic Earth model without external forces but with considering the complex forms of compliances and the electromagnetic coupling between the core and mantle. Based on the present knowledge of the Chandler wobble(CW) and Free Core Nutation(FCN), we provide a set of complete compliances which could be used for reference in further investigations. There are eight rotational normal mode solutions, four of which might exist in nature. However, in reality only two of these four solutions correspond to the present motion status of the prograde CW and the retrograde FCN. On one hand, our numerical calculations show that the periods and quality factors(Qs) of CW and FCN are respectively 434.90 and 429.86 mean solar days(d) and 76.56 and 23988.47 under frequency-dependent assumption, and the triaxiality prolongs CW about 0.01 d and has hardly effect on FCN. On the other hand, we analyze the sensibility of compliances and electromagnetic coupling parameter on the periods and Qs of CW and FCN and find the sensitive parameters with respect to them.
文摘For realizing of long text information hiding and covert communication, a binary watermark sequence was obtained firstly from a text file and encoded by a redundant encoding method. Then, two neighboring blocks were selected at each time from the Hilbert scanning sequence of carrier image blocks, and transformed by 1-level discrete wavelet transformation (DWT). And then the double block based JNDs (just noticeable difference) were calculated with a visual model. According to the different codes of each two watermark bits, the average values of two corresponding detail sub-bands were modified by using one of JNDs to hide information into carrier image. The experimental results show that the hidden information is invisible to human eyes, and the algorithm is robust to some common image processing operations. The conclusion is that the algorithm is effective and practical.
文摘Sentence classification is the process of categorizing a sentence based on the context of the sentence.Sentence categorization requires more semantic highlights than other tasks,such as dependence parsing,which requires more syntactic elements.Most existing strategies focus on the general semantics of a conversation without involving the context of the sentence,recognizing the progress and comparing impacts.An ensemble pre-trained language model was taken up here to classify the conversation sentences from the conversation corpus.The conversational sentences are classified into four categories:information,question,directive,and commission.These classification label sequences are for analyzing the conversation progress and predicting the pecking order of the conversation.Ensemble of Bidirectional Encoder for Representation of Transformer(BERT),Robustly Optimized BERT pretraining Approach(RoBERTa),Generative Pre-Trained Transformer(GPT),DistilBERT and Generalized Autoregressive Pretraining for Language Understanding(XLNet)models are trained on conversation corpus with hyperparameters.Hyperparameter tuning approach is carried out for better performance on sentence classification.This Ensemble of Pre-trained Language Models with a Hyperparameter Tuning(EPLM-HT)system is trained on an annotated conversation dataset.The proposed approach outperformed compared to the base BERT,GPT,DistilBERT and XLNet transformer models.The proposed ensemble model with the fine-tuned parameters achieved an F1_score of 0.88.
