Seismic time-frequency(TF)transforms are essential tools in reservoir interpretation and signal processing,particularly for characterizing frequency variations in non-stationary seismic data.Recently,sparse TF trans-f...Seismic time-frequency(TF)transforms are essential tools in reservoir interpretation and signal processing,particularly for characterizing frequency variations in non-stationary seismic data.Recently,sparse TF trans-forms,which leverage sparse coding(SC),have gained significant attention in the geosciences due to their ability to achieve high TF resolution.However,the iterative approaches typically employed in sparse TF transforms are computationally intensive,making them impractical for real seismic data analysis.To address this issue,we propose an interpretable convolutional sparse coding(CSC)network to achieve high TF resolution.The proposed model is generated based on the traditional short-time Fourier transform(STFT)transform and a modified UNet,named ULISTANet.In this design,we replace the conventional convolutional layers of the UNet with learnable iterative shrinkage thresholding algorithm(LISTA)blocks,a specialized form of CSC.The LISTA block,which evolves from the traditional iterative shrinkage thresholding algorithm(ISTA),is optimized for extracting sparse features more effectively.Furthermore,we create a synthetic dataset featuring complex frequency-modulated signals to train ULISTANet.Finally,the proposed method’s performance is subsequently validated using both synthetic and field data,demonstrating its potential for enhanced seismic data analysis.展开更多
The machinery fault signal is a typical non-Gaussian and non-stationary process. The fault signal can be described by SaS distribution model because of the presence of impulses.Time-frequency distribution is a useful ...The machinery fault signal is a typical non-Gaussian and non-stationary process. The fault signal can be described by SaS distribution model because of the presence of impulses.Time-frequency distribution is a useful tool to extract helpful information of the machinery fault signal. Various fractional lower order(FLO) time-frequency distribution methods have been proposed based on fractional lower order statistics, which include fractional lower order short time Fourier transform(FLO-STFT), fractional lower order Wigner-Ville distributions(FLO-WVDs), fractional lower order Cohen class time-frequency distributions(FLO-CDs), fractional lower order adaptive kernel time-frequency distributions(FLO-AKDs) and adaptive fractional lower order time-frequency auto-regressive moving average(FLO-TFARMA) model time-frequency representation method.The methods and the exiting methods based on second order statistics in SaS distribution environments are compared, simulation results show that the new methods have better performances than the existing methods. The advantages and disadvantages of the improved time-frequency methods have been summarized.Last, the new methods are applied to analyze the outer race fault signals, the results illustrate their good performances.展开更多
A localized parametric time-sheared Gabor atom is derived by convolving a linear frequency modulated factor, modulating in frequency and translating in time to a dilated Gaussian function, which is the generalization ...A localized parametric time-sheared Gabor atom is derived by convolving a linear frequency modulated factor, modulating in frequency and translating in time to a dilated Gaussian function, which is the generalization of Gabor atom and is more delicate for matching most of the signals encountered in practice, especially for those having frequency dispersion characteristics. The time-frequency distribution of this atom concentrates in its time center and frequency center along energy curve, with the curve being oblique to a certain extent along the time axis. A novel parametric adaptive time-frequency distribution based on a set of the derived atoms is then proposed using a adaptive signal subspace decomposition method in frequency domain, which is non-negative time-frequency energy distribution and free of cross-term interference for multicomponent signals. The results of numerical simulation manifest the effectiveness of the approach in time-frequency representation and signal de-noising processing.展开更多
This paper proposes a new method for estimating the parameter of maneuvering targets based on sparse time-frequency transform in over-the-horizon radar（OTHR）. In this method, the sparse time-frequency distribution o...This paper proposes a new method for estimating the parameter of maneuvering targets based on sparse time-frequency transform in over-the-horizon radar（OTHR）. In this method, the sparse time-frequency distribution of the radar echo is obtained by solving a sparse optimization problem based on the short-time Fourier transform. Then Hough transform is employed to estimate the parameter of the targets. The proposed algorithm has the following advantages： Compared with the Wigner-Hough transform method, the computational complexity of the sparse optimization is low due to the application of fast Fourier transform（FFT）. And the computational cost of Hough transform is also greatly reduced because of the sparsity of the time-frequency distribution. Compared with the high order ambiguity function（HAF） method, the proposed method improves in terms of precision and robustness to noise. Simulation results show that compared with the HAF method, the required SNR and relative mean square error are 8 dB lower and 50 dB lower respectively in the proposed method. While processing the field experiment data, the execution time of Hough transform in the proposed method is only 4% of the Wigner-Hough transform method.展开更多
A new time-frequency representation called Dopplerlet transform, which uses the dilated, translated and modulated windowed Doppler signals as its basis functions, is proposed, and the Fourier transform, short-time Fou...A new time-frequency representation called Dopplerlet transform, which uses the dilated, translated and modulated windowed Doppler signals as its basis functions, is proposed, and the Fourier transform, short-time Fourier transform (including Gabor transform), wavelet transform, and chirplet transform are formulated in one framework of Dopplerlet transform accordingly.It is proved that the matching pursuits based on Dopplerlet basis functions are convergent, and that the energy of residual signals yielded in the decomposition process decays exponentially. Simulation results show that the matching pursuits with Dopplerlet basis functions can characterize compactly a nonstationary signal.展开更多
Let F_(1)be the virtual field consisting of one element and(Q,I)a string pair.In this paper,we study the representations of string pairs over the virtual field F_(1).It is proved that an indecomposable F_(1)-represent...Let F_(1)be the virtual field consisting of one element and(Q,I)a string pair.In this paper,we study the representations of string pairs over the virtual field F_(1).It is proved that an indecomposable F_(1)-representation is either a string representation or a band representation by using the coefficient quivers.It is worth noting that for a given band and a positive integer,there exists a unique band representation up to isomorphism.展开更多
As a new research direction in contemporary cognitive science,predictive processing surpasses traditional computational representation and embodied cognition and has emerged as a new paradigm in cognitive science rese...As a new research direction in contemporary cognitive science,predictive processing surpasses traditional computational representation and embodied cognition and has emerged as a new paradigm in cognitive science research.The predictive processing theory advocates that the brain is a hierarchical predictive model based on Bayesian inference,and its purpose is to minimize the difference between the predicted world and the actual world,so as to minimize the prediction error.Predictive processing is therefore essentially a context-dependent model representation,an adaptive representational system designed to achieve its cognitive goals through the minimization of prediction error.展开更多
In global navigation satellite system denial environment,cross-view geo-localization based on image retrieval presents an exceedingly critical visual localization solution for Unmanned Aerial Vehicle(UAV)systems.The e...In global navigation satellite system denial environment,cross-view geo-localization based on image retrieval presents an exceedingly critical visual localization solution for Unmanned Aerial Vehicle(UAV)systems.The essence of cross-view geo-localization resides in matching images containing the same geographical targets from disparate platforms,such as UAV-view and satellite-view images.However,images of the same geographical targets may suffer from occlusions and geometric distortions due to variations in the capturing platform,view,and timing.The existing methods predominantly extract features by segmenting feature maps,which overlook the holistic semantic distribution and structural information of objects,resulting in loss of image information.To address these challenges,dilated neighborhood attention Transformer is employed as the feature extraction backbone,and Multi-feature representations based on Multi-scale Hierarchical Contextual Aggregation(MMHCA)is proposed.In the proposed MMHCA method,the multiscale hierarchical contextual aggregation method is utilized to extract contextual information from local to global across various granularity levels,establishing feature associations of contextual information with global and local information in the image.Subsequently,the multi-feature representations method is utilized to obtain rich discriminative feature information,bolstering the robustness of model in scenarios characterized by positional shifts,varying distances,and scale ambiguities.Comprehensive experiments conducted on the extensively utilized University-1652 and SUES-200 benchmarks indicate that the MMHCA method surpasses the existing techniques.showing outstanding results in UAV localization and navigation.展开更多
Establishing and maintaining protected areas is a pivotal strategy for attaining the post-2020 biodiversity target. The conservation objectives of protected areas have shifted from a narrow emphasis on biodiversity to...Establishing and maintaining protected areas is a pivotal strategy for attaining the post-2020 biodiversity target. The conservation objectives of protected areas have shifted from a narrow emphasis on biodiversity to encompass broader considerations such as ecosystem stability, community resilience to climate change, and enhancement of human well-being. Given these multifaceted objectives, it is imperative to judiciously allocate resources to effectively conserve biodiversity by identifying strategically significant areas for conservation, particularly for mountainous areas. In this study, we evaluated the representativeness of the protected area network in the Qin ling Mountains concerning species diversity, ecosystem services, climate stability and ecological stability. The results indicate that some of the ecological indicators are spatially correlated with topographic gradient effects. The conservation priority areas predominantly lie in the northern foothills, the southeastern, and southwestern parts of the Qinling Mountain with areas concentrated at altitudes between 1,500-2,000 m and slopes between 40°-50° as hotspots. The conservation priority areas identified through the framework of inclusive conservation optimization account for 22.9 % of the Qinling Mountain. Existing protected areas comprise only 6.1 % of the Qinling Mountain and 13.18 % of the conservation priority areas. This will play an important role in achiev ing sustainable development in the region and in meeting the post-2020 biodiversity target. The framework can advance the different objectives of achieving a quadruple win and can also be extended to other regions.展开更多
Physics-informed neural networks(PINNs)are promising to replace conventional mesh-based partial tial differen-equation(PDE)solvers by offering more accurate and flexible PDE solutions.However,PINNs are hampered by the...Physics-informed neural networks(PINNs)are promising to replace conventional mesh-based partial tial differen-equation(PDE)solvers by offering more accurate and flexible PDE solutions.However,PINNs are hampered by the relatively slow convergence and the need to perform additional,potentially expensive training for new PDE parameters.To solve this limitation,we introduce LatentPINN,a framework that utilizes latent representations of the PDE parameters as additional(to the coordinates)inputs into PINNs and allows for training over the distribution of these parameters.Motivated by the recent progress on generative models,we promote using latent diffusion models to learn compressed latent representations of the distribution of PDE parameters as they act as input parameters for NN functional solutions.We use a two-stage training scheme in which,in the first stage,we learn the latent representations for the distribution of PDE parameters.In the second stage,we train a physics-informed neural network over inputs given by randomly drawn samples from the coordinate space within the solution domain and samples from the learned latent representation of the PDE parameters.Considering their importance in capturing evolving interfaces and fronts in various fields,we test the approach on a class of level set equations given,for example,by the nonlinear Eikonal equation.We share results corresponding to three Eikonal parameters(velocity models)sets.The proposed method performs well on new phase velocity models without the need for any additional training.展开更多
Considering that the algorithm accuracy of the traditional sparse representation models is not high under the influence of multiple complex environmental factors,this study focuses on the improvement of feature extrac...Considering that the algorithm accuracy of the traditional sparse representation models is not high under the influence of multiple complex environmental factors,this study focuses on the improvement of feature extraction and model construction.Firstly,the convolutional neural network(CNN)features of the face are extracted by the trained deep learning network.Next,the steady-state and dynamic classifiers for face recognition are constructed based on the CNN features and Haar features respectively,with two-stage sparse representation introduced in the process of constructing the steady-state classifier and the feature templates with high reliability are dynamically selected as alternative templates from the sparse representation template dictionary constructed using the CNN features.Finally,the results of face recognition are given based on the classification results of the steady-state classifier and the dynamic classifier together.Based on this,the feature weights of the steady-state classifier template are adjusted in real time and the dictionary set is dynamically updated to reduce the probability of irrelevant features entering the dictionary set.The average recognition accuracy of this method is 94.45%on the CMU PIE face database and 96.58%on the AR face database,which is significantly improved compared with that of the traditional face recognition methods.展开更多
Clustering is a pivotal data analysis method for deciphering the charge transport properties of single molecules in break junction experiments.However,given the high dimensionality and variability of the data,feature ...Clustering is a pivotal data analysis method for deciphering the charge transport properties of single molecules in break junction experiments.However,given the high dimensionality and variability of the data,feature extraction remains a bottleneck in the development of efficient clustering methods.In this regard,extensive research over the past two decades has focused on feature engineering and dimensionality reduction in break junction conductance.However,extracting highly relevant features without expert knowledge remains an unresolved challenge.To address this issue,we propose a deep clustering method driven by task-oriented representation learning(CTRL)in which the clustering module serves as a guide for the representation learning(RepL)module.First,we determine an optimal autoencoder(AE)structure through a neural architecture search(NAS)to ensure efficient RepL;second,the RepL process is guided by a joint training strategy that combines AE reconstruction loss with the clustering objective.The results demonstrate that CTRL achieves excellent performance on both the generated and experimental data.Further inspection of the RepL step reveals that joint training robustly learns more compact features than the unconstrained AE or traditional dimensionality reduction methods,significantly reducing misclustering possibilities.Our method provides a general end-to-end automatic clustering solution for analyzing single-molecule break junction data.展开更多
Background Three-dimensional(3D)shape representation using mesh data is essential in various applications,such as virtual reality and simulation technologies.Current methods for extracting features from mesh edges or ...Background Three-dimensional(3D)shape representation using mesh data is essential in various applications,such as virtual reality and simulation technologies.Current methods for extracting features from mesh edges or faces struggle with complex 3D models because edge-based approaches miss global contexts and face-based methods overlook variations in adjacent areas,which affects the overall precision.To address these issues,we propose the Feature Discrimination and Context Propagation Network(FDCPNet),which is a novel approach that synergistically integrates local and global features in mesh datasets.Methods FDCPNet is composed of two modules:(1)the Feature Discrimination Module,which employs an attention mechanism to enhance the identification of key local features,and(2)the Context Propagation Module,which enriches key local features by integrating global contextual information,thereby facilitating a more detailed and comprehensive representation of crucial areas within the mesh model.Results Experiments on popular datasets validated the effectiveness of FDCPNet,showing an improvement in the classification accuracy over the baseline MeshNet.Furthermore,even with reduced mesh face numbers and limited training data,FDCPNet achieved promising results,demonstrating its robustness in scenarios of variable complexity.展开更多
This paper considers the notions of common sense and interobjectivity to articulate an understanding of how different cultural realities give rise to different construals of scientific phenomena across distinct cultur...This paper considers the notions of common sense and interobjectivity to articulate an understanding of how different cultural realities give rise to different construals of scientific phenomena across distinct cultures. Our main focus in this paper is on the social sciences. We propose a quadrant of different cultural–scientific stances from which the study of social phenomena is possible, based on the emic–etic dimension pertaining to the study of culture from contrasting perspectives. Although the emic–etic distinction is normal y applied in fields within the science of culture, it is proposed here that the distinction is in some ways germane to scientific practice in general, making it amenable for use in a culture of science(CoS) programme. The four perspectives that emerge from the quadrant are illustrated using exemplars. Different aspects of CoS—that is, scientific practice, scientific conventions and representations of science—are then discussed in further detail, including in two tables illustrating points of convergence and divergence between the East and West when it comes to different aspects of CoS.展开更多
Sparse representation has been highly successful in various tasks related to image processing and computer vision.For ancient mural image inpainting,traditional group sparse representation models usually lead to struc...Sparse representation has been highly successful in various tasks related to image processing and computer vision.For ancient mural image inpainting,traditional group sparse representation models usually lead to structure blur and line discontinuity due to the construction of similarity group solely based on the Euclidean distance and the randomness of dictionary initialization.To address the aforementioned issues,an improved curvature Gabor transform and group sparse representation(CGabor-GSR)model for ancient Dunhuang mural inpainting is proposed.To begin with,mutual information is introduced to weight the Euclidean distance,and then the weighted Euclidean distance acts as a new standard of similarity group.Subsequently,to mitigate the randomness of dictionary initialization,a curvature Gabor wavelet transform is proposed to extract the features and initialize the feature dictionary with dimension reduction based on principal component analysis(PCA).Ultimately,singular value decomposition(SVD)and split Bregman iteration(SBI)can be used to resolve the CGabor-GSR model to reconstruct the mural images.Experimental results on Dunhuang mural inpainting demonstrate tha the proposed CGabor-GSR achieves a better performance than compared algorithms in both objective and visual evaluation.展开更多
Phase classification has a clear guiding significance for the design of high entropy alloys.For mutually exclusive and non-mutually exclusive classifications,the composition descriptors,commonly used physical paramete...Phase classification has a clear guiding significance for the design of high entropy alloys.For mutually exclusive and non-mutually exclusive classifications,the composition descriptors,commonly used physical parameter descriptors,elemental-property descriptors,and descriptors extracted from the periodic table representation(PTR)by the convolutional neural network were collected.Appropriate selection among features with rich information is helpful for phase classification.Based on random forest,the accuracy of the four-label classification and balanced accuracy of the five-label classification were improved to be 0.907 and 0.876,respectively.The roles of the four important features were summarized by interpretability analysis,and a new important feature was found.The model extrapolation ability and the influence of Mo were demonstrated by phase prediction in(CoFeNiMn)_(1-x)Mo_(x).The phase information is helpful for the hardness prediction,the classification results were coupled with the PTR of hardness data,and the prediction error(the root mean square error)was reduced to 56.69.展开更多
Infrared and visible image fusion technology integrates the thermal radiation information of infrared images with the texture details of visible images to generate more informative fused images.However,existing method...Infrared and visible image fusion technology integrates the thermal radiation information of infrared images with the texture details of visible images to generate more informative fused images.However,existing methods often fail to distinguish salient objects from background regions,leading to detail suppression in salient regions due to global fusion strategies.This study presents a mask-guided latent low-rank representation fusion method to address this issue.First,the GrabCut algorithm is employed to extract a saliency mask,distinguishing salient regions from background regions.Then,latent low-rank representation(LatLRR)is applied to extract deep image features,enhancing key information extraction.In the fusion stage,a weighted fusion strategy strengthens infrared thermal information and visible texture details in salient regions,while an average fusion strategy improves background smoothness and stability.Experimental results on the TNO dataset demonstrate that the proposed method achieves superior performance in SPI,MI,Qabf,PSNR,and EN metrics,effectively preserving salient target details while maintaining balanced background information.Compared to state-of-the-art fusion methods,our approach achieves more stable and visually consistent fusion results.The fusion code is available on GitHub at:https://github.com/joyzhen1/Image(accessed on 15 January 2025).展开更多
Artificial intelligence(AI)researchers and cheminformatics specialists strive to identify effective drug precursors while optimizing costs and accelerating development processes.Digital molecular representation plays ...Artificial intelligence(AI)researchers and cheminformatics specialists strive to identify effective drug precursors while optimizing costs and accelerating development processes.Digital molecular representation plays a crucial role in achieving this objective by making molecules machine-readable,thereby enhancing the accuracy of molecular prediction tasks and facilitating evidence-based decision making.This study presents a comprehensive review of small molecular representations and AI-driven drug discovery downstream tasks utilizing these representations.The research methodology begins with the compilation of small molecule databases,followed by an analysis of fundamental molecular representations and the models that learn these representations from initial forms,capturing patterns and salient features across extensive chemical spaces.The study then examines various drug discovery downstream tasks,including drug-target interaction(DTI)prediction,drug-target affinity(DTA)prediction,drug property(DP)prediction,and drug generation,all based on learned representations.The analysis concludes by highlighting challenges and opportunities associated with machine learning(ML)methods for molecular representation and improving downstream task performance.Additionally,the representation of small molecules and AI-based downstream tasks demonstrates significant potential in identifying traditional Chinese medicine(TCM)medicinal substances and facilitating TCM target discovery.展开更多
基金supported by the National Natural Science Foundation of China under Grant 42474139the Key Research and Development Program of Shaanxi under Grant 2024GX-YBXM-067.
文摘Seismic time-frequency(TF)transforms are essential tools in reservoir interpretation and signal processing,particularly for characterizing frequency variations in non-stationary seismic data.Recently,sparse TF trans-forms,which leverage sparse coding(SC),have gained significant attention in the geosciences due to their ability to achieve high TF resolution.However,the iterative approaches typically employed in sparse TF transforms are computationally intensive,making them impractical for real seismic data analysis.To address this issue,we propose an interpretable convolutional sparse coding(CSC)network to achieve high TF resolution.The proposed model is generated based on the traditional short-time Fourier transform(STFT)transform and a modified UNet,named ULISTANet.In this design,we replace the conventional convolutional layers of the UNet with learnable iterative shrinkage thresholding algorithm(LISTA)blocks,a specialized form of CSC.The LISTA block,which evolves from the traditional iterative shrinkage thresholding algorithm(ISTA),is optimized for extracting sparse features more effectively.Furthermore,we create a synthetic dataset featuring complex frequency-modulated signals to train ULISTANet.Finally,the proposed method’s performance is subsequently validated using both synthetic and field data,demonstrating its potential for enhanced seismic data analysis.
基金supported by the National Natural Science Foundation of China(61261046,61362038)the Natural Science Foundation of Jiangxi Province(20142BAB207006,20151BAB207013)+2 种基金the Science and Technology Project of Provincial Education Department of Jiangxi Province(GJJ14738,GJJ14739)the Research Foundation of Health Department of Jiangxi Province(20175561)the Science and Technology Project of Jiujiang University(2016KJ001,2016KJ002)
文摘The machinery fault signal is a typical non-Gaussian and non-stationary process. The fault signal can be described by SaS distribution model because of the presence of impulses.Time-frequency distribution is a useful tool to extract helpful information of the machinery fault signal. Various fractional lower order(FLO) time-frequency distribution methods have been proposed based on fractional lower order statistics, which include fractional lower order short time Fourier transform(FLO-STFT), fractional lower order Wigner-Ville distributions(FLO-WVDs), fractional lower order Cohen class time-frequency distributions(FLO-CDs), fractional lower order adaptive kernel time-frequency distributions(FLO-AKDs) and adaptive fractional lower order time-frequency auto-regressive moving average(FLO-TFARMA) model time-frequency representation method.The methods and the exiting methods based on second order statistics in SaS distribution environments are compared, simulation results show that the new methods have better performances than the existing methods. The advantages and disadvantages of the improved time-frequency methods have been summarized.Last, the new methods are applied to analyze the outer race fault signals, the results illustrate their good performances.
基金This project was supported by the National Natural Science Foundation of China (60472102)Shanghai Leading Academic Discipline Project (T0103).
文摘A localized parametric time-sheared Gabor atom is derived by convolving a linear frequency modulated factor, modulating in frequency and translating in time to a dilated Gaussian function, which is the generalization of Gabor atom and is more delicate for matching most of the signals encountered in practice, especially for those having frequency dispersion characteristics. The time-frequency distribution of this atom concentrates in its time center and frequency center along energy curve, with the curve being oblique to a certain extent along the time axis. A novel parametric adaptive time-frequency distribution based on a set of the derived atoms is then proposed using a adaptive signal subspace decomposition method in frequency domain, which is non-negative time-frequency energy distribution and free of cross-term interference for multicomponent signals. The results of numerical simulation manifest the effectiveness of the approach in time-frequency representation and signal de-noising processing.
基金supported by the National Natural Science Foundation of China(611011726137118461301262)
文摘This paper proposes a new method for estimating the parameter of maneuvering targets based on sparse time-frequency transform in over-the-horizon radar（OTHR）. In this method, the sparse time-frequency distribution of the radar echo is obtained by solving a sparse optimization problem based on the short-time Fourier transform. Then Hough transform is employed to estimate the parameter of the targets. The proposed algorithm has the following advantages： Compared with the Wigner-Hough transform method, the computational complexity of the sparse optimization is low due to the application of fast Fourier transform（FFT）. And the computational cost of Hough transform is also greatly reduced because of the sparsity of the time-frequency distribution. Compared with the high order ambiguity function（HAF） method, the proposed method improves in terms of precision and robustness to noise. Simulation results show that compared with the HAF method, the required SNR and relative mean square error are 8 dB lower and 50 dB lower respectively in the proposed method. While processing the field experiment data, the execution time of Hough transform in the proposed method is only 4% of the Wigner-Hough transform method.
基金Supported by the National Natural Science Fundation of China(Grant No.69775009)
文摘A new time-frequency representation called Dopplerlet transform, which uses the dilated, translated and modulated windowed Doppler signals as its basis functions, is proposed, and the Fourier transform, short-time Fourier transform (including Gabor transform), wavelet transform, and chirplet transform are formulated in one framework of Dopplerlet transform accordingly.It is proved that the matching pursuits based on Dopplerlet basis functions are convergent, and that the energy of residual signals yielded in the decomposition process decays exponentially. Simulation results show that the matching pursuits with Dopplerlet basis functions can characterize compactly a nonstationary signal.
文摘Let F_(1)be the virtual field consisting of one element and(Q,I)a string pair.In this paper,we study the representations of string pairs over the virtual field F_(1).It is proved that an indecomposable F_(1)-representation is either a string representation or a band representation by using the coefficient quivers.It is worth noting that for a given band and a positive integer,there exists a unique band representation up to isomorphism.
基金supported by the National Social Science Fund of China’s project‘Philosophical Research on the Challenge of Artificial Cognition to Natural Cognition’(grant number 21&ZD061)
文摘As a new research direction in contemporary cognitive science,predictive processing surpasses traditional computational representation and embodied cognition and has emerged as a new paradigm in cognitive science research.The predictive processing theory advocates that the brain is a hierarchical predictive model based on Bayesian inference,and its purpose is to minimize the difference between the predicted world and the actual world,so as to minimize the prediction error.Predictive processing is therefore essentially a context-dependent model representation,an adaptive representational system designed to achieve its cognitive goals through the minimization of prediction error.
基金supported by the National Natural Science Foundation of China(Nos.12072027,62103052,61603346 and 62103379)the Henan Key Laboratory of General Aviation Technology,China(No.ZHKF-230201)+3 种基金the Funding for the Open Research Project of the Rotor Aerodynamics Key Laboratory,China(No.RAL20200101)the Key Research and Development Program of Henan Province,China(Nos.241111222000 and 241111222900)the Key Science and Technology Program of Henan Province,China(No.232102220067)the Scholarship Funding from the China Scholarship Council(No.202206030079).
文摘In global navigation satellite system denial environment,cross-view geo-localization based on image retrieval presents an exceedingly critical visual localization solution for Unmanned Aerial Vehicle(UAV)systems.The essence of cross-view geo-localization resides in matching images containing the same geographical targets from disparate platforms,such as UAV-view and satellite-view images.However,images of the same geographical targets may suffer from occlusions and geometric distortions due to variations in the capturing platform,view,and timing.The existing methods predominantly extract features by segmenting feature maps,which overlook the holistic semantic distribution and structural information of objects,resulting in loss of image information.To address these challenges,dilated neighborhood attention Transformer is employed as the feature extraction backbone,and Multi-feature representations based on Multi-scale Hierarchical Contextual Aggregation(MMHCA)is proposed.In the proposed MMHCA method,the multiscale hierarchical contextual aggregation method is utilized to extract contextual information from local to global across various granularity levels,establishing feature associations of contextual information with global and local information in the image.Subsequently,the multi-feature representations method is utilized to obtain rich discriminative feature information,bolstering the robustness of model in scenarios characterized by positional shifts,varying distances,and scale ambiguities.Comprehensive experiments conducted on the extensively utilized University-1652 and SUES-200 benchmarks indicate that the MMHCA method surpasses the existing techniques.showing outstanding results in UAV localization and navigation.
基金supported by the National Natural Science Foun-dation of China(Grant No.72349002).
文摘Establishing and maintaining protected areas is a pivotal strategy for attaining the post-2020 biodiversity target. The conservation objectives of protected areas have shifted from a narrow emphasis on biodiversity to encompass broader considerations such as ecosystem stability, community resilience to climate change, and enhancement of human well-being. Given these multifaceted objectives, it is imperative to judiciously allocate resources to effectively conserve biodiversity by identifying strategically significant areas for conservation, particularly for mountainous areas. In this study, we evaluated the representativeness of the protected area network in the Qin ling Mountains concerning species diversity, ecosystem services, climate stability and ecological stability. The results indicate that some of the ecological indicators are spatially correlated with topographic gradient effects. The conservation priority areas predominantly lie in the northern foothills, the southeastern, and southwestern parts of the Qinling Mountain with areas concentrated at altitudes between 1,500-2,000 m and slopes between 40°-50° as hotspots. The conservation priority areas identified through the framework of inclusive conservation optimization account for 22.9 % of the Qinling Mountain. Existing protected areas comprise only 6.1 % of the Qinling Mountain and 13.18 % of the conservation priority areas. This will play an important role in achiev ing sustainable development in the region and in meeting the post-2020 biodiversity target. The framework can advance the different objectives of achieving a quadruple win and can also be extended to other regions.
基金King Abdullah University of Science and Technol-ogy(KAUST)for supporting this research and the Seismic Wave Anal-ysis group for the supportive and encouraging environment.
文摘Physics-informed neural networks(PINNs)are promising to replace conventional mesh-based partial tial differen-equation(PDE)solvers by offering more accurate and flexible PDE solutions.However,PINNs are hampered by the relatively slow convergence and the need to perform additional,potentially expensive training for new PDE parameters.To solve this limitation,we introduce LatentPINN,a framework that utilizes latent representations of the PDE parameters as additional(to the coordinates)inputs into PINNs and allows for training over the distribution of these parameters.Motivated by the recent progress on generative models,we promote using latent diffusion models to learn compressed latent representations of the distribution of PDE parameters as they act as input parameters for NN functional solutions.We use a two-stage training scheme in which,in the first stage,we learn the latent representations for the distribution of PDE parameters.In the second stage,we train a physics-informed neural network over inputs given by randomly drawn samples from the coordinate space within the solution domain and samples from the learned latent representation of the PDE parameters.Considering their importance in capturing evolving interfaces and fronts in various fields,we test the approach on a class of level set equations given,for example,by the nonlinear Eikonal equation.We share results corresponding to three Eikonal parameters(velocity models)sets.The proposed method performs well on new phase velocity models without the need for any additional training.
基金the financial support from Natural Science Foundation of Gansu Province(Nos.22JR5RA217,22JR5RA216)Lanzhou Science and Technology Program(No.2022-2-111)+1 种基金Lanzhou University of Arts and Sciences School Innovation Fund Project(No.XJ2022000103)Lanzhou College of Arts and Sciences 2023 Talent Cultivation Quality Improvement Project(No.2023-ZL-jxzz-03)。
文摘Considering that the algorithm accuracy of the traditional sparse representation models is not high under the influence of multiple complex environmental factors,this study focuses on the improvement of feature extraction and model construction.Firstly,the convolutional neural network(CNN)features of the face are extracted by the trained deep learning network.Next,the steady-state and dynamic classifiers for face recognition are constructed based on the CNN features and Haar features respectively,with two-stage sparse representation introduced in the process of constructing the steady-state classifier and the feature templates with high reliability are dynamically selected as alternative templates from the sparse representation template dictionary constructed using the CNN features.Finally,the results of face recognition are given based on the classification results of the steady-state classifier and the dynamic classifier together.Based on this,the feature weights of the steady-state classifier template are adjusted in real time and the dictionary set is dynamically updated to reduce the probability of irrelevant features entering the dictionary set.The average recognition accuracy of this method is 94.45%on the CMU PIE face database and 96.58%on the AR face database,which is significantly improved compared with that of the traditional face recognition methods.
基金supported by Guangxi Science and Technology Program(No.GuiKeAD23026291)Guangxi Science and Technology Major Project(No.AA22068057).
文摘Clustering is a pivotal data analysis method for deciphering the charge transport properties of single molecules in break junction experiments.However,given the high dimensionality and variability of the data,feature extraction remains a bottleneck in the development of efficient clustering methods.In this regard,extensive research over the past two decades has focused on feature engineering and dimensionality reduction in break junction conductance.However,extracting highly relevant features without expert knowledge remains an unresolved challenge.To address this issue,we propose a deep clustering method driven by task-oriented representation learning(CTRL)in which the clustering module serves as a guide for the representation learning(RepL)module.First,we determine an optimal autoencoder(AE)structure through a neural architecture search(NAS)to ensure efficient RepL;second,the RepL process is guided by a joint training strategy that combines AE reconstruction loss with the clustering objective.The results demonstrate that CTRL achieves excellent performance on both the generated and experimental data.Further inspection of the RepL step reveals that joint training robustly learns more compact features than the unconstrained AE or traditional dimensionality reduction methods,significantly reducing misclustering possibilities.Our method provides a general end-to-end automatic clustering solution for analyzing single-molecule break junction data.
基金Supported by the National Key R&D Program of China(2022YFC3803600).
文摘Background Three-dimensional(3D)shape representation using mesh data is essential in various applications,such as virtual reality and simulation technologies.Current methods for extracting features from mesh edges or faces struggle with complex 3D models because edge-based approaches miss global contexts and face-based methods overlook variations in adjacent areas,which affects the overall precision.To address these issues,we propose the Feature Discrimination and Context Propagation Network(FDCPNet),which is a novel approach that synergistically integrates local and global features in mesh datasets.Methods FDCPNet is composed of two modules:(1)the Feature Discrimination Module,which employs an attention mechanism to enhance the identification of key local features,and(2)the Context Propagation Module,which enriches key local features by integrating global contextual information,thereby facilitating a more detailed and comprehensive representation of crucial areas within the mesh model.Results Experiments on popular datasets validated the effectiveness of FDCPNet,showing an improvement in the classification accuracy over the baseline MeshNet.Furthermore,even with reduced mesh face numbers and limited training data,FDCPNet achieved promising results,demonstrating its robustness in scenarios of variable complexity.
文摘This paper considers the notions of common sense and interobjectivity to articulate an understanding of how different cultural realities give rise to different construals of scientific phenomena across distinct cultures. Our main focus in this paper is on the social sciences. We propose a quadrant of different cultural–scientific stances from which the study of social phenomena is possible, based on the emic–etic dimension pertaining to the study of culture from contrasting perspectives. Although the emic–etic distinction is normal y applied in fields within the science of culture, it is proposed here that the distinction is in some ways germane to scientific practice in general, making it amenable for use in a culture of science(CoS) programme. The four perspectives that emerge from the quadrant are illustrated using exemplars. Different aspects of CoS—that is, scientific practice, scientific conventions and representations of science—are then discussed in further detail, including in two tables illustrating points of convergence and divergence between the East and West when it comes to different aspects of CoS.
基金supported by National Natural Science Foundation of China(No.61963023)Humanities and Social Sciences Youth Foundation of Ministry of Education(No.19YJC760012)Lanzhou Jiaotong University Basic Top-Notch Personnel Project(No.2022JC36).
文摘Sparse representation has been highly successful in various tasks related to image processing and computer vision.For ancient mural image inpainting,traditional group sparse representation models usually lead to structure blur and line discontinuity due to the construction of similarity group solely based on the Euclidean distance and the randomness of dictionary initialization.To address the aforementioned issues,an improved curvature Gabor transform and group sparse representation(CGabor-GSR)model for ancient Dunhuang mural inpainting is proposed.To begin with,mutual information is introduced to weight the Euclidean distance,and then the weighted Euclidean distance acts as a new standard of similarity group.Subsequently,to mitigate the randomness of dictionary initialization,a curvature Gabor wavelet transform is proposed to extract the features and initialize the feature dictionary with dimension reduction based on principal component analysis(PCA).Ultimately,singular value decomposition(SVD)and split Bregman iteration(SBI)can be used to resolve the CGabor-GSR model to reconstruct the mural images.Experimental results on Dunhuang mural inpainting demonstrate tha the proposed CGabor-GSR achieves a better performance than compared algorithms in both objective and visual evaluation.
基金supported by the National Natural Science Foundation of China(Nos.51671075,51971086)the Natural Science Foundation of Heilongjiang Province,China(No.LH2022E081)。
文摘Phase classification has a clear guiding significance for the design of high entropy alloys.For mutually exclusive and non-mutually exclusive classifications,the composition descriptors,commonly used physical parameter descriptors,elemental-property descriptors,and descriptors extracted from the periodic table representation(PTR)by the convolutional neural network were collected.Appropriate selection among features with rich information is helpful for phase classification.Based on random forest,the accuracy of the four-label classification and balanced accuracy of the five-label classification were improved to be 0.907 and 0.876,respectively.The roles of the four important features were summarized by interpretability analysis,and a new important feature was found.The model extrapolation ability and the influence of Mo were demonstrated by phase prediction in(CoFeNiMn)_(1-x)Mo_(x).The phase information is helpful for the hardness prediction,the classification results were coupled with the PTR of hardness data,and the prediction error(the root mean square error)was reduced to 56.69.
基金supported by Universiti Teknologi MARA through UiTM MyRA Research Grant,600-RMC 5/3/GPM(053/2022).
文摘Infrared and visible image fusion technology integrates the thermal radiation information of infrared images with the texture details of visible images to generate more informative fused images.However,existing methods often fail to distinguish salient objects from background regions,leading to detail suppression in salient regions due to global fusion strategies.This study presents a mask-guided latent low-rank representation fusion method to address this issue.First,the GrabCut algorithm is employed to extract a saliency mask,distinguishing salient regions from background regions.Then,latent low-rank representation(LatLRR)is applied to extract deep image features,enhancing key information extraction.In the fusion stage,a weighted fusion strategy strengthens infrared thermal information and visible texture details in salient regions,while an average fusion strategy improves background smoothness and stability.Experimental results on the TNO dataset demonstrate that the proposed method achieves superior performance in SPI,MI,Qabf,PSNR,and EN metrics,effectively preserving salient target details while maintaining balanced background information.Compared to state-of-the-art fusion methods,our approach achieves more stable and visually consistent fusion results.The fusion code is available on GitHub at:https://github.com/joyzhen1/Image(accessed on 15 January 2025).
基金supported by the Shenzhen Key Laboratory of Intelligent Bioinformatics(No.ZDSYS20220422103800001)the Shenzhen Science and Technology Program(No.JCYJ20230807140709020)+2 种基金National Natural Science Foundation of China(Nos.62402489,U22A2041,and 62373172)the China Postdoctoral Science Foundation(No.2023M743688)Guangdong Basic and Applied Basic Research Foundation(Nos.2024A1515011960 and 2023A1515110570)。
文摘Artificial intelligence(AI)researchers and cheminformatics specialists strive to identify effective drug precursors while optimizing costs and accelerating development processes.Digital molecular representation plays a crucial role in achieving this objective by making molecules machine-readable,thereby enhancing the accuracy of molecular prediction tasks and facilitating evidence-based decision making.This study presents a comprehensive review of small molecular representations and AI-driven drug discovery downstream tasks utilizing these representations.The research methodology begins with the compilation of small molecule databases,followed by an analysis of fundamental molecular representations and the models that learn these representations from initial forms,capturing patterns and salient features across extensive chemical spaces.The study then examines various drug discovery downstream tasks,including drug-target interaction(DTI)prediction,drug-target affinity(DTA)prediction,drug property(DP)prediction,and drug generation,all based on learned representations.The analysis concludes by highlighting challenges and opportunities associated with machine learning(ML)methods for molecular representation and improving downstream task performance.Additionally,the representation of small molecules and AI-based downstream tasks demonstrates significant potential in identifying traditional Chinese medicine(TCM)medicinal substances and facilitating TCM target discovery.
