The advent of self-attention mechanisms within Transformer models has significantly propelled the advancement of deep learning algorithms,yielding outstanding achievements across diverse domains.Nonetheless,self-atten...The advent of self-attention mechanisms within Transformer models has significantly propelled the advancement of deep learning algorithms,yielding outstanding achievements across diverse domains.Nonetheless,self-attention mechanisms falter when applied to datasets with intricate semantic content and extensive dependency structures.In response,this paper introduces a Diffusion Sampling and Label-Driven Co-attention Neural Network(DSLD),which adopts a diffusion sampling method to capture more comprehensive semantic information of the data.Additionally,themodel leverages the joint correlation information of labels and data to introduce the computation of text representation,correcting semantic representationbiases in thedata,andincreasing the accuracyof semantic representation.Ultimately,the model computes the corresponding classification results by synthesizing these rich data semantic representations.Experiments on seven benchmark datasets show that our proposed model achieves competitive results compared to state-of-the-art methods.展开更多
Physics-informed neural networks(PINNs)have become an attractive machine learning framework for obtaining solutions to partial differential equations(PDEs).PINNs embed initial,boundary,and PDE constraints into the los...Physics-informed neural networks(PINNs)have become an attractive machine learning framework for obtaining solutions to partial differential equations(PDEs).PINNs embed initial,boundary,and PDE constraints into the loss function.The performance of PINNs is generally affected by both training and sampling.Specifically,training methods focus on how to overcome the training difficulties caused by the special PDE residual loss of PINNs,and sampling methods are concerned with the location and distribution of the sampling points upon which evaluations of PDE residual loss are accomplished.However,a common problem among these original PINNs is that they omit special temporal information utilization during the training or sampling stages when dealing with an important PDE category,namely,time-dependent PDEs,where temporal information plays a key role in the algorithms used.There is one method,called Causal PINN,that considers temporal causality at the training level but not special temporal utilization at the sampling level.Incorporating temporal knowledge into sampling remains to be studied.To fill this gap,we propose a novel temporal causality-based adaptive sampling method that dynamically determines the sampling ratio according to both PDE residual and temporal causality.By designing a sampling ratio determined by both residual loss and temporal causality to control the number and location of sampled points in each temporal sub-domain,we provide a practical solution by incorporating temporal information into sampling.Numerical experiments of several nonlinear time-dependent PDEs,including the Cahn–Hilliard,Korteweg–de Vries,Allen–Cahn and wave equations,show that our proposed sampling method can improve the performance.We demonstrate that using such a relatively simple sampling method can improve prediction performance by up to two orders of magnitude compared with the results from other methods,especially when points are limited.展开更多
Video-text retrieval (VTR) is an essential task in multimodal learning, aiming to bridge the semantic gap between visual and textual data. Effective video frame sampling plays a crucial role in improving retrieval per...Video-text retrieval (VTR) is an essential task in multimodal learning, aiming to bridge the semantic gap between visual and textual data. Effective video frame sampling plays a crucial role in improving retrieval performance, as it determines the quality of the visual content representation. Traditional sampling methods, such as uniform sampling and optical flow-based techniques, often fail to capture the full semantic range of videos, leading to redundancy and inefficiencies. In this work, we propose CLIP4Video-Sampling: Global Semantics-Guided Multi-Granularity Frame Sampling for Video-Text Retrieval, a global semantics-guided multi-granularity frame sampling strategy designed to optimize both computational efficiency and retrieval accuracy. By integrating multi-scale global and local temporal sampling and leveraging the CLIP (Contrastive Language-Image Pre-training) model’s powerful feature extraction capabilities, our method significantly outperforms existing approaches in both zero-shot and fine-tuned video-text retrieval tasks on popular datasets. CLIP4Video-Sampling reduces redundancy, ensures keyframe coverage, and serves as an adaptable pre-processing module for multimodal models.展开更多
The formation of the mammalian nervous system and its maturation into sensory,motor,cognitive,and behavioral circuits is a complex process that begins during early embryogenesis and lasts until young adulthood.Impaire...The formation of the mammalian nervous system and its maturation into sensory,motor,cognitive,and behavioral circuits is a complex process that begins during early embryogenesis and lasts until young adulthood.Impaired neurodevelopment can result in various neurological and psychiatric conditions,jointly named neurodevelopmental disorders(NDDs).展开更多
The members of the fourth subgroup of R2R3-MYB(Sg4 members)are well-known inhibitors of phenylpropanoid and lignin synthesis pathways.The C2 domain is closely related to the transcriptional inhibitory activity of Sg4 ...The members of the fourth subgroup of R2R3-MYB(Sg4 members)are well-known inhibitors of phenylpropanoid and lignin synthesis pathways.The C2 domain is closely related to the transcriptional inhibitory activity of Sg4 members.Phosphorylation modification enhances the transcriptional inhibitory activity of Sg4 members.Here,we identified a phosphorylation site on the C2 domain of Cs MYB4a from tea plants(Camellia sinensis).A mitogen-activated protein kinase(MAPK),named Cs MPK3-2,phosphorylated this site on the C2 domain of Cs MYB4a.Further experiments revealed that phosphorylation of Cs MYB4a weakened its ability to inhibit the gene expression of PAL,C4H,and 4CL in the phenylpropanoid pathway and activated the expression of transcription factor YABBY5,maintaining the adaxial-abaxial polarity of the leaf.Knocking out Nt YAB5 in Cs MYB4a transgenic tobacco partially repaired the leaf wrinkling phenotype caused by Cs MYB4a.The C1 domain exhibited an activation function when the C2 domain of Cs MYB4a was phosphorylated by Cs MPK3-2,causing this reversal phenomenon.These results enrich our understanding of the regulatory diversity of Sg4 members.展开更多
It is essential to investigate the light field camera parameters for the accurate flame temperature measurement because the sampling characteristics of the flame radiation can be varied with them. In this study, novel...It is essential to investigate the light field camera parameters for the accurate flame temperature measurement because the sampling characteristics of the flame radiation can be varied with them. In this study, novel indices of the light field camera were proposed to investigate the directional and spatial sampling characteristics of the flame radiation. Effects of light field camera parameters such as focal length and magnification of the main lens, focal length and magnification of the microlens were investigated. It was observed that the sampling characteristics of the flame are varied with the different parameters of the light field camera. The optimized parameters of the light field camera were then proposed for the flame radiation sampling. The larger sampling angle(23 times larger) is achieved by the optimized parameters compared to the commercial light field camera parameters. A non-negative least square(NNLS) algorithm was used to reconstruct the flame temperature. The reconstruction accuracy was also evaluated by the optimized parameters. The results suggested that the optimized parameters can provide higher reconstruction accuracy for axisymmetric and non-symmetric flame conditions in comparison to the commercial light field camera.展开更多
China＇s continental deposition basins are characterized by complex geological structures and various reservoir lithologies. Therefore, high precision exploration methods are needed. High density spatial sampling is a...China＇s continental deposition basins are characterized by complex geological structures and various reservoir lithologies. Therefore, high precision exploration methods are needed. High density spatial sampling is a new technology to increase the accuracy of seismic exploration. We briefly discuss point source and receiver technology, analyze the high density spatial sampling in situ method, introduce the symmetric sampling principles presented by Gijs J. O. Vermeer, and discuss high density spatial sampling technology from the point of view of wave field continuity. We emphasize the analysis of the high density spatial sampling characteristics, including the high density first break advantages for investigation of near surface structure, improving static correction precision, the use of dense receiver spacing at short offsets to increase the effective coverage at shallow depth, and the accuracy of reflection imaging. Coherent noise is not aliased and the noise analysis precision and suppression increases as a result. High density spatial sampling enhances wave field continuity and the accuracy of various mathematical transforms, which benefits wave field separation. Finally, we point out that the difficult part of high density spatial sampling technology is the data processing. More research needs to be done on the methods of analyzing and processing huge amounts of seismic data.展开更多
基金the Communication University of China(CUC230A013)the Fundamental Research Funds for the Central Universities.
文摘The advent of self-attention mechanisms within Transformer models has significantly propelled the advancement of deep learning algorithms,yielding outstanding achievements across diverse domains.Nonetheless,self-attention mechanisms falter when applied to datasets with intricate semantic content and extensive dependency structures.In response,this paper introduces a Diffusion Sampling and Label-Driven Co-attention Neural Network(DSLD),which adopts a diffusion sampling method to capture more comprehensive semantic information of the data.Additionally,themodel leverages the joint correlation information of labels and data to introduce the computation of text representation,correcting semantic representationbiases in thedata,andincreasing the accuracyof semantic representation.Ultimately,the model computes the corresponding classification results by synthesizing these rich data semantic representations.Experiments on seven benchmark datasets show that our proposed model achieves competitive results compared to state-of-the-art methods.
基金Project supported by the Key National Natural Science Foundation of China(Grant No.62136005)the National Natural Science Foundation of China(Grant Nos.61922087,61906201,and 62006238)。
文摘Physics-informed neural networks(PINNs)have become an attractive machine learning framework for obtaining solutions to partial differential equations(PDEs).PINNs embed initial,boundary,and PDE constraints into the loss function.The performance of PINNs is generally affected by both training and sampling.Specifically,training methods focus on how to overcome the training difficulties caused by the special PDE residual loss of PINNs,and sampling methods are concerned with the location and distribution of the sampling points upon which evaluations of PDE residual loss are accomplished.However,a common problem among these original PINNs is that they omit special temporal information utilization during the training or sampling stages when dealing with an important PDE category,namely,time-dependent PDEs,where temporal information plays a key role in the algorithms used.There is one method,called Causal PINN,that considers temporal causality at the training level but not special temporal utilization at the sampling level.Incorporating temporal knowledge into sampling remains to be studied.To fill this gap,we propose a novel temporal causality-based adaptive sampling method that dynamically determines the sampling ratio according to both PDE residual and temporal causality.By designing a sampling ratio determined by both residual loss and temporal causality to control the number and location of sampled points in each temporal sub-domain,we provide a practical solution by incorporating temporal information into sampling.Numerical experiments of several nonlinear time-dependent PDEs,including the Cahn–Hilliard,Korteweg–de Vries,Allen–Cahn and wave equations,show that our proposed sampling method can improve the performance.We demonstrate that using such a relatively simple sampling method can improve prediction performance by up to two orders of magnitude compared with the results from other methods,especially when points are limited.
文摘Video-text retrieval (VTR) is an essential task in multimodal learning, aiming to bridge the semantic gap between visual and textual data. Effective video frame sampling plays a crucial role in improving retrieval performance, as it determines the quality of the visual content representation. Traditional sampling methods, such as uniform sampling and optical flow-based techniques, often fail to capture the full semantic range of videos, leading to redundancy and inefficiencies. In this work, we propose CLIP4Video-Sampling: Global Semantics-Guided Multi-Granularity Frame Sampling for Video-Text Retrieval, a global semantics-guided multi-granularity frame sampling strategy designed to optimize both computational efficiency and retrieval accuracy. By integrating multi-scale global and local temporal sampling and leveraging the CLIP (Contrastive Language-Image Pre-training) model’s powerful feature extraction capabilities, our method significantly outperforms existing approaches in both zero-shot and fine-tuned video-text retrieval tasks on popular datasets. CLIP4Video-Sampling reduces redundancy, ensures keyframe coverage, and serves as an adaptable pre-processing module for multimodal models.
基金supported by Danish National Research Foundation(#DNRF133)(to AN)。
文摘The formation of the mammalian nervous system and its maturation into sensory,motor,cognitive,and behavioral circuits is a complex process that begins during early embryogenesis and lasts until young adulthood.Impaired neurodevelopment can result in various neurological and psychiatric conditions,jointly named neurodevelopmental disorders(NDDs).
基金supported by the joint funds of National Natural Science Foundation of China(Grant Nos.U21A20232,32372756,32072621)。
文摘The members of the fourth subgroup of R2R3-MYB(Sg4 members)are well-known inhibitors of phenylpropanoid and lignin synthesis pathways.The C2 domain is closely related to the transcriptional inhibitory activity of Sg4 members.Phosphorylation modification enhances the transcriptional inhibitory activity of Sg4 members.Here,we identified a phosphorylation site on the C2 domain of Cs MYB4a from tea plants(Camellia sinensis).A mitogen-activated protein kinase(MAPK),named Cs MPK3-2,phosphorylated this site on the C2 domain of Cs MYB4a.Further experiments revealed that phosphorylation of Cs MYB4a weakened its ability to inhibit the gene expression of PAL,C4H,and 4CL in the phenylpropanoid pathway and activated the expression of transcription factor YABBY5,maintaining the adaxial-abaxial polarity of the leaf.Knocking out Nt YAB5 in Cs MYB4a transgenic tobacco partially repaired the leaf wrinkling phenotype caused by Cs MYB4a.The C1 domain exhibited an activation function when the C2 domain of Cs MYB4a was phosphorylated by Cs MPK3-2,causing this reversal phenomenon.These results enrich our understanding of the regulatory diversity of Sg4 members.
基金supported by the National Natural Science Foundation of China(Grant Nos.51676044 and 51327803)the Social Development Project of Jiangsu Province,China(Grant No.BE20187053)+1 种基金the Postgraduate Research and Practice Innovation Program of Jiangsu Province,China(Grant No.KYCX170081)China Scholarship Council
文摘It is essential to investigate the light field camera parameters for the accurate flame temperature measurement because the sampling characteristics of the flame radiation can be varied with them. In this study, novel indices of the light field camera were proposed to investigate the directional and spatial sampling characteristics of the flame radiation. Effects of light field camera parameters such as focal length and magnification of the main lens, focal length and magnification of the microlens were investigated. It was observed that the sampling characteristics of the flame are varied with the different parameters of the light field camera. The optimized parameters of the light field camera were then proposed for the flame radiation sampling. The larger sampling angle(23 times larger) is achieved by the optimized parameters compared to the commercial light field camera parameters. A non-negative least square(NNLS) algorithm was used to reconstruct the flame temperature. The reconstruction accuracy was also evaluated by the optimized parameters. The results suggested that the optimized parameters can provide higher reconstruction accuracy for axisymmetric and non-symmetric flame conditions in comparison to the commercial light field camera.
文摘China＇s continental deposition basins are characterized by complex geological structures and various reservoir lithologies. Therefore, high precision exploration methods are needed. High density spatial sampling is a new technology to increase the accuracy of seismic exploration. We briefly discuss point source and receiver technology, analyze the high density spatial sampling in situ method, introduce the symmetric sampling principles presented by Gijs J. O. Vermeer, and discuss high density spatial sampling technology from the point of view of wave field continuity. We emphasize the analysis of the high density spatial sampling characteristics, including the high density first break advantages for investigation of near surface structure, improving static correction precision, the use of dense receiver spacing at short offsets to increase the effective coverage at shallow depth, and the accuracy of reflection imaging. Coherent noise is not aliased and the noise analysis precision and suppression increases as a result. High density spatial sampling enhances wave field continuity and the accuracy of various mathematical transforms, which benefits wave field separation. Finally, we point out that the difficult part of high density spatial sampling technology is the data processing. More research needs to be done on the methods of analyzing and processing huge amounts of seismic data.
