Imputation of missing data has long been an important topic and an essential application for intelligent transportation systems(ITS)in the real world.As a state-of-the-art generative model,the diffusion model has prov...Imputation of missing data has long been an important topic and an essential application for intelligent transportation systems(ITS)in the real world.As a state-of-the-art generative model,the diffusion model has proven highly successful in image generation,speech generation,time series modelling etc.and now opens a new avenue for traffic data imputation.In this paper,we propose a conditional diffusion model,called the implicit-explicit diffusion model,for traffic data imputation.This model exploits both the implicit and explicit feature of the data simultaneously.More specifically,we design two types of feature extraction modules,one to capture the implicit dependencies hidden in the raw data at multiple time scales and the other to obtain the long-term temporal dependencies of the time series.This approach not only inherits the advantages of the diffusion model for estimating missing data,but also takes into account the multiscale correlation inherent in traffic data.To illustrate the performance of the model,extensive experiments are conducted on three real-world time series datasets using different missing rates.The experimental results demonstrate that the model improves imputation accuracy and generalization capability.展开更多
With the growing application of intelligent robots in service,manufacturing,and medical fields,efficient and natural interaction between humans and robots has become key to improving collaboration efficiency and user ...With the growing application of intelligent robots in service,manufacturing,and medical fields,efficient and natural interaction between humans and robots has become key to improving collaboration efficiency and user experience.Gesture recognition,as an intuitive and contactless interaction method,can overcome the limitations of traditional interfaces and enable real-time control and feedback of robot movements and behaviors.This study first reviews mainstream gesture recognition algorithms and their application on different sensing platforms(RGB cameras,depth cameras,and inertial measurement units).It then proposes a gesture recognition method based on multimodal feature fusion and a lightweight deep neural network that balances recognition accuracy with computational efficiency.At system level,a modular human-robot interaction architecture is constructed,comprising perception,decision,and execution layers,and gesture commands are transmitted and mapped to robot actions in real time via the ROS communication protocol.Through multiple comparative experiments on public gesture datasets and a self-collected dataset,the proposed method’s superiority is validated in terms of accuracy,response latency,and system robustness,while user-experience tests assess the interface’s usability.The results provide a reliable technical foundation for robot collaboration and service in complex scenarios,offering broad prospects for practical application and deployment.展开更多
Knowledge distillation(KD)is an emerging model compression technique for learning compact object detector models.Previous KD often focused solely on distilling from the logits layer or the feature intermediate layers,...Knowledge distillation(KD)is an emerging model compression technique for learning compact object detector models.Previous KD often focused solely on distilling from the logits layer or the feature intermediate layers,which may limit the comprehensive learning of the student network.Additionally,the imbalance between the foreground and background also affects the performance of the model.To address these issues,this paper employs feature-based distillation to enhance the detection performance of the bounding box localization part,and logit-based distillation to improve the detection performance of the category prediction part.Specifically,for the intermediate layer feature distillation,we introduce feature resampling to reduce the risk of the student model merely imitating the teacher model.At the same time,we incorporate a Spatial Attention Mechanism(SAM)to highlight the foreground features learned by the student model.In terms of output layer feature distillation,we divide the traditional distillation targets into target-class objects and non-target-class objects,aiming to improve overall distillation performance.Furthermore,we introduce a one-to-many matching distillation strategy based on Feature Alignment Module(FAM),which further enhances the studentmodel’s feature representation ability,making its feature distribution closer to that of the teacher model,and thus demonstrating superior localization and classification capabilities in object detection tasks.Experimental results demonstrate that our proposedmethodology outperforms conventional distillation techniques in terms of object detecting performance.展开更多
Content-based 3D model retrieval is of great help to facilitate the reuse of existing designs and to inspire designers during conceptual design. However, there is still a gap to apply it in industry due to the low tim...Content-based 3D model retrieval is of great help to facilitate the reuse of existing designs and to inspire designers during conceptual design. However, there is still a gap to apply it in industry due to the low time efficiency. This paper presents two new methods with high efficiency to build a Content-based 3D model retrieval system. First, an improvement is made on the ＂Shape Distribution （D2）＂ algorithm, and a new algorithm named ＂Quick D2＂ is proposed. Four sample 3D mechanical models are used in an experiment to compare the time cost of the two algorithms. The result indicates that the time cost of Quick D2 is much lower than that of D2, while the descriptors extracted by the two algorithms are almost the same. Second, an expandable 3D model repository index method with high performance, namely, RBK index, is presented. On the basis of RBK index, the search space is pruned effectively during the search process, leading to a speed up of the whole system. The factors that influence the values of the key parameters of RBK index are discussed and an experimental method to find the optimal values of the key parameters is given. Finally, ＂3D Searcher＂, a content-based 3D model retrieval system is developed. By using the methods proposed, the time cost for the system to respond one query online is reduced by 75% on average. The system has been implemented in a manufacturing enterprise, and practical query examples during a case of the automobile rear axle design are also shown. The research method presented shows a new research perspective and can effectively improve the content-based 3D model retrieval efficiency.展开更多
After more than 70 years of evolution,great achievements have been made in machine translation.Especially in recent years,translation quality has been greatly improved with the emergence of neural machine translation(...After more than 70 years of evolution,great achievements have been made in machine translation.Especially in recent years,translation quality has been greatly improved with the emergence of neural machine translation(NMT).In this article,we first review the history of machine translation from rule-based machine translation to example-based machine translation and statistical machine translation.We then introduce NMT in more detail,including the basic framework and the current dominant framework,Transformer,as well as multilingual translation models to deal with the data sparseness problem.In addition,we introduce cutting-edge simultaneous translation methods that achieve a balance between translation quality and latency.We then describe various products and applications of machine translation.At the end of this article,we briefly discuss challenges and future research directions in this field.展开更多
The paper presents a novel benefit based query processing strategy for efficient query routing. Based on DHT as the overlay network, it first applies Nash equilibrium to construct the optimal peer group based on the c...The paper presents a novel benefit based query processing strategy for efficient query routing. Based on DHT as the overlay network, it first applies Nash equilibrium to construct the optimal peer group based on the correlations of keywords and coverage and overlap of the peers to decrease the time cost, and then presents a two-layered architecture for query processing that utilizes Bloom filter as compact representation to reduce the bandwidth consumption. Extensive experiments conducted on a real world dataset have demonstrated that our approach obviously decreases the processing time, while improves the precision and recall as well.展开更多
预训练语言模型(pre-trained languages model,PTLM)在自然语言处理(natural language processing,NLP)领域取得了令人瞩目的成功,并由此引发了下游任务从监督学习到预训练-微调范式的转变。在此之后,一系列预训练模型的创新研究涌现出...预训练语言模型(pre-trained languages model,PTLM)在自然语言处理(natural language processing,NLP)领域取得了令人瞩目的成功,并由此引发了下游任务从监督学习到预训练-微调范式的转变。在此之后,一系列预训练模型的创新研究涌现出来。本文系统性、全面的回顾了自然语言处理的代表性工作和最新进展,并按照类别系统性的介绍了自然语言处理领域的预训练模型。首先我们简要介绍了预训练模型,以及不同的模型特点和框架。之后,我们介绍并分析了预训练模型的影响和挑战以及下游任务中的应用。最后,我们简要总结并阐述了预训练模型未来的研究方向。展开更多
The goal of street-to-aerial cross-view image geo-localization is to determine the location of the query street-view image by retrieving the aerial-view image from the same place.The drastic viewpoint and appearance g...The goal of street-to-aerial cross-view image geo-localization is to determine the location of the query street-view image by retrieving the aerial-view image from the same place.The drastic viewpoint and appearance gap between the aerial-view and the street-view images brings a huge challenge against this task.In this paper,we propose a novel multiscale attention encoder to capture the multiscale contextual information of the aerial/street-view images.To bridge the domain gap between these two view images,we first use an inverse polar transform to make the street-view images approximately aligned with the aerial-view images.Then,the explored multiscale attention encoder is applied to convert the image into feature representation with the guidance of the learnt multiscale information.Finally,we propose a novel global mining strategy to enable the network to pay more attention to hard negative exemplars.Experiments on standard benchmark datasets show that our approach obtains 81.39%top-1 recall rate on the CVUSA dataset and 71.52%on the CVACT dataset,achieving the state-of-the-art performance and outperforming most of the existing methods significantly.展开更多
We introduce a new code for cosmological simulations, PHo To Ns, which incorporates features for performing massive cosmological simulations on heterogeneous high performance computer(HPC) systems and threads oriented...We introduce a new code for cosmological simulations, PHo To Ns, which incorporates features for performing massive cosmological simulations on heterogeneous high performance computer(HPC) systems and threads oriented programming. PHo To Ns adopts a hybrid scheme to compute gravitational force, with the conventional Particle-Mesh(PM) algorithm to compute the long-range force,the Tree algorithm to compute the short range force and the direct summation Particle-Particle(PP) algorithm to compute gravity from very close particles. A self-similar space filling a Peano-Hilbert curve is used to decompose the computing domain. Threads programming is advantageously used to more flexibly manage the domain communication, PM calculation and synchronization, as well as Dual Tree Traversal on the CPU+MIC platform. PHo To Ns scales well and efficiency of the PP kernel achieves68.6% of peak performance on MIC and 74.4% on CPU platforms. We also test the accuracy of the code against the much used Gadget-2 in the community and found excellent agreement.展开更多
Physics-informed neural networks(PINNs)are proved methods that are effective in solving some strongly nonlinear partial differential equations(PDEs),e.g.,Navier-Stokes equations,with a small amount of boundary or inte...Physics-informed neural networks(PINNs)are proved methods that are effective in solving some strongly nonlinear partial differential equations(PDEs),e.g.,Navier-Stokes equations,with a small amount of boundary or interior data.However,the feasibility of applying PINNs to the flow at moderate or high Reynolds numbers has rarely been reported.The present paper proposes an artificial viscosity(AV)-based PINN for solving the forward and inverse flow problems.Specifically,the AV used in PINNs is inspired by the entropy viscosity method developed in conventional computational fluid dynamics(CFD)to stabilize the simulation of flow at high Reynolds numbers.The newly developed PINN is used to solve the forward problem of the two-dimensional steady cavity flow at Re=1000 and the inverse problem derived from two-dimensional film boiling.The results show that the AV augmented PINN can solve both problems with good accuracy and substantially reduce the inference errors in the forward problem.展开更多
Recent advances in spatially resolved transcriptomic technologies have enabled unprecedented opportunities to elucidate tissue architecture and function in situ.Spatial transcriptomics can provide multimodal and compl...Recent advances in spatially resolved transcriptomic technologies have enabled unprecedented opportunities to elucidate tissue architecture and function in situ.Spatial transcriptomics can provide multimodal and complementary information simultaneously,including gene expression profiles,spatial locations,and histology images.However,most existing methods have limitations in efficiently utilizing spatial information and matched high-resolution histology images.To fully leverage the multi-modal information,we propose a SPAtially embedded Deep Attentional graph Clustering(SpaDAC)method to identify spatial domains while reconstructing denoised gene expression profiles.This method can efficiently learn the low-dimensional embeddings for spatial transcriptomics data by constructing multi-view graph modules to capture both spatial location connectives and morphological connectives.Benchmark results demonstrate that SpaDAC outperforms other algorithms on several recent spatial transcriptomics datasets.SpaDAC is a valuable tool for spatial domain detection,facilitating the comprehension of tissue architecture and cellular microenvironment.The source code of SpaDAC is freely available at Github(https://github.com/huoyuying/SpaDAC.git).展开更多
This paper analyzes the characteristics of emotion state and group behavior in the evacuation process.During the emergency evacuation,emotion state and group behavior are interacting with each other,and indivisible.Th...This paper analyzes the characteristics of emotion state and group behavior in the evacuation process.During the emergency evacuation,emotion state and group behavior are interacting with each other,and indivisible.The emotion spread model with the effect of group behavior,and the leader-follower model with the effect of emotion state are proposed.On this basis,exit choice strategies with the effect of emotion state and group behavior are proposed.Fusing emotion spread model,leader-follower model,and exit choice strategies into a cellular automata(CA)-based pedestrian simulation model,we simulate the evacuation process in a multi-exit case.Simulation results indicate that panic emotion and group behavior are two negative influence factors for pedestrian evacuation.Compared with panic emotion or group behavior only,pedestrian evacuation efficiency with the effects of both is lower.展开更多
An obstacle perception system for intelligent vehicle is proposed.The proposed system combines the stereo version technique and the deep learning network model,and is applied to obstacle perception tasks in complex en...An obstacle perception system for intelligent vehicle is proposed.The proposed system combines the stereo version technique and the deep learning network model,and is applied to obstacle perception tasks in complex environment.In this paper,we provide a complete system design project,which includes the hardware parameters,software framework,algorithm principle,and optimization method.In addition,special experiments are designed to demonstrate that the performance of the proposed system meets the requirements of actual application.The experiment results show that the proposed system is valid to both standard obstacles and non-standard obstacles,and suitable for different weather and lighting conditions in complex environment.It announces that the proposed system is flexible and robust to the intelligent vehicle.展开更多
Taking autonomous driving and driverless as the research object,we discuss and define intelligent high-precision map.Intelligent high-precision map is considered as a key link of future travel,a carrier of real-time p...Taking autonomous driving and driverless as the research object,we discuss and define intelligent high-precision map.Intelligent high-precision map is considered as a key link of future travel,a carrier of real-time perception of traffic resources in the entire space-time range,and the criterion for the operation and control of the whole process of the vehicle.As a new form of map,it has distinctive features in terms of cartography theory and application requirements compared with traditional navigation electronic maps.Thus,it is necessary to analyze and discuss its key features and problems to promote the development of research and application of intelligent high-precision map.Accordingly,we propose an information transmission model based on the cartography theory and combine the wheeled robot’s control flow in practical application.Next,we put forward the data logic structure of intelligent high-precision map,and analyze its application in autonomous driving.Then,we summarize the computing mode of“Crowdsourcing+Edge-Cloud Collaborative Computing”,and carry out key technical analysis on how to improve the quality of crowdsourced data.We also analyze the effective application scenarios of intelligent high-precision map in the future.Finally,we present some thoughts and suggestions for the future development of this field.展开更多
Traditional Chinese medicine prescription is one of the treasures of traditional Chinese medicine(TCM).There are tens of thousands TCM prescriptions accumulated in the past thousands of years,corresponding to differen...Traditional Chinese medicine prescription is one of the treasures of traditional Chinese medicine(TCM).There are tens of thousands TCM prescriptions accumulated in the past thousands of years,corresponding to different diseases,symptoms and therapeutic goals.The correspondences are so complicated that there is an urgent need to leverage new technologies such as artificial intelligence(AI)to analyze,understand and utilize them effectively.In this paper,we present a brief overview of this direction,where current research progress on TCM prescription powered by AI is summarized.Our summarization focuses on three aspects,TCM prescription mining that aims at understanding the TCM prescription,TCM prescription or herb knowledge base construction that aims at extracting knowledge to support the TCM prescription-related study,and TCM prescription discovery that aims at utilizing AI technologies to further energize TCM.It is encouraging to see that steady progress in this direction has been made recently.Besides,a toy experiment on image-based TCM herb recognition by using convolutional neural networks is also conducted.It basically verifies that it is promising to use AI technologies to address challenging tasks in TCM.We also point out several research topics that could be cooperatively performed by researchers from the two disciplines.展开更多
Identical-by-descent(IBD)is a fundamental genomic characteristic in population genetics and has been widely used for population history reconstruction.However,limited by the nature of IBD,which could only capture the ...Identical-by-descent(IBD)is a fundamental genomic characteristic in population genetics and has been widely used for population history reconstruction.However,limited by the nature of IBD,which could only capture the relationship between two individuals/haplotypes,existing IBD-based history inference is constrained to two populations.In this study,we propose a framework by leveraging IBD sharing in multipopulation and develop a method,MatrixiBD,to reconstruct recent multi-population migration history.Specifically,we employ the structured coalescent theory to precisely model the genealogical process and then estimate the IBD sharing across multiple populations.Within our model,we establish a theoretical connection between migration history and IBD sharing.Our method is rigorously evaluated through simulations,revealing its remarkable accuracy and robustness.Furthermore,we apply MatrixiBD to Central and South Asia in the Human Genome Diversity Project and successfully reconstruct the recent migration history of three closely related populations in South Asia.By taking into account the IBD sharing across multiple populations simultaneously,MatrixlBD enables us to attain clearer and more comprehensive insights into the history of regions characterized by complex migration dynamics,providing a holistic perspective on intricate patterns embedded within the recent population migration history.展开更多
Background Gesture recognition has attracted significant attention because of its wide range of potential applications.Although multi-modal gesture recognition has made significant progress in recent years,a popular m...Background Gesture recognition has attracted significant attention because of its wide range of potential applications.Although multi-modal gesture recognition has made significant progress in recent years,a popular method still is simply fusing prediction scores at the end of each branch,which often ignores complementary features among different modalities in the early stage and does not fuse the complementary features into a more discriminative feature.Methods This paper proposes an Adaptive Cross-modal Weighting(ACmW)scheme to exploit complementarity features from RGB-D data in this study.The scheme learns relations among different modalities by combining the features of different data streams.The proposed ACmW module contains two key functions:(1)fusing complementary features from multiple streams through an adaptive one-dimensional convolution;and(2)modeling the correlation of multi-stream complementary features in the time dimension.Through the effective combination of these two functional modules,the proposed ACmW can automatically analyze the relationship between the complementary features from different streams,and can fuse them in the spatial and temporal dimensions.Results Extensive experiments validate the effectiveness of the proposed method,and show that our method outperforms state-of-the-art methods on IsoGD and NVGesture.展开更多
Artificial intelligence will transform the faces of many industry sectors,and also creates a large amount of opportunities.We have witnessed,in the past 10 to 20years,a change within the network industry.I also belive...Artificial intelligence will transform the faces of many industry sectors,and also creates a large amount of opportunities.We have witnessed,in the past 10 to 20years,a change within the network industry.I also belive that something big will happen in the next 10 to 20years.We hold great expectations to the展开更多
基金partially supported by the National Natural Science Foundation of China(62271485)the SDHS Science and Technology Project(HS2023B044)
文摘Imputation of missing data has long been an important topic and an essential application for intelligent transportation systems(ITS)in the real world.As a state-of-the-art generative model,the diffusion model has proven highly successful in image generation,speech generation,time series modelling etc.and now opens a new avenue for traffic data imputation.In this paper,we propose a conditional diffusion model,called the implicit-explicit diffusion model,for traffic data imputation.This model exploits both the implicit and explicit feature of the data simultaneously.More specifically,we design two types of feature extraction modules,one to capture the implicit dependencies hidden in the raw data at multiple time scales and the other to obtain the long-term temporal dependencies of the time series.This approach not only inherits the advantages of the diffusion model for estimating missing data,but also takes into account the multiscale correlation inherent in traffic data.To illustrate the performance of the model,extensive experiments are conducted on three real-world time series datasets using different missing rates.The experimental results demonstrate that the model improves imputation accuracy and generalization capability.
文摘With the growing application of intelligent robots in service,manufacturing,and medical fields,efficient and natural interaction between humans and robots has become key to improving collaboration efficiency and user experience.Gesture recognition,as an intuitive and contactless interaction method,can overcome the limitations of traditional interfaces and enable real-time control and feedback of robot movements and behaviors.This study first reviews mainstream gesture recognition algorithms and their application on different sensing platforms(RGB cameras,depth cameras,and inertial measurement units).It then proposes a gesture recognition method based on multimodal feature fusion and a lightweight deep neural network that balances recognition accuracy with computational efficiency.At system level,a modular human-robot interaction architecture is constructed,comprising perception,decision,and execution layers,and gesture commands are transmitted and mapped to robot actions in real time via the ROS communication protocol.Through multiple comparative experiments on public gesture datasets and a self-collected dataset,the proposed method’s superiority is validated in terms of accuracy,response latency,and system robustness,while user-experience tests assess the interface’s usability.The results provide a reliable technical foundation for robot collaboration and service in complex scenarios,offering broad prospects for practical application and deployment.
基金funded by National Natural Science Foundation of China(61603245).
文摘Knowledge distillation(KD)is an emerging model compression technique for learning compact object detector models.Previous KD often focused solely on distilling from the logits layer or the feature intermediate layers,which may limit the comprehensive learning of the student network.Additionally,the imbalance between the foreground and background also affects the performance of the model.To address these issues,this paper employs feature-based distillation to enhance the detection performance of the bounding box localization part,and logit-based distillation to improve the detection performance of the category prediction part.Specifically,for the intermediate layer feature distillation,we introduce feature resampling to reduce the risk of the student model merely imitating the teacher model.At the same time,we incorporate a Spatial Attention Mechanism(SAM)to highlight the foreground features learned by the student model.In terms of output layer feature distillation,we divide the traditional distillation targets into target-class objects and non-target-class objects,aiming to improve overall distillation performance.Furthermore,we introduce a one-to-many matching distillation strategy based on Feature Alignment Module(FAM),which further enhances the studentmodel’s feature representation ability,making its feature distribution closer to that of the teacher model,and thus demonstrating superior localization and classification capabilities in object detection tasks.Experimental results demonstrate that our proposedmethodology outperforms conventional distillation techniques in terms of object detecting performance.
基金supported by National Natural Science Foundation of China(Grant No. 51175287)National Science and Technology Major Project(Grant No. 2011ZX02403)
文摘Content-based 3D model retrieval is of great help to facilitate the reuse of existing designs and to inspire designers during conceptual design. However, there is still a gap to apply it in industry due to the low time efficiency. This paper presents two new methods with high efficiency to build a Content-based 3D model retrieval system. First, an improvement is made on the ＂Shape Distribution （D2）＂ algorithm, and a new algorithm named ＂Quick D2＂ is proposed. Four sample 3D mechanical models are used in an experiment to compare the time cost of the two algorithms. The result indicates that the time cost of Quick D2 is much lower than that of D2, while the descriptors extracted by the two algorithms are almost the same. Second, an expandable 3D model repository index method with high performance, namely, RBK index, is presented. On the basis of RBK index, the search space is pruned effectively during the search process, leading to a speed up of the whole system. The factors that influence the values of the key parameters of RBK index are discussed and an experimental method to find the optimal values of the key parameters is given. Finally, ＂3D Searcher＂, a content-based 3D model retrieval system is developed. By using the methods proposed, the time cost for the system to respond one query online is reduced by 75% on average. The system has been implemented in a manufacturing enterprise, and practical query examples during a case of the automobile rear axle design are also shown. The research method presented shows a new research perspective and can effectively improve the content-based 3D model retrieval efficiency.
文摘After more than 70 years of evolution,great achievements have been made in machine translation.Especially in recent years,translation quality has been greatly improved with the emergence of neural machine translation(NMT).In this article,we first review the history of machine translation from rule-based machine translation to example-based machine translation and statistical machine translation.We then introduce NMT in more detail,including the basic framework and the current dominant framework,Transformer,as well as multilingual translation models to deal with the data sparseness problem.In addition,we introduce cutting-edge simultaneous translation methods that achieve a balance between translation quality and latency.We then describe various products and applications of machine translation.At the end of this article,we briefly discuss challenges and future research directions in this field.
基金Supported by the National Natural Science Foundation of China (60673139, 60473073, 60573090)
文摘The paper presents a novel benefit based query processing strategy for efficient query routing. Based on DHT as the overlay network, it first applies Nash equilibrium to construct the optimal peer group based on the correlations of keywords and coverage and overlap of the peers to decrease the time cost, and then presents a two-layered architecture for query processing that utilizes Bloom filter as compact representation to reduce the bandwidth consumption. Extensive experiments conducted on a real world dataset have demonstrated that our approach obviously decreases the processing time, while improves the precision and recall as well.
文摘预训练语言模型(pre-trained languages model,PTLM)在自然语言处理(natural language processing,NLP)领域取得了令人瞩目的成功,并由此引发了下游任务从监督学习到预训练-微调范式的转变。在此之后,一系列预训练模型的创新研究涌现出来。本文系统性、全面的回顾了自然语言处理的代表性工作和最新进展,并按照类别系统性的介绍了自然语言处理领域的预训练模型。首先我们简要介绍了预训练模型,以及不同的模型特点和框架。之后,我们介绍并分析了预训练模型的影响和挑战以及下游任务中的应用。最后,我们简要总结并阐述了预训练模型未来的研究方向。
基金National Natural Science Foundation of China,Grant/Award Number:62106177supported by the Central University Basic Research Fund of China(No.2042020KF0016)supported by the supercomputing system in the Supercomputing Center of Wuhan University.
文摘The goal of street-to-aerial cross-view image geo-localization is to determine the location of the query street-view image by retrieving the aerial-view image from the same place.The drastic viewpoint and appearance gap between the aerial-view and the street-view images brings a huge challenge against this task.In this paper,we propose a novel multiscale attention encoder to capture the multiscale contextual information of the aerial/street-view images.To bridge the domain gap between these two view images,we first use an inverse polar transform to make the street-view images approximately aligned with the aerial-view images.Then,the explored multiscale attention encoder is applied to convert the image into feature representation with the guidance of the learnt multiscale information.Finally,we propose a novel global mining strategy to enable the network to pay more attention to hard negative exemplars.Experiments on standard benchmark datasets show that our approach obtains 81.39%top-1 recall rate on the CVUSA dataset and 71.52%on the CVACT dataset,achieving the state-of-the-art performance and outperforming most of the existing methods significantly.
基金support from the National Key Program for Science and Technology Research and Development (2017YFB0203300)the National Natural Science Foundation of China (Grant Nos. 11403035, 11425312 and 11573030)support from Royal Society Newton advanced Fellowships
文摘We introduce a new code for cosmological simulations, PHo To Ns, which incorporates features for performing massive cosmological simulations on heterogeneous high performance computer(HPC) systems and threads oriented programming. PHo To Ns adopts a hybrid scheme to compute gravitational force, with the conventional Particle-Mesh(PM) algorithm to compute the long-range force,the Tree algorithm to compute the short range force and the direct summation Particle-Particle(PP) algorithm to compute gravity from very close particles. A self-similar space filling a Peano-Hilbert curve is used to decompose the computing domain. Threads programming is advantageously used to more flexibly manage the domain communication, PM calculation and synchronization, as well as Dual Tree Traversal on the CPU+MIC platform. PHo To Ns scales well and efficiency of the PP kernel achieves68.6% of peak performance on MIC and 74.4% on CPU platforms. We also test the accuracy of the code against the much used Gadget-2 in the community and found excellent agreement.
基金Project supported by the Fundamental Research Funds for the Central Universities of China(No.DUT21RC(3)063)the National Natural Science Foundation of China(No.51720105007)the Baidu Foundation(No.ghfund202202014542)。
文摘Physics-informed neural networks(PINNs)are proved methods that are effective in solving some strongly nonlinear partial differential equations(PDEs),e.g.,Navier-Stokes equations,with a small amount of boundary or interior data.However,the feasibility of applying PINNs to the flow at moderate or high Reynolds numbers has rarely been reported.The present paper proposes an artificial viscosity(AV)-based PINN for solving the forward and inverse flow problems.Specifically,the AV used in PINNs is inspired by the entropy viscosity method developed in conventional computational fluid dynamics(CFD)to stabilize the simulation of flow at high Reynolds numbers.The newly developed PINN is used to solve the forward problem of the two-dimensional steady cavity flow at Re=1000 and the inverse problem derived from two-dimensional film boiling.The results show that the AV augmented PINN can solve both problems with good accuracy and substantially reduce the inference errors in the forward problem.
基金supported by National Natural Science Foundation of China(62003028).X.L.was supported by a Scholarship from the China Scholarship Council.
文摘Recent advances in spatially resolved transcriptomic technologies have enabled unprecedented opportunities to elucidate tissue architecture and function in situ.Spatial transcriptomics can provide multimodal and complementary information simultaneously,including gene expression profiles,spatial locations,and histology images.However,most existing methods have limitations in efficiently utilizing spatial information and matched high-resolution histology images.To fully leverage the multi-modal information,we propose a SPAtially embedded Deep Attentional graph Clustering(SpaDAC)method to identify spatial domains while reconstructing denoised gene expression profiles.This method can efficiently learn the low-dimensional embeddings for spatial transcriptomics data by constructing multi-view graph modules to capture both spatial location connectives and morphological connectives.Benchmark results demonstrate that SpaDAC outperforms other algorithms on several recent spatial transcriptomics datasets.SpaDAC is a valuable tool for spatial domain detection,facilitating the comprehension of tissue architecture and cellular microenvironment.The source code of SpaDAC is freely available at Github(https://github.com/huoyuying/SpaDAC.git).
基金Project supported by the National Key Research and Development Program of China(Grant No.2017YFC0803903)the National Natural Science Foundation of China(Grant No.62003182)。
文摘This paper analyzes the characteristics of emotion state and group behavior in the evacuation process.During the emergency evacuation,emotion state and group behavior are interacting with each other,and indivisible.The emotion spread model with the effect of group behavior,and the leader-follower model with the effect of emotion state are proposed.On this basis,exit choice strategies with the effect of emotion state and group behavior are proposed.Fusing emotion spread model,leader-follower model,and exit choice strategies into a cellular automata(CA)-based pedestrian simulation model,we simulate the evacuation process in a multi-exit case.Simulation results indicate that panic emotion and group behavior are two negative influence factors for pedestrian evacuation.Compared with panic emotion or group behavior only,pedestrian evacuation efficiency with the effects of both is lower.
基金supported by the National Natural Science Foundation of China(61673381)the National Key R&D Program of China(2018AAA0103103)the Science and Technology Development Fund(0024/2018/A1)。
文摘An obstacle perception system for intelligent vehicle is proposed.The proposed system combines the stereo version technique and the deep learning network model,and is applied to obstacle perception tasks in complex environment.In this paper,we provide a complete system design project,which includes the hardware parameters,software framework,algorithm principle,and optimization method.In addition,special experiments are designed to demonstrate that the performance of the proposed system meets the requirements of actual application.The experiment results show that the proposed system is valid to both standard obstacles and non-standard obstacles,and suitable for different weather and lighting conditions in complex environment.It announces that the proposed system is flexible and robust to the intelligent vehicle.
基金National Key Research and Development Program(No.2018YFB1305001)Major Consulting and Research Project of Chinese Academy of Engineering(No.2018-ZD-02-07)。
文摘Taking autonomous driving and driverless as the research object,we discuss and define intelligent high-precision map.Intelligent high-precision map is considered as a key link of future travel,a carrier of real-time perception of traffic resources in the entire space-time range,and the criterion for the operation and control of the whole process of the vehicle.As a new form of map,it has distinctive features in terms of cartography theory and application requirements compared with traditional navigation electronic maps.Thus,it is necessary to analyze and discuss its key features and problems to promote the development of research and application of intelligent high-precision map.Accordingly,we propose an information transmission model based on the cartography theory and combine the wheeled robot’s control flow in practical application.Next,we put forward the data logic structure of intelligent high-precision map,and analyze its application in autonomous driving.Then,we summarize the computing mode of“Crowdsourcing+Edge-Cloud Collaborative Computing”,and carry out key technical analysis on how to improve the quality of crowdsourced data.We also analyze the effective application scenarios of intelligent high-precision map in the future.Finally,we present some thoughts and suggestions for the future development of this field.
基金the National Natural Science Foundation of China(No.2019YFC1710400,2019YFC1710404).
文摘Traditional Chinese medicine prescription is one of the treasures of traditional Chinese medicine(TCM).There are tens of thousands TCM prescriptions accumulated in the past thousands of years,corresponding to different diseases,symptoms and therapeutic goals.The correspondences are so complicated that there is an urgent need to leverage new technologies such as artificial intelligence(AI)to analyze,understand and utilize them effectively.In this paper,we present a brief overview of this direction,where current research progress on TCM prescription powered by AI is summarized.Our summarization focuses on three aspects,TCM prescription mining that aims at understanding the TCM prescription,TCM prescription or herb knowledge base construction that aims at extracting knowledge to support the TCM prescription-related study,and TCM prescription discovery that aims at utilizing AI technologies to further energize TCM.It is encouraging to see that steady progress in this direction has been made recently.Besides,a toy experiment on image-based TCM herb recognition by using convolutional neural networks is also conducted.It basically verifies that it is promising to use AI technologies to address challenging tasks in TCM.We also point out several research topics that could be cooperatively performed by researchers from the two disciplines.
基金supported by the Fundamental Research Funds for the Central Universities(2023JBMC011)the National Natural Science Foundation of China(NSFC)Grant(12271026)the Beijing Natural Science Foundation Grant(L222051).
文摘Identical-by-descent(IBD)is a fundamental genomic characteristic in population genetics and has been widely used for population history reconstruction.However,limited by the nature of IBD,which could only capture the relationship between two individuals/haplotypes,existing IBD-based history inference is constrained to two populations.In this study,we propose a framework by leveraging IBD sharing in multipopulation and develop a method,MatrixiBD,to reconstruct recent multi-population migration history.Specifically,we employ the structured coalescent theory to precisely model the genealogical process and then estimate the IBD sharing across multiple populations.Within our model,we establish a theoretical connection between migration history and IBD sharing.Our method is rigorously evaluated through simulations,revealing its remarkable accuracy and robustness.Furthermore,we apply MatrixiBD to Central and South Asia in the Human Genome Diversity Project and successfully reconstruct the recent migration history of three closely related populations in South Asia.By taking into account the IBD sharing across multiple populations simultaneously,MatrixlBD enables us to attain clearer and more comprehensive insights into the history of regions characterized by complex migration dynamics,providing a holistic perspective on intricate patterns embedded within the recent population migration history.
基金the Chinese National Natural Science Foundation Projects(61961160704,61876179)the Key Project of the General Logistics Department(ASW17C001)the Science and Technology Development Fund of Macao(0010/2019/AFJ,0025/2019/AKP).
文摘Background Gesture recognition has attracted significant attention because of its wide range of potential applications.Although multi-modal gesture recognition has made significant progress in recent years,a popular method still is simply fusing prediction scores at the end of each branch,which often ignores complementary features among different modalities in the early stage and does not fuse the complementary features into a more discriminative feature.Methods This paper proposes an Adaptive Cross-modal Weighting(ACmW)scheme to exploit complementarity features from RGB-D data in this study.The scheme learns relations among different modalities by combining the features of different data streams.The proposed ACmW module contains two key functions:(1)fusing complementary features from multiple streams through an adaptive one-dimensional convolution;and(2)modeling the correlation of multi-stream complementary features in the time dimension.Through the effective combination of these two functional modules,the proposed ACmW can automatically analyze the relationship between the complementary features from different streams,and can fuse them in the spatial and temporal dimensions.Results Extensive experiments validate the effectiveness of the proposed method,and show that our method outperforms state-of-the-art methods on IsoGD and NVGesture.
文摘Artificial intelligence will transform the faces of many industry sectors,and also creates a large amount of opportunities.We have witnessed,in the past 10 to 20years,a change within the network industry.I also belive that something big will happen in the next 10 to 20years.We hold great expectations to the
