Facial beauty analysis is an important topic in human society.It may be used as a guidance for face beautification applications such as cosmetic surgery.Deep neural networks(DNNs)have recently been adopted for facial ...Facial beauty analysis is an important topic in human society.It may be used as a guidance for face beautification applications such as cosmetic surgery.Deep neural networks(DNNs)have recently been adopted for facial beauty analysis and have achieved remarkable performance.However,most existing DNN-based models regard facial beauty analysis as a normal classification task.They ignore important prior knowledge in traditional machine learning models which illustrate the significant contribution of the geometric features in facial beauty analysis.To be specific,landmarks of the whole face and facial organs are introduced to extract geometric features to make the decision.Inspired by this,we introduce a novel dual-branch network for facial beauty analysis:one branch takes the Swin Transformer as the backbone to model the full face and global patterns,and another branch focuses on the masked facial organs with the residual network to model the local patterns of certain facial parts.Additionally,the designed multi-scale feature fusion module can further facilitate our network to learn complementary semantic information between the two branches.In model optimisation,we propose a hybrid loss function,where especially geometric regulation is introduced by regressing the facial landmarks and it can force the extracted features to convey facial geometric features.Experiments performed on the SCUT-FBP5500 dataset and the SCUT-FBP dataset demonstrate that our model outperforms the state-of-the-art convolutional neural networks models,which proves the effectiveness of the proposed geometric regularisation and dual-branch structure with the hybrid network.To the best of our knowledge,this is the first study to introduce a Vision Transformer into the facial beauty analysis task.展开更多
Diabetic retinopathy(DR),the main cause of irreversible blindness,is one of the most common complications of diabetes.At present,deep convolutional neural networks have achieved promising performance in automatic DR d...Diabetic retinopathy(DR),the main cause of irreversible blindness,is one of the most common complications of diabetes.At present,deep convolutional neural networks have achieved promising performance in automatic DR detection tasks.The convolution operation of methods is a local cross-correlation operation,whose receptive field de-termines the size of the local neighbourhood for processing.However,for retinal fundus photographs,there is not only the local information but also long-distance dependence between the lesion features(e.g.hemorrhages and exudates)scattered throughout the whole image.The proposed method incorporates correlations between long-range patches into the deep learning framework to improve DR detection.Patch-wise re-lationships are used to enhance the local patch features since lesions of DR usually appear as plaques.The Long-Range unit in the proposed network with a residual structure can be flexibly embedded into other trained networks.Extensive experimental results demon-strate that the proposed approach can achieve higher accuracy than existing state-of-the-art models on Messidor and EyePACS datasets.展开更多
Dynamic Simultaneous Localization and Mapping(SLAM)in visual scenes is currently a major research area in fields such as robot navigation and autonomous driving.However,in the face of complex real-world envi-ronments,...Dynamic Simultaneous Localization and Mapping(SLAM)in visual scenes is currently a major research area in fields such as robot navigation and autonomous driving.However,in the face of complex real-world envi-ronments,current dynamic SLAM systems struggle to achieve precise localization and map construction.With the advancement of deep learning,there has been increasing interest in the development of deep learning-based dynamic SLAM visual odometry in recent years,and more researchers are turning to deep learning techniques to address the challenges of dynamic SLAM.Compared to dynamic SLAM systems based on deep learning methods such as object detection and semantic segmentation,dynamic SLAM systems based on instance segmentation can not only detect dynamic objects in the scene but also distinguish different instances of the same type of object,thereby reducing the impact of dynamic objects on the SLAM system’s positioning.This article not only introduces traditional dynamic SLAM systems based on mathematical models but also provides a comprehensive analysis of existing instance segmentation algorithms and dynamic SLAM systems based on instance segmentation,comparing and summarizing their advantages and disadvantages.Through comparisons on datasets,it is found that instance segmentation-based methods have significant advantages in accuracy and robustness in dynamic environments.However,the real-time performance of instance segmentation algorithms hinders the widespread application of dynamic SLAM systems.In recent years,the rapid development of single-stage instance segmentationmethods has brought hope for the widespread application of dynamic SLAM systems based on instance segmentation.Finally,possible future research directions and improvementmeasures are discussed for reference by relevant professionals.展开更多
Although selective nanofiltration(SNF)and selective electrodialysis(SED)have been widely adopted in the field of Mg^(2+)/Li^(+)separation,their differences have not been illustrated systematically.In this study,for th...Although selective nanofiltration(SNF)and selective electrodialysis(SED)have been widely adopted in the field of Mg^(2+)/Li^(+)separation,their differences have not been illustrated systematically.In this study,for the first time,SNF and SED processes in continuous mode were studied for Li+fractionation from the same brine with high Mg/Li ratios and their differences were discussed in detail.For a fair analysis of the two processes,typical factors were optimized.Specifically,the optimal operating pressure and feed flow rate for SNF were 2.4 MPa and 140 L·h^(-1),respectively,while the optimal cell-pair voltage and replenishment flow rate for SED were 1.0 V and 14 L·h^(-1),respectively.Although the Li^(+)fractionation capacity of the two processes were similar,the selectivity coefficient of SNF was 24.7% higher than that of SED and,thus,the Mg/Li ratio in purified stream of the former was 19.0% lower than that of the latter.Due to higher ion driving force,SED had clear advantages in recovery ratio and concentration effects.Meanwhile,the specific energy consumption of SED was 20.1% lower than that of SNF.This study provided a better understanding and guidance for the application and improvement of the two technologies.展开更多
The in-core self-powered neutron detector(SPND)acts as a key measuring device for the monitoring of parameters and evaluation of the operating conditions of nuclear reactors.Prompt detection and tolerance of faulty SP...The in-core self-powered neutron detector(SPND)acts as a key measuring device for the monitoring of parameters and evaluation of the operating conditions of nuclear reactors.Prompt detection and tolerance of faulty SPNDs are indispensable for reliable reactor management.To completely extract the correlated state information of SPNDs,we constructed a twin model based on a generalized regression neural network(GRNN)that represents the common relationships among overall signals.Faulty SPNDs were determined because of the functional concordance of the twin model and real monitoring sys-tems,which calculated the error probability distribution between the model outputs and real values.Fault detection follows a tolerance phase to reinforce the stability of the twin model in the case of massive failures.A weighted K-nearest neighbor model was employed to reasonably reconstruct the values of the faulty signals and guarantee data purity.The experimental evaluation of the proposed method showed promising results,with excellent output consistency and high detection accuracy for both single-and multiple-point faulty SPNDs.For unexpected excessive failures,the proposed tolerance approach can efficiently repair fault behaviors and enhance the prediction performance of the twin model.展开更多
Difenoconazole(DIF)is a representative variety of broad-spectrum triazole fungicides and liposoluble pesticides.However,the water solubility of DIF is so poor that its application is limited in plant protection.In add...Difenoconazole(DIF)is a representative variety of broad-spectrum triazole fungicides and liposoluble pesticides.However,the water solubility of DIF is so poor that its application is limited in plant protection.In addition,the conventional formulations of DIF always contain abundant organic solvents,which may cause pollution of the environment.In this study,two DIF/cyclodextrins(CDs)inclusion complexes(ICs)were successfully prepared,which were DIF/β-CD IC and DIF/hydroxypropyl-β-CD IC(DIF/HP-β-CD IC).The effect of cyclodextrins on the water solubility and the antifungal effect of liposoluble DIF pesticide were investigated.According to the phase solubility test,the molar ratio and apparent stability constant of ICs were obtained.Fourier transform infrared spectroscopy,thermal gravity analysis,X-ray diffraction and scanning electron microscopy were used systematically to characterize the formation and characteristics of ICs.The results noted that DIF successfully entered the cavities of two CDs.In addition,the antifungal effect test proved the better performance of DIF/HP-β-CD IC,which exceeded that of DIF emulsifiable concentrate.Therefore,our study provides informative direction for the intelligent use of liposoluble pesticides with cyclodextrins to develop water-based environmentally friendly formulations.展开更多
Antiferromagnets offer great potential for high-speed data processing applications,as they can expend spintronic devices from a static storage and gigahertz frequency range to the terahertz range.However,their zero ne...Antiferromagnets offer great potential for high-speed data processing applications,as they can expend spintronic devices from a static storage and gigahertz frequency range to the terahertz range.However,their zero net magnetization makes them difficult to manipulate and detect.In recent years,there has been a lot of attention given to the ultrafast manipulation of magnetic order using ultra-short single laser pulses,but it remains unknown whether a similar scenario can be observed in antiferromagnets.In this work,we demonstrate the manipulation of antiferromagnets with a single femtosecond laser pulse in perpendicular exchange-biased Co/Ir Mn/Co Gd trilayers.We study the dual exchange bias interlayer interaction in quasi-static conditions and competition in ultrafast antiferromagnet rearrangement.Our results show that,compared to conventional ferromagnetic/antiferromagnetic systems,the Ir Mn antiferromagnet can be ultrafast and efficiently manipulated by the coupled Co Gd ferrimagnetic layer,which paves the way for potential energy-efficient spintronic devices.展开更多
Visual simultaneous localization and mapping(SLAM)is crucial in robotics and autonomous driving.However,traditional visual SLAM faces challenges in dynamic environments.To address this issue,researchers have proposed ...Visual simultaneous localization and mapping(SLAM)is crucial in robotics and autonomous driving.However,traditional visual SLAM faces challenges in dynamic environments.To address this issue,researchers have proposed semantic SLAM,which combines object detection,semantic segmentation,instance segmentation,and visual SLAM.Despite the growing body of literature on semantic SLAM,there is currently a lack of comprehensive research on the integration of object detection and visual SLAM.Therefore,this study aims to gather information from multiple databases and review relevant literature using specific keywords.It focuses on visual SLAM based on object detection,covering different aspects.Firstly,it discusses the current research status and challenges in this field,highlighting methods for incorporating semantic information from object detection networks into mileage measurement,closed-loop detection,and map construction.It also compares the characteristics and performance of various visual SLAM object detection algorithms.Lastly,it provides an outlook on future research directions and emerging trends in visual SLAM.Research has shown that visual SLAM based on object detection has significant improvements compared to traditional SLAM in dynamic point removal,data association,point cloud segmentation,and other technologies.It can improve the robustness and accuracy of the entire SLAM system and can run in real time.With the continuous optimization of algorithms and the improvement of hardware level,object visual SLAM has great potential for development.展开更多
In this paper,the path integral solutions for a general n-dimensional stochastic differential equa-tions(SDEs)withα-stable Lévy noise are derived and verified.Firstly,the governing equations for the solutions of...In this paper,the path integral solutions for a general n-dimensional stochastic differential equa-tions(SDEs)withα-stable Lévy noise are derived and verified.Firstly,the governing equations for the solutions of n-dimensional SDEs under the excitation ofα-stable Lévy noise are obtained through the characteristic function of stochastic processes.Then,the short-time transition probability density func-tion of the path integral solution is derived based on the Chapman-Kolmogorov-Smoluchowski(CKS)equation and the characteristic function,and its correctness is demonstrated by proving that it satis-fies the governing equation of the solution of the SDE,which is also called the Fokker-Planck-Kolmogorov equation.Besides,illustrative examples are numerically considered for highlighting the feasibility of the proposed path integral method,and the pertinent Monte Carlo solution is also calculated to show its correctness and effectiveness.展开更多
The robotics industry has seen rapid development in recent years due to the Corona Virus Disease 2019.With the development of sensors and smart devices,factories and enterprises have accumulated a large amount of data...The robotics industry has seen rapid development in recent years due to the Corona Virus Disease 2019.With the development of sensors and smart devices,factories and enterprises have accumulated a large amount of data in their daily production,which creates extremely favorable conditions for robots to perform machine learning.However,in recent years,people’s awareness of data privacy has been increasing,leading to the inability to circulate data between different enterprises,resulting in the emergence of data silos.The emergence of federated learning provides a feasible solution to this problem,and the combination of federated learning and multi-robot systems can break down data silos and improve the overall performance of robots.However,as scholars have studied more deeply,they found that federated learning has very limited privacy protection.Therefore,how to protect data privacy from infringement remains an important issue.In this paper,we first give a brief introduction to the current development of multi-robot and federated learning;second,we review three aspects of privacy protection methods commonly used,privacy protection methods for multi-robot,and Other Problems Faced by Multi-robot Systems,focusing on method comparisons and challenges;and finally draw conclusions and predict possible future research directions.展开更多
Deep learning has been widely used in the field of mammographic image classification owing to its superiority in automatic feature extraction.However,general deep learning models cannot achieve very satisfactory class...Deep learning has been widely used in the field of mammographic image classification owing to its superiority in automatic feature extraction.However,general deep learning models cannot achieve very satisfactory classification results on mammographic images because these models are not specifically designed for mammographic images and do not take the specific traits of these images into account.To exploit the essential discriminant information of mammographic images,we propose a novel classification method based on a convolutional neural network.Specifically,the proposed method designs two branches to extract the discriminative features from mammographic images from the mediolateral oblique and craniocaudal(CC)mammographic views.The features extracted from the two-view mammographic images contain complementary information that enables breast cancer to be more easily distinguished.Moreover,the attention block is introduced to capture the channel-wise information by adjusting the weight of each feature map,which is beneficial to emphasising the important features of mammographic images.Furthermore,we add a penalty term based on the fuzzy cluster algorithm to the cross-entropy function,which improves the generalisation ability of the classification model by maximising the interclass distance and minimising the intraclass distance of the samples.The experimental results on The Digital database for Screening Mammography INbreast and MIAS mammography databases illustrate that the proposed method achieves the best classification performance and is more robust than the compared state-ofthe-art classification methods.展开更多
Compared with the space on the ground,if there is a fire in the urban complex underground space,the loss will be greatly harmful.In addition,the complex underground space is usually connected with other large space ar...Compared with the space on the ground,if there is a fire in the urban complex underground space,the loss will be greatly harmful.In addition,the complex underground space is usually connected with other large space areas and densely populated.Once a fire occurs in the complex underground space,it will cause huge property losses and casualties.In order to reduce the risk of fire,it is necessary to deeply understand the development rules and characteristics of fire in the complex underground space of the city.This article has mainly carried on the following work:(I)A particularly complex model of the multi‐storey subway station was built.On this basis,three groups of comparative experiments were conducted to study the effects of fire power,fire location and smoke control system on fire development,and the conclusion that fire location is the most important factor for fire development was obtained;(II)In order to explore the entire space fire and the local space fire,CFD(Computational Fluid Dynamics)is used to build a large‐size fire model and a small‐size fire model respectively;(III)Multiple detector data as temperature slices were built,and it is expected to make full use of the simulation data to deduce the important index of fire location in the early stage of fire.All of the works in this paper will provide reference experimental data for the prevention and firefighting of a sudden fire in the complex underground space.展开更多
Nonlinear and stochastic dynamics is essential and practical in almost all branches of natural science and engineering.It has been a central subject to understand various complex dynamics,such as random vibration,stoc...Nonlinear and stochastic dynamics is essential and practical in almost all branches of natural science and engineering.It has been a central subject to understand various complex dynamics,such as random vibration,stochastic transition,synchronization,et al.,in the areas of mechanical and aerospace engineering,physics and chemistry.For example,the stochastic resonance has been utilized effectively in mechanical fault diagnosis and weak signal detection.Naturally,many methods have been developed to get the stochastic responses,including stochastic averaging method.展开更多
基金Shenzhen Science and Technology Program,Grant/Award Number:ZDSYS20211021111415025Shenzhen Institute of Artificial Intelligence and Robotics for SocietyYouth Science and Technology Talents Development Project of Guizhou Education Department,Grant/Award Number:QianJiaoheKYZi[2018]459。
文摘Facial beauty analysis is an important topic in human society.It may be used as a guidance for face beautification applications such as cosmetic surgery.Deep neural networks(DNNs)have recently been adopted for facial beauty analysis and have achieved remarkable performance.However,most existing DNN-based models regard facial beauty analysis as a normal classification task.They ignore important prior knowledge in traditional machine learning models which illustrate the significant contribution of the geometric features in facial beauty analysis.To be specific,landmarks of the whole face and facial organs are introduced to extract geometric features to make the decision.Inspired by this,we introduce a novel dual-branch network for facial beauty analysis:one branch takes the Swin Transformer as the backbone to model the full face and global patterns,and another branch focuses on the masked facial organs with the residual network to model the local patterns of certain facial parts.Additionally,the designed multi-scale feature fusion module can further facilitate our network to learn complementary semantic information between the two branches.In model optimisation,we propose a hybrid loss function,where especially geometric regulation is introduced by regressing the facial landmarks and it can force the extracted features to convey facial geometric features.Experiments performed on the SCUT-FBP5500 dataset and the SCUT-FBP dataset demonstrate that our model outperforms the state-of-the-art convolutional neural networks models,which proves the effectiveness of the proposed geometric regularisation and dual-branch structure with the hybrid network.To the best of our knowledge,this is the first study to introduce a Vision Transformer into the facial beauty analysis task.
基金National Natural Science Foundation of China,Grant/Award Numbers:62001141,62272319Science,Technology and Innovation Commission of Shenzhen Municipality,Grant/Award Numbers:GJHZ20210705141812038,JCYJ20210324094413037,JCYJ20210324131800002,RCBS20210609103820029Stable Support Projects for Shenzhen Higher Education Institutions,Grant/Award Number:20220715183602001。
文摘Diabetic retinopathy(DR),the main cause of irreversible blindness,is one of the most common complications of diabetes.At present,deep convolutional neural networks have achieved promising performance in automatic DR detection tasks.The convolution operation of methods is a local cross-correlation operation,whose receptive field de-termines the size of the local neighbourhood for processing.However,for retinal fundus photographs,there is not only the local information but also long-distance dependence between the lesion features(e.g.hemorrhages and exudates)scattered throughout the whole image.The proposed method incorporates correlations between long-range patches into the deep learning framework to improve DR detection.Patch-wise re-lationships are used to enhance the local patch features since lesions of DR usually appear as plaques.The Long-Range unit in the proposed network with a residual structure can be flexibly embedded into other trained networks.Extensive experimental results demon-strate that the proposed approach can achieve higher accuracy than existing state-of-the-art models on Messidor and EyePACS datasets.
基金the National Natural Science Foundation of China(No.62063006)the Natural Science Foundation of Guangxi Province(No.2023GXNS-FAA026025)+3 种基金the Innovation Fund of Chinese Universities Industry-University-Research(ID:2021RYC06005)the Research Project for Young andMiddle-Aged Teachers in Guangxi Universi-ties(ID:2020KY15013)the Special Research Project of Hechi University(ID:2021GCC028)financially supported by the Project of Outstanding Thousand Young Teachers’Training in Higher Education Institutions of Guangxi,Guangxi Colleges and Universities Key Laboratory of AI and Information Processing(Hechi University),Education Department of Guangxi Zhuang Autonomous Region.
文摘Dynamic Simultaneous Localization and Mapping(SLAM)in visual scenes is currently a major research area in fields such as robot navigation and autonomous driving.However,in the face of complex real-world envi-ronments,current dynamic SLAM systems struggle to achieve precise localization and map construction.With the advancement of deep learning,there has been increasing interest in the development of deep learning-based dynamic SLAM visual odometry in recent years,and more researchers are turning to deep learning techniques to address the challenges of dynamic SLAM.Compared to dynamic SLAM systems based on deep learning methods such as object detection and semantic segmentation,dynamic SLAM systems based on instance segmentation can not only detect dynamic objects in the scene but also distinguish different instances of the same type of object,thereby reducing the impact of dynamic objects on the SLAM system’s positioning.This article not only introduces traditional dynamic SLAM systems based on mathematical models but also provides a comprehensive analysis of existing instance segmentation algorithms and dynamic SLAM systems based on instance segmentation,comparing and summarizing their advantages and disadvantages.Through comparisons on datasets,it is found that instance segmentation-based methods have significant advantages in accuracy and robustness in dynamic environments.However,the real-time performance of instance segmentation algorithms hinders the widespread application of dynamic SLAM systems.In recent years,the rapid development of single-stage instance segmentationmethods has brought hope for the widespread application of dynamic SLAM systems based on instance segmentation.Finally,possible future research directions and improvementmeasures are discussed for reference by relevant professionals.
基金financial support by the National Key Research and Development Program of China(2017YFC0404003)the Tianjin Natural Science Foundation(21JCZDJC00270)+3 种基金the China Postdoctoral Science Foundation(2021M701875)the Tianjin Special Project of Ecological Environment Management Science and Technology(18ZXSZSF00050)the Tianjin Science and Technology Support Project(19YFZCSF00760)the Fundamental Research Funds for the Central Universities(63221312).
文摘Although selective nanofiltration(SNF)and selective electrodialysis(SED)have been widely adopted in the field of Mg^(2+)/Li^(+)separation,their differences have not been illustrated systematically.In this study,for the first time,SNF and SED processes in continuous mode were studied for Li+fractionation from the same brine with high Mg/Li ratios and their differences were discussed in detail.For a fair analysis of the two processes,typical factors were optimized.Specifically,the optimal operating pressure and feed flow rate for SNF were 2.4 MPa and 140 L·h^(-1),respectively,while the optimal cell-pair voltage and replenishment flow rate for SED were 1.0 V and 14 L·h^(-1),respectively.Although the Li^(+)fractionation capacity of the two processes were similar,the selectivity coefficient of SNF was 24.7% higher than that of SED and,thus,the Mg/Li ratio in purified stream of the former was 19.0% lower than that of the latter.Due to higher ion driving force,SED had clear advantages in recovery ratio and concentration effects.Meanwhile,the specific energy consumption of SED was 20.1% lower than that of SNF.This study provided a better understanding and guidance for the application and improvement of the two technologies.
基金supported by the Natural Science Foundation of Fujian Province,China(No.2022J01566).
文摘The in-core self-powered neutron detector(SPND)acts as a key measuring device for the monitoring of parameters and evaluation of the operating conditions of nuclear reactors.Prompt detection and tolerance of faulty SPNDs are indispensable for reliable reactor management.To completely extract the correlated state information of SPNDs,we constructed a twin model based on a generalized regression neural network(GRNN)that represents the common relationships among overall signals.Faulty SPNDs were determined because of the functional concordance of the twin model and real monitoring sys-tems,which calculated the error probability distribution between the model outputs and real values.Fault detection follows a tolerance phase to reinforce the stability of the twin model in the case of massive failures.A weighted K-nearest neighbor model was employed to reasonably reconstruct the values of the faulty signals and guarantee data purity.The experimental evaluation of the proposed method showed promising results,with excellent output consistency and high detection accuracy for both single-and multiple-point faulty SPNDs.For unexpected excessive failures,the proposed tolerance approach can efficiently repair fault behaviors and enhance the prediction performance of the twin model.
文摘Difenoconazole(DIF)is a representative variety of broad-spectrum triazole fungicides and liposoluble pesticides.However,the water solubility of DIF is so poor that its application is limited in plant protection.In addition,the conventional formulations of DIF always contain abundant organic solvents,which may cause pollution of the environment.In this study,two DIF/cyclodextrins(CDs)inclusion complexes(ICs)were successfully prepared,which were DIF/β-CD IC and DIF/hydroxypropyl-β-CD IC(DIF/HP-β-CD IC).The effect of cyclodextrins on the water solubility and the antifungal effect of liposoluble DIF pesticide were investigated.According to the phase solubility test,the molar ratio and apparent stability constant of ICs were obtained.Fourier transform infrared spectroscopy,thermal gravity analysis,X-ray diffraction and scanning electron microscopy were used systematically to characterize the formation and characteristics of ICs.The results noted that DIF successfully entered the cavities of two CDs.In addition,the antifungal effect test proved the better performance of DIF/HP-β-CD IC,which exceeded that of DIF emulsifiable concentrate.Therefore,our study provides informative direction for the intelligent use of liposoluble pesticides with cyclodextrins to develop water-based environmentally friendly formulations.
基金National Key Research and Development Program of China(Grant No.2022YFB4400200)the National Natural Science Foundation of China(Grant Nos.12104030,12104031,and 61627813)+10 种基金the Program of Introducing Talents of Discipline to Universities(Grant No.B16001)the Beijing Municipal Science and Technology Project(Grant No.Z201100004220002)China Postdoctoral Science Foundation(Grant No.2022M710320)China Scholarship Councilsupported by the ANR-15-CE24-0009 UMAMI and the ANR-20-CE09-0013by the Institute Carnot ICEEL for the project“Optic-switch”and Matelasby the Région Grand Estby the Metropole Grand Nancyby the impact project LUE-N4Spart of the French PIA project“Lorraine Universitéd’Excellence,”reference ANR-15-IDEX-04-LUEby the“FEDERFSE Lorraine et Massif Vosges 2014-2020,”a European Union Program。
文摘Antiferromagnets offer great potential for high-speed data processing applications,as they can expend spintronic devices from a static storage and gigahertz frequency range to the terahertz range.However,their zero net magnetization makes them difficult to manipulate and detect.In recent years,there has been a lot of attention given to the ultrafast manipulation of magnetic order using ultra-short single laser pulses,but it remains unknown whether a similar scenario can be observed in antiferromagnets.In this work,we demonstrate the manipulation of antiferromagnets with a single femtosecond laser pulse in perpendicular exchange-biased Co/Ir Mn/Co Gd trilayers.We study the dual exchange bias interlayer interaction in quasi-static conditions and competition in ultrafast antiferromagnet rearrangement.Our results show that,compared to conventional ferromagnetic/antiferromagnetic systems,the Ir Mn antiferromagnet can be ultrafast and efficiently manipulated by the coupled Co Gd ferrimagnetic layer,which paves the way for potential energy-efficient spintronic devices.
基金the National Natural Science Foundation of China(No.62063006)to the Natural Science Foundation of Guangxi Province(No.2023GXNS-FAA026025)+3 种基金to the Innovation Fund of Chinese Universities Industry-University-Research(ID:2021RYC06005)to the Research Project for Young and Middle-aged Teachers in Guangxi Universities(ID:2020KY15013)to the Special Research Project of Hechi University(ID:2021GCC028)supported by the Project of Outstanding Thousand Young Teachers’Training in Higher Education Institutions of Guangxi,Guangxi Colleges and Universities Key Laboratory of AI and Information Processing(Hechi University),Education Department of Guangxi Zhuang Autonomous Region.
文摘Visual simultaneous localization and mapping(SLAM)is crucial in robotics and autonomous driving.However,traditional visual SLAM faces challenges in dynamic environments.To address this issue,researchers have proposed semantic SLAM,which combines object detection,semantic segmentation,instance segmentation,and visual SLAM.Despite the growing body of literature on semantic SLAM,there is currently a lack of comprehensive research on the integration of object detection and visual SLAM.Therefore,this study aims to gather information from multiple databases and review relevant literature using specific keywords.It focuses on visual SLAM based on object detection,covering different aspects.Firstly,it discusses the current research status and challenges in this field,highlighting methods for incorporating semantic information from object detection networks into mileage measurement,closed-loop detection,and map construction.It also compares the characteristics and performance of various visual SLAM object detection algorithms.Lastly,it provides an outlook on future research directions and emerging trends in visual SLAM.Research has shown that visual SLAM based on object detection has significant improvements compared to traditional SLAM in dynamic point removal,data association,point cloud segmentation,and other technologies.It can improve the robustness and accuracy of the entire SLAM system and can run in real time.With the continuous optimization of algorithms and the improvement of hardware level,object visual SLAM has great potential for development.
基金This work was supported by the Key International(Regional)Joint Research Program of the National Natural Science Foundation of China(No.12120101002).
文摘In this paper,the path integral solutions for a general n-dimensional stochastic differential equa-tions(SDEs)withα-stable Lévy noise are derived and verified.Firstly,the governing equations for the solutions of n-dimensional SDEs under the excitation ofα-stable Lévy noise are obtained through the characteristic function of stochastic processes.Then,the short-time transition probability density func-tion of the path integral solution is derived based on the Chapman-Kolmogorov-Smoluchowski(CKS)equation and the characteristic function,and its correctness is demonstrated by proving that it satis-fies the governing equation of the solution of the SDE,which is also called the Fokker-Planck-Kolmogorov equation.Besides,illustrative examples are numerically considered for highlighting the feasibility of the proposed path integral method,and the pertinent Monte Carlo solution is also calculated to show its correctness and effectiveness.
基金the National Natural Science Foundation of China(No.62063006)to the Natural Science Foundation of Guangxi Province(No.2023GXNSFAA026025)+2 种基金to the Innovation Fund of Chinese Universities Industry-University-Research(ID:2021RYC06005)to the Research Project for Young and Middle-Aged Teachers in Guangxi Universities(ID:2020KY15013)to the Special Research Project of Hechi University(ID:2021GCC028).
文摘The robotics industry has seen rapid development in recent years due to the Corona Virus Disease 2019.With the development of sensors and smart devices,factories and enterprises have accumulated a large amount of data in their daily production,which creates extremely favorable conditions for robots to perform machine learning.However,in recent years,people’s awareness of data privacy has been increasing,leading to the inability to circulate data between different enterprises,resulting in the emergence of data silos.The emergence of federated learning provides a feasible solution to this problem,and the combination of federated learning and multi-robot systems can break down data silos and improve the overall performance of robots.However,as scholars have studied more deeply,they found that federated learning has very limited privacy protection.Therefore,how to protect data privacy from infringement remains an important issue.In this paper,we first give a brief introduction to the current development of multi-robot and federated learning;second,we review three aspects of privacy protection methods commonly used,privacy protection methods for multi-robot,and Other Problems Faced by Multi-robot Systems,focusing on method comparisons and challenges;and finally draw conclusions and predict possible future research directions.
基金Guangdong Basic and Applied Basic Research Foundation,Grant/Award Number:2019A1515110582Shenzhen Key Laboratory of Visual Object Detection and Recognition,Grant/Award Number:ZDSYS20190902093015527National Natural Science Foundation of China,Grant/Award Number:61876051。
文摘Deep learning has been widely used in the field of mammographic image classification owing to its superiority in automatic feature extraction.However,general deep learning models cannot achieve very satisfactory classification results on mammographic images because these models are not specifically designed for mammographic images and do not take the specific traits of these images into account.To exploit the essential discriminant information of mammographic images,we propose a novel classification method based on a convolutional neural network.Specifically,the proposed method designs two branches to extract the discriminative features from mammographic images from the mediolateral oblique and craniocaudal(CC)mammographic views.The features extracted from the two-view mammographic images contain complementary information that enables breast cancer to be more easily distinguished.Moreover,the attention block is introduced to capture the channel-wise information by adjusting the weight of each feature map,which is beneficial to emphasising the important features of mammographic images.Furthermore,we add a penalty term based on the fuzzy cluster algorithm to the cross-entropy function,which improves the generalisation ability of the classification model by maximising the interclass distance and minimising the intraclass distance of the samples.The experimental results on The Digital database for Screening Mammography INbreast and MIAS mammography databases illustrate that the proposed method achieves the best classification performance and is more robust than the compared state-ofthe-art classification methods.
基金supported by Shenzhen Science and Technology Innovation Commission(NO.KCXFZ20211020163402004).
文摘Compared with the space on the ground,if there is a fire in the urban complex underground space,the loss will be greatly harmful.In addition,the complex underground space is usually connected with other large space areas and densely populated.Once a fire occurs in the complex underground space,it will cause huge property losses and casualties.In order to reduce the risk of fire,it is necessary to deeply understand the development rules and characteristics of fire in the complex underground space of the city.This article has mainly carried on the following work:(I)A particularly complex model of the multi‐storey subway station was built.On this basis,three groups of comparative experiments were conducted to study the effects of fire power,fire location and smoke control system on fire development,and the conclusion that fire location is the most important factor for fire development was obtained;(II)In order to explore the entire space fire and the local space fire,CFD(Computational Fluid Dynamics)is used to build a large‐size fire model and a small‐size fire model respectively;(III)Multiple detector data as temperature slices were built,and it is expected to make full use of the simulation data to deduce the important index of fire location in the early stage of fire.All of the works in this paper will provide reference experimental data for the prevention and firefighting of a sudden fire in the complex underground space.
文摘Nonlinear and stochastic dynamics is essential and practical in almost all branches of natural science and engineering.It has been a central subject to understand various complex dynamics,such as random vibration,stochastic transition,synchronization,et al.,in the areas of mechanical and aerospace engineering,physics and chemistry.For example,the stochastic resonance has been utilized effectively in mechanical fault diagnosis and weak signal detection.Naturally,many methods have been developed to get the stochastic responses,including stochastic averaging method.
