Purpose:This paper reports on a scientometric analysis bolstered by human-in-the-loop,domain experts,to examine the field of metal-organic frameworks(MOFs)research.Scientometric analyses reveal the intellectual landsc...Purpose:This paper reports on a scientometric analysis bolstered by human-in-the-loop,domain experts,to examine the field of metal-organic frameworks(MOFs)research.Scientometric analyses reveal the intellectual landscape of a field.The study engaged MOF scientists in the design and review of our research workflow.MOF materials are an essential component in next-generation renewable energy storage and biomedical technologies.The research approach demonstrates how engaging experts,via human-in-the-loop processes,can help develop a comprehensive view of a field’s research trends,influential works,and specialized topics.Design/methodology/approach:Ascientometric analysis was conducted,integrating natural language processing(NLP),topic modeling,and network analysis methods.The analytical approach was enhanced through a human-in-the-loop iterative process involving MOF research scientists at selected intervals.MOF researcher feedback was incorporated into our method.The data sample included 65,209 MOF research articles.Python3 and software tool VOSviewer were used to perform the analysis.Findings:The findings demonstrate the value of including domain experts in research workflows,refinement,and interpretation of results.At each stage of the analysis,the MOF researchers contributed to interpreting the results and method refinements targeting our focus Research evolution of metal organic frameworks:A scientometric approach with human-in-the-loop on MOF research.This study identified influential works and their themes.Our findings also underscore four main MOF research directions and applications.Research limitations:This study is limited by the sample(articles identified and referenced by the Cambridge Structural Database)that informed our analysis.Practical implications:Our findings contribute to addressing the current gap in fully mapping out the comprehensive landscape of MOF research.Additionally,the results will help domain scientists target future research directions.Originality/value:To the best of our knowledge,the number of publications collected for analysis exceeds those of previous studies.This enabled us to explore a more extensive body of MOF research compared to previous studies.Another contribution of our work is the iterative engagement of domain scientists,who brought in-depth,expert interpretation to the data analysis,helping hone the study.展开更多
This paper considers the human-in-the-loop leader-following consensus control problem of multi-agent systems(MASs)with unknown matched nonlinear functions and actuator faults.It is assumed that a human operator contro...This paper considers the human-in-the-loop leader-following consensus control problem of multi-agent systems(MASs)with unknown matched nonlinear functions and actuator faults.It is assumed that a human operator controls the MASs via sending the command signal to a non-autonomous leader which generates the desired trajectory.Moreover,the leader’s input is nonzero and not available to all followers.By using neural networks and fault estimators to approximate unknown nonlinear dynamics and identify the actuator faults,respectively,the neighborhood observer-based neural fault-tolerant controller with dynamic coupling gains is designed.It is proved that the state of each follower can synchronize with the leader’s state under a directed graph and all signals in the closed-loop system are guaranteed to be cooperatively uniformly ultimately bounded.Finally,simulation results are presented for verifying the effectiveness of the proposed control method.展开更多
With the worldwide rapid development of 5 G networks, haptic communications, a key use case of the 5 G, has attracted increasing attentions nowadays. Its human-in-the-loop nature makes quality of experience(QoE) the l...With the worldwide rapid development of 5 G networks, haptic communications, a key use case of the 5 G, has attracted increasing attentions nowadays. Its human-in-the-loop nature makes quality of experience(QoE) the leading performance indicator of the system design. A vast number of high quality works were published on user-level, application-level and network-level QoE-oriented designs in haptic communications. In this paper, we present an overview of the recent research activities in this progressive research area. We start from the QoE modeling of human haptic perceptions, followed by the application-level QoE management mechanisms based on these QoE models. High fidelity haptic communications require an orchestra of QoE designs in the application level and the quality of service(QoS) support in the network level. Hence, we also review the state-of-the-art QoS-related QoE management strategies in haptic communications, especially the QoS-related QoE modeling which guides the resource allocation design of the communication network. In addition to a thorough survey of the literature, we also present the open challenges in this research area. We believe that our review and findings in this paper not only provide a timely summary of prevailing research in this area, but also help to inspire new QoE-related research opportunities in haptic communications.展开更多
Neural networks have demonstrated exceptional performance across a range of applications.Yet,their training often demands substantial time and data resources,presenting a challenge for autonomous robots operating in r...Neural networks have demonstrated exceptional performance across a range of applications.Yet,their training often demands substantial time and data resources,presenting a challenge for autonomous robots operating in real-world environments where real-time learning is difficult.To mitigate this constraint,we propose a novel human-in-the-loop framework that harnesses human expertise to mitigate the learning challenges of autonomous robots.Our approach centers on directly incorporating human knowledge and insights into the robot’s learning pipeline.The proposed framework incorporates a mechanism for autonomous learning from the environment via reinforcement learning,utilizing a pre-trained model that encapsulates human knowledge as its foundation.By integrating human-provided knowledge and evaluation,we aim to bridge the division between human intuition and machine learning capabilities.Through a series of collision avoidance experiments,we validated that incorporating human knowledge significantly improves both learning efficiency and generalization capabilities.This collaborative learning paradigm enables robots to utilize human common sense and domain-specific expertise,resulting in faster convergence and better performance in complex environments.This research contributes to the development of more efficient and adaptable autonomous robots and seeks to analyze how humans can effectively participate in robot learning and the effects of such participation,illuminating the intricate interplay between human cognition and artificial intelligence.展开更多
The exponential growth of astronomical datasets provides an unprecedented opportunity for humans to gain insight into the Universe.However,effectively analyzing this vast amount of data poses a significant challenge.I...The exponential growth of astronomical datasets provides an unprecedented opportunity for humans to gain insight into the Universe.However,effectively analyzing this vast amount of data poses a significant challenge.In response,astronomers are turning to deep learning techniques,but these methods are limited by their specific training sets,leading to considerable duplicate workloads.To overcome this issue,we built a framework for the general analysis of galaxy images based on a large vision model(LVM)plus downstream tasks(DST),including galaxy morphological classification,image restoration object detection,parameter extraction,and more.Considering the low signal-to-noise ratios of galaxy images and the imbalanced distribution of galaxy categories,we designed our LVM to incorporate a Human-in-the-loop(HITL)module,which leverages human knowledge to enhance the reliability and interpretability of processing galaxy images interactively.The proposed framework exhibits notable fewshot learning capabilities and versatile adaptability for all the abovementioned tasks on galaxy images in the DESI Legacy Imaging Surveys.In particular,for the object detection task,which was trained using 1000 data points,our DST in the LVM achieved an accuracy of 96.7%,while ResNet50 plus Mask R-CNN reached an accuracy of 93.1%.For morphological classification,to obtain an area under the curve(AUC)of~0.9,LVM plus DST and HITL only requested 1/50 of the training sets that ResNet18 requested.In addition,multimodal data can be integrated,which creates possibilities for conducting joint analyses with datasets spanning diverse domains in the era of multi-messenger astronomy.展开更多
Automated pulmonary nodule detection using computed tomography scans is vital in the early diagnosis of lung cancer.Although extensive well-performed methods have been proposed for this task,they suffer from the domai...Automated pulmonary nodule detection using computed tomography scans is vital in the early diagnosis of lung cancer.Although extensive well-performed methods have been proposed for this task,they suffer from the domain shift issue between training and test images.Unsupervised domain adaptation(UDA)methods provide a promising means to mitigate the domain variance;however,their performance is still limited since no target domain supervision is introduced.To make the pulmonary nodule detection algorithm more applicable in clinical practice and further boost the performance across domains,we propose a human-in-the-loop method in a semi-supervised fashion to enhance the model generalization ability when transferred from source domain to target domain.Specifically,we first train a detector model on source domain,and then the pre-trained detector is utilized with our proposed uncertainty-guided sample selection scheme(USSS)to find a few target domain samples worth annotating most and obtain their human annotations.Finally,the annotated and the rest unlabeled target domain samples are used together to refine the pre-trained model via our proposed zoom-in and zoom-out constraint(ZZC)strategy.We evaluate our method on the Nodule Analysis 2016(LUNA16)and TianChi datasets.Experimental results show that our method surpasses recent competitive methods on source domain and also achieves surprising performance on target domain.展开更多
In this paper, the problem of pre-specified performance fault-tolerant cluster consensus control and fault direction identification is solved for the human-in-the-loop(HIL) swarm unmanned aerial vehicles(UAVs) in the ...In this paper, the problem of pre-specified performance fault-tolerant cluster consensus control and fault direction identification is solved for the human-in-the-loop(HIL) swarm unmanned aerial vehicles(UAVs) in the presence of possible nonidentical and unknown direction faults(NUDFs) in the yaw channel.The control strategy begins with the design of a pre-specified performance event-triggered observer for each individual UAV.These observers estimate the outputs of the human controlled UAVs, and simultaneously achieve the distributed design of actual control signals as well as cluster consensus of the observer output.It is worth mentioning that these observers require neither the high-order derivatives of the human controlled UAVs' output nor a priori knowledge of the initial conditions. The fault-tolerant controller realizes the pre-specified performance output regulation through error transformation and the Nussbaum function. It should be pointed out that there are no chattering caused by the jump of the Nussbaum function when a reverse fault occurs. In addition, to provide a basis for further solving the problem of physical malfunctions, a fault direction identification algorithm is proposed to accurately identify whether a reverse fault has occurred. Simulation results verify the effectiveness of the proposed control and fault direction identification strategies when the reverse faults occur.展开更多
The dynamic event-triggered(DET)formation control problem of a class of stochastic nonlinear multi-agent systems(MASs)with full state constraints is investigated in this article.Supposing that the human operator sends...The dynamic event-triggered(DET)formation control problem of a class of stochastic nonlinear multi-agent systems(MASs)with full state constraints is investigated in this article.Supposing that the human operator sends commands to the leader as control input signals,all followers keep formation through network topology communication.Under the command-filter-based backstepping technique,the radial basis function neural networks(RBF NNs)and the barrier Lyapunov function(BLF)are utilized to resolve the problems of unknown nonlinear terms and full state constraints,respectively.Furthermore,a DET control mechanism is proposed to reduce the occupation of communication bandwidth.The presented distributed formation control strategy guarantees that all signals of the MASs are semi-globally uniformly ultimately bounded(SGUUB)in probability.Finally,the feasibility of the theoretical research result is demonstrated by a simulation example.展开更多
This paper proposes the concept and framework of smart operating system based on the artificial intelligence(AI)techniques. The demands and the potential applications of AI technologies in power system control centers...This paper proposes the concept and framework of smart operating system based on the artificial intelligence(AI)techniques. The demands and the potential applications of AI technologies in power system control centers is discussed in the beginning of the paper. The discussion is based on the results of a field study in the Tianjin Power System Control Center in China. According to the study, one problem in power systems is that the power system analysis system in the control center is not fast and powerful enough to help the operators in time to deal with the incidents in the power system. Another issue in current power system control center is that the operation tickets are compiled manually by the operators, so that it is less efficient and human errors cannot be avoided. Based on these problems, a framework of the smart operating robot is proposed in this paper, which includes an intelligent power system analysis system and a smart operation ticket compiling system to solve the two problems in power system control centers. The proposed framework is mainly based on the AI techniques, especially the neural network with deep learning, since it is faster and more capable of dealing with the highly nonlinear and complex power system.展开更多
Mapping grasps from human to anthropomorphic robotic hands is an open issue in research,because the master hand and the slave hand have dissimilar kinematics.This paper proposes a hybrid mapping method to solve this p...Mapping grasps from human to anthropomorphic robotic hands is an open issue in research,because the master hand and the slave hand have dissimilar kinematics.This paper proposes a hybrid mapping method to solve this problem.In the proposed method,fingers in the master and the slave hands are divided into vital and synergic fingers according to their contribution to the grasping task.The tip of the vital finger of the master hand is first mapped to that of the slave hand while ensuring that both are in simultaneous contact with the object to be grasped.Following postural synergy theory,joints of the other synergic fingers of the slave hand are then used to generate an anthropomorphic grasping configuration according to the shape of the object to be grasped.Following this,a human-guided impedance controller is used to reduce the pre-grasping error and realize compliant interaction with the environment.The proposed hybrid mapping method can not only generate the posture of the humanoid envelope but can also carry out impedance-adaptive matching.It was evaluated using simulations and an experiment involving an anthropomorphic robotic slave hand.展开更多
Rapid advancement of digital technologies has resulted in an acceleration of cyber-physical systems for autonomous mobile robots to improve energy asset management activities within inspection,maintenance and repair.W...Rapid advancement of digital technologies has resulted in an acceleration of cyber-physical systems for autonomous mobile robots to improve energy asset management activities within inspection,maintenance and repair.Within this systems-based approach,the role of the human-in-the-loop has also increased leading to cyber-physical-human systems requiring real-time interaction of robotics and digital twins with a human operator.Subject to existing network systems and physical systems,cyber-physical-human systems face enormous challenges requiring further investigation.This review presents the state-of-the-art in discovery,design,development and deployment of cyber-physical-human systems for mobile robots in energy asset management.To address dominant concepts and misconceptions in this area,key terminologies,system concepts and applications are presented.Then a state-of-the-art review with associated trends for several applications within academic and industrial sectors is presented where current practises and limitations are then discussed.Finally,future opportunities are explored alongside highlighted concepts providing a pathway for rapid adoption and improved key performance indicators of mobile fleets for facility operators and those in the wider community.展开更多
Despite significant progress in autonomous vehicles(AVs),the development of driving policies that ensure both the safety of AVs and traffic flow efficiency has not yet been fully explored.In this paper,we propose an e...Despite significant progress in autonomous vehicles(AVs),the development of driving policies that ensure both the safety of AVs and traffic flow efficiency has not yet been fully explored.In this paper,we propose an enhanced human-in-the-loop reinforcement learning method,termed the Human as AI mentor-based deep reinforcement learning(HAIM-DRL)framework,which facilitates safe and efficient autonomous driving in mixed traffic platoon.Drawing inspiration from the human learning process,we first introduce an innovative learning paradigm that effectively injects human intelligence into AI,termed Human as AI mentor(HAIM).In this paradigm,the human expert serves as a mentor to the AI agent.While allowing the agent to sufficiently explore uncertain environments,the human expert can take control in dangerous situations and demonstrate correct actions to avoid potential accidents.On the other hand,the agent could be guided to minimize traffic flow disturbance,thereby optimizing traffic flow efficiency.In detail,HAIM-DRL leverages data collected from free exploration and partial human demonstrations as its two training sources.Remarkably,we circumvent the intricate process of manually designing reward functions;instead,we directly derive proxy state-action values from partial human demonstrations to guide the agents’policy learning.Additionally,we employ a minimal intervention technique to reduce the human mentor’s cognitive load.Comparative results show that HAIM-DRL outperforms traditional methods in driving safety,sampling efficiency,mitigation of traffic flow disturbance,and generalizability to unseen traffic scenarios.展开更多
Human-in-the-loop(HiTL)control is promising for the cooperative control problem of multi-agent systems(MASs)under the complicated environment.By considering the effect of human intelligence and decision making,the sys...Human-in-the-loop(HiTL)control is promising for the cooperative control problem of multi-agent systems(MASs)under the complicated environment.By considering the effect of human intelligence and decision making,the system robustness and security are notably enhanced.Hence,a distributed fixed-time tracking control problem is investigated in this paper for heterogeneous MASs based on the HiTL idea.First,a lemma of practically fixed-time stable is given where an explicit relationship of settling time and convergence domain is clearly shown.Then,under the framework of the adaptive backstepping approach,a series of modified intermediate control signals is designed to avoid the singularity problem by taking advantage of power transformation,fuzzy logic systems,and inequality schemes.Finally,the numerical example and comparison results are utilized to testify the effectiveness of the proposed method.展开更多
We present a set of configurable Web service and interactive tools,s-ProvFlow,for managing and exploiting records tracking data lineage during workflow runs.It facilitates detailed analysis of single executions.It hel...We present a set of configurable Web service and interactive tools,s-ProvFlow,for managing and exploiting records tracking data lineage during workflow runs.It facilitates detailed analysis of single executions.It helps users manage complex tasks by exposing the relationships between data,people,equipment and workflow runs intended to combine productively.Its logical model extends the PROV standard to precisely record parallel data-streaming applications.Its metadata handling encourages users to capture the application context by specifying how application attributes,often using standard vocabularies,should be added.These metadata records immediately help productivity as the interactive tools support their use in selection and bulk operations.Users rapidly appreciate the power of the encoded semantics as they reap the benefits.This improves the quality of provenance for users and management.Which in turn facilitates analysis of collections of runs,enabling users to manage results and validate procedures.It fosters reuse of data and methods and facilitates diagnostic investigations and optimisations.We present S-ProvFlow's use by scientists,research engineers and managers as part of the DARE hyper-platform as they create,validate and use their data-driven scientific workflows.展开更多
The philosophy of building energy management is going through a paradigm change from traditional,often inefficient,user-controlled systems to one that is centrally automated with the aid of IoT-enabled technologies.In...The philosophy of building energy management is going through a paradigm change from traditional,often inefficient,user-controlled systems to one that is centrally automated with the aid of IoT-enabled technologies.In this context,occupants’perceived control and building automation may seem to be in conflict.The inquiry of this study is rooted in a proposition that while building automation and centralized control systems are assumed to provide indoor comfort and conserve energy use,limiting occupants’control over their work environment may result in dissatisfaction,and in turn decrease productivity.For assessing this hypothesis,data from the post-occupancy evaluation survey of a smart building in a university in Australia was used to analyze the relationships between perceived control,satisfaction,and perceived productivity.Using structural equation modeling,we have found a positive direct effect of occupants’perceived control on overall satisfaction with their working area.Meanwhile,perceived control exerts an influence on perceived productivity through satisfaction.Furthermore,a field experiment conducted in the same building revealed the potential impact that occupant controllability can have on energy saving.We changed the default light settings from automatic on-and-offto manual-on and automatic-off,letting occupants choose themselves whether to switch the light on or not.Interestingly,about half of the participants usually kept the lights off,preferring daylight in their rooms.This also resulted in a reduction in lighting electricity use by 17.8%without any upfront investment and major technical modification.These findings emphasize the important role of perceived control on occupant satisfaction and productivity,as well as on the energy-saving potential of the user-in-the-loop automation of buildings.展开更多
文摘Purpose:This paper reports on a scientometric analysis bolstered by human-in-the-loop,domain experts,to examine the field of metal-organic frameworks(MOFs)research.Scientometric analyses reveal the intellectual landscape of a field.The study engaged MOF scientists in the design and review of our research workflow.MOF materials are an essential component in next-generation renewable energy storage and biomedical technologies.The research approach demonstrates how engaging experts,via human-in-the-loop processes,can help develop a comprehensive view of a field’s research trends,influential works,and specialized topics.Design/methodology/approach:Ascientometric analysis was conducted,integrating natural language processing(NLP),topic modeling,and network analysis methods.The analytical approach was enhanced through a human-in-the-loop iterative process involving MOF research scientists at selected intervals.MOF researcher feedback was incorporated into our method.The data sample included 65,209 MOF research articles.Python3 and software tool VOSviewer were used to perform the analysis.Findings:The findings demonstrate the value of including domain experts in research workflows,refinement,and interpretation of results.At each stage of the analysis,the MOF researchers contributed to interpreting the results and method refinements targeting our focus Research evolution of metal organic frameworks:A scientometric approach with human-in-the-loop on MOF research.This study identified influential works and their themes.Our findings also underscore four main MOF research directions and applications.Research limitations:This study is limited by the sample(articles identified and referenced by the Cambridge Structural Database)that informed our analysis.Practical implications:Our findings contribute to addressing the current gap in fully mapping out the comprehensive landscape of MOF research.Additionally,the results will help domain scientists target future research directions.Originality/value:To the best of our knowledge,the number of publications collected for analysis exceeds those of previous studies.This enabled us to explore a more extensive body of MOF research compared to previous studies.Another contribution of our work is the iterative engagement of domain scientists,who brought in-depth,expert interpretation to the data analysis,helping hone the study.
基金This work was partially supported by the National Natural Science Foundation of China(62033003,62003098)the Local Innovative and Research Teams Project of Guangdong Special Support Program(2019BT02X353)the China Postdoctoral Science Foundation(2019M662813,2020T130124).
文摘This paper considers the human-in-the-loop leader-following consensus control problem of multi-agent systems(MASs)with unknown matched nonlinear functions and actuator faults.It is assumed that a human operator controls the MASs via sending the command signal to a non-autonomous leader which generates the desired trajectory.Moreover,the leader’s input is nonzero and not available to all followers.By using neural networks and fault estimators to approximate unknown nonlinear dynamics and identify the actuator faults,respectively,the neighborhood observer-based neural fault-tolerant controller with dynamic coupling gains is designed.It is proved that the state of each follower can synchronize with the leader’s state under a directed graph and all signals in the closed-loop system are guaranteed to be cooperatively uniformly ultimately bounded.Finally,simulation results are presented for verifying the effectiveness of the proposed control method.
文摘With the worldwide rapid development of 5 G networks, haptic communications, a key use case of the 5 G, has attracted increasing attentions nowadays. Its human-in-the-loop nature makes quality of experience(QoE) the leading performance indicator of the system design. A vast number of high quality works were published on user-level, application-level and network-level QoE-oriented designs in haptic communications. In this paper, we present an overview of the recent research activities in this progressive research area. We start from the QoE modeling of human haptic perceptions, followed by the application-level QoE management mechanisms based on these QoE models. High fidelity haptic communications require an orchestra of QoE designs in the application level and the quality of service(QoS) support in the network level. Hence, we also review the state-of-the-art QoS-related QoE management strategies in haptic communications, especially the QoS-related QoE modeling which guides the resource allocation design of the communication network. In addition to a thorough survey of the literature, we also present the open challenges in this research area. We believe that our review and findings in this paper not only provide a timely summary of prevailing research in this area, but also help to inspire new QoE-related research opportunities in haptic communications.
基金supported by the research funding of Waseda University,Japan.
文摘Neural networks have demonstrated exceptional performance across a range of applications.Yet,their training often demands substantial time and data resources,presenting a challenge for autonomous robots operating in real-world environments where real-time learning is difficult.To mitigate this constraint,we propose a novel human-in-the-loop framework that harnesses human expertise to mitigate the learning challenges of autonomous robots.Our approach centers on directly incorporating human knowledge and insights into the robot’s learning pipeline.The proposed framework incorporates a mechanism for autonomous learning from the environment via reinforcement learning,utilizing a pre-trained model that encapsulates human knowledge as its foundation.By integrating human-provided knowledge and evaluation,we aim to bridge the division between human intuition and machine learning capabilities.Through a series of collision avoidance experiments,we validated that incorporating human knowledge significantly improves both learning efficiency and generalization capabilities.This collaborative learning paradigm enables robots to utilize human common sense and domain-specific expertise,resulting in faster convergence and better performance in complex environments.This research contributes to the development of more efficient and adaptable autonomous robots and seeks to analyze how humans can effectively participate in robot learning and the effects of such participation,illuminating the intricate interplay between human cognition and artificial intelligence.
基金the support from the National Natural Science Foundation of China(Grant Nos.12173027,12303105,12173062)the National Key R&D Program of China(Grant Nos.2023YFF0725300,2022YFF0503402)+5 种基金the Science Research Grants from the Square Kilometre Array(SKA)(2020SKA0110100)the Science Research Grants from the China Manned Space Project(Grant Nos.CMS-CSST-2021-A01,CMS-CSST-2021-A07,CMS-CSST-2021-B05)the CAS Project for Young Scientists in Basic ResearchChina(Grant No.YSBR-062)supported by the Young Data Scientist Project of the National Astronomical Data Centerthe Program of Science and Education Integration at the School of Astronomy and Space Science,University of Chinese Academy of Sciences,China。
文摘The exponential growth of astronomical datasets provides an unprecedented opportunity for humans to gain insight into the Universe.However,effectively analyzing this vast amount of data poses a significant challenge.In response,astronomers are turning to deep learning techniques,but these methods are limited by their specific training sets,leading to considerable duplicate workloads.To overcome this issue,we built a framework for the general analysis of galaxy images based on a large vision model(LVM)plus downstream tasks(DST),including galaxy morphological classification,image restoration object detection,parameter extraction,and more.Considering the low signal-to-noise ratios of galaxy images and the imbalanced distribution of galaxy categories,we designed our LVM to incorporate a Human-in-the-loop(HITL)module,which leverages human knowledge to enhance the reliability and interpretability of processing galaxy images interactively.The proposed framework exhibits notable fewshot learning capabilities and versatile adaptability for all the abovementioned tasks on galaxy images in the DESI Legacy Imaging Surveys.In particular,for the object detection task,which was trained using 1000 data points,our DST in the LVM achieved an accuracy of 96.7%,while ResNet50 plus Mask R-CNN reached an accuracy of 93.1%.For morphological classification,to obtain an area under the curve(AUC)of~0.9,LVM plus DST and HITL only requested 1/50 of the training sets that ResNet18 requested.In addition,multimodal data can be integrated,which creates possibilities for conducting joint analyses with datasets spanning diverse domains in the era of multi-messenger astronomy.
基金supported in part by the National Natural Science Foundation of China(No.62171377)in part by the National Key R&D Program of China(Nos.2022YFC2009903 and 2022YFC2009900)in part by the Ningbo Clinical Research Center for Medical Imaging(No.2021L003)(Open Project 2022LYKFZD06).
文摘Automated pulmonary nodule detection using computed tomography scans is vital in the early diagnosis of lung cancer.Although extensive well-performed methods have been proposed for this task,they suffer from the domain shift issue between training and test images.Unsupervised domain adaptation(UDA)methods provide a promising means to mitigate the domain variance;however,their performance is still limited since no target domain supervision is introduced.To make the pulmonary nodule detection algorithm more applicable in clinical practice and further boost the performance across domains,we propose a human-in-the-loop method in a semi-supervised fashion to enhance the model generalization ability when transferred from source domain to target domain.Specifically,we first train a detector model on source domain,and then the pre-trained detector is utilized with our proposed uncertainty-guided sample selection scheme(USSS)to find a few target domain samples worth annotating most and obtain their human annotations.Finally,the annotated and the rest unlabeled target domain samples are used together to refine the pre-trained model via our proposed zoom-in and zoom-out constraint(ZZC)strategy.We evaluate our method on the Nodule Analysis 2016(LUNA16)and TianChi datasets.Experimental results show that our method surpasses recent competitive methods on source domain and also achieves surprising performance on target domain.
基金supported in part by the National Natural Science Foundation of China(62173028,62233015,62173024)the Guangdong Basic and Applied Basic Research Foundation(2024A1515011493)+3 种基金the Science,Technology&Innovation Project of Xiongan New Area(2023XAGG0062)Beijing Natural Science Foundation(4232060)the International Scientists Project,Beijing Natural Science Foundation(IS23065)the Brazilian Research Council(303289/2022-8)
文摘In this paper, the problem of pre-specified performance fault-tolerant cluster consensus control and fault direction identification is solved for the human-in-the-loop(HIL) swarm unmanned aerial vehicles(UAVs) in the presence of possible nonidentical and unknown direction faults(NUDFs) in the yaw channel.The control strategy begins with the design of a pre-specified performance event-triggered observer for each individual UAV.These observers estimate the outputs of the human controlled UAVs, and simultaneously achieve the distributed design of actual control signals as well as cluster consensus of the observer output.It is worth mentioning that these observers require neither the high-order derivatives of the human controlled UAVs' output nor a priori knowledge of the initial conditions. The fault-tolerant controller realizes the pre-specified performance output regulation through error transformation and the Nussbaum function. It should be pointed out that there are no chattering caused by the jump of the Nussbaum function when a reverse fault occurs. In addition, to provide a basis for further solving the problem of physical malfunctions, a fault direction identification algorithm is proposed to accurately identify whether a reverse fault has occurred. Simulation results verify the effectiveness of the proposed control and fault direction identification strategies when the reverse faults occur.
基金supported in part by the National Natural Science Foundation of China(62121004,62033003,61973091,62203119)the Local Innovative and Research Teams Project of Guangdong Special Support Program(2019BT02X353)+1 种基金the Natural Science Foundation of Guangdong Province(2023A1515011527,2022A1515011506)the China National Postdoctoral Program(BX20220095,2022M710826).
文摘The dynamic event-triggered(DET)formation control problem of a class of stochastic nonlinear multi-agent systems(MASs)with full state constraints is investigated in this article.Supposing that the human operator sends commands to the leader as control input signals,all followers keep formation through network topology communication.Under the command-filter-based backstepping technique,the radial basis function neural networks(RBF NNs)and the barrier Lyapunov function(BLF)are utilized to resolve the problems of unknown nonlinear terms and full state constraints,respectively.Furthermore,a DET control mechanism is proposed to reduce the occupation of communication bandwidth.The presented distributed formation control strategy guarantees that all signals of the MASs are semi-globally uniformly ultimately bounded(SGUUB)in probability.Finally,the feasibility of the theoretical research result is demonstrated by a simulation example.
基金supported by State Grid Corporation of China(SGCC)Science and Technolgy Project(SGTJDK00DWJS1700060)
文摘This paper proposes the concept and framework of smart operating system based on the artificial intelligence(AI)techniques. The demands and the potential applications of AI technologies in power system control centers is discussed in the beginning of the paper. The discussion is based on the results of a field study in the Tianjin Power System Control Center in China. According to the study, one problem in power systems is that the power system analysis system in the control center is not fast and powerful enough to help the operators in time to deal with the incidents in the power system. Another issue in current power system control center is that the operation tickets are compiled manually by the operators, so that it is less efficient and human errors cannot be avoided. Based on these problems, a framework of the smart operating robot is proposed in this paper, which includes an intelligent power system analysis system and a smart operation ticket compiling system to solve the two problems in power system control centers. The proposed framework is mainly based on the AI techniques, especially the neural network with deep learning, since it is faster and more capable of dealing with the highly nonlinear and complex power system.
基金supported in part by the China National Key Research and Development Program under Grant no.2020YFC2007801in part by the National Natural Science Foundation of China under Grant no.U1813209.
文摘Mapping grasps from human to anthropomorphic robotic hands is an open issue in research,because the master hand and the slave hand have dissimilar kinematics.This paper proposes a hybrid mapping method to solve this problem.In the proposed method,fingers in the master and the slave hands are divided into vital and synergic fingers according to their contribution to the grasping task.The tip of the vital finger of the master hand is first mapped to that of the slave hand while ensuring that both are in simultaneous contact with the object to be grasped.Following postural synergy theory,joints of the other synergic fingers of the slave hand are then used to generate an anthropomorphic grasping configuration according to the shape of the object to be grasped.Following this,a human-guided impedance controller is used to reduce the pre-grasping error and realize compliant interaction with the environment.The proposed hybrid mapping method can not only generate the posture of the humanoid envelope but can also carry out impedance-adaptive matching.It was evaluated using simulations and an experiment involving an anthropomorphic robotic slave hand.
基金supported by the following UK Research and In-novation(UKRI)projects:Robotics and Artificial Intelligence for Nu-clear(RAIN)EP/R026084/1,Robotics for Nuclear Environments(RNE)EP/P01366X/1UK Robotics and Artificial Intelligence Hub for Offshore Energy Asset Integrity Management EP/R026173/1.
文摘Rapid advancement of digital technologies has resulted in an acceleration of cyber-physical systems for autonomous mobile robots to improve energy asset management activities within inspection,maintenance and repair.Within this systems-based approach,the role of the human-in-the-loop has also increased leading to cyber-physical-human systems requiring real-time interaction of robotics and digital twins with a human operator.Subject to existing network systems and physical systems,cyber-physical-human systems face enormous challenges requiring further investigation.This review presents the state-of-the-art in discovery,design,development and deployment of cyber-physical-human systems for mobile robots in energy asset management.To address dominant concepts and misconceptions in this area,key terminologies,system concepts and applications are presented.Then a state-of-the-art review with associated trends for several applications within academic and industrial sectors is presented where current practises and limitations are then discussed.Finally,future opportunities are explored alongside highlighted concepts providing a pathway for rapid adoption and improved key performance indicators of mobile fleets for facility operators and those in the wider community.
文摘Despite significant progress in autonomous vehicles(AVs),the development of driving policies that ensure both the safety of AVs and traffic flow efficiency has not yet been fully explored.In this paper,we propose an enhanced human-in-the-loop reinforcement learning method,termed the Human as AI mentor-based deep reinforcement learning(HAIM-DRL)framework,which facilitates safe and efficient autonomous driving in mixed traffic platoon.Drawing inspiration from the human learning process,we first introduce an innovative learning paradigm that effectively injects human intelligence into AI,termed Human as AI mentor(HAIM).In this paradigm,the human expert serves as a mentor to the AI agent.While allowing the agent to sufficiently explore uncertain environments,the human expert can take control in dangerous situations and demonstrate correct actions to avoid potential accidents.On the other hand,the agent could be guided to minimize traffic flow disturbance,thereby optimizing traffic flow efficiency.In detail,HAIM-DRL leverages data collected from free exploration and partial human demonstrations as its two training sources.Remarkably,we circumvent the intricate process of manually designing reward functions;instead,we directly derive proxy state-action values from partial human demonstrations to guide the agents’policy learning.Additionally,we employ a minimal intervention technique to reduce the human mentor’s cognitive load.Comparative results show that HAIM-DRL outperforms traditional methods in driving safety,sampling efficiency,mitigation of traffic flow disturbance,and generalizability to unseen traffic scenarios.
基金the National Natural Science Foundation of China(Grant Nos.62373208,62003097,62033003,61873139,62103214 and 62203245)the Talent Introduction and Cultivation Plan for Youth Innovation of Universities in Shandong Province。
文摘Human-in-the-loop(HiTL)control is promising for the cooperative control problem of multi-agent systems(MASs)under the complicated environment.By considering the effect of human intelligence and decision making,the system robustness and security are notably enhanced.Hence,a distributed fixed-time tracking control problem is investigated in this paper for heterogeneous MASs based on the HiTL idea.First,a lemma of practically fixed-time stable is given where an explicit relationship of settling time and convergence domain is clearly shown.Then,under the framework of the adaptive backstepping approach,a series of modified intermediate control signals is designed to avoid the singularity problem by taking advantage of power transformation,fuzzy logic systems,and inequality schemes.Finally,the numerical example and comparison results are utilized to testify the effectiveness of the proposed method.
基金supported by the EU FP7-Infrastructure project VERCE(No.283543)EU H2020 project DARE(No.777413).
文摘We present a set of configurable Web service and interactive tools,s-ProvFlow,for managing and exploiting records tracking data lineage during workflow runs.It facilitates detailed analysis of single executions.It helps users manage complex tasks by exposing the relationships between data,people,equipment and workflow runs intended to combine productively.Its logical model extends the PROV standard to precisely record parallel data-streaming applications.Its metadata handling encourages users to capture the application context by specifying how application attributes,often using standard vocabularies,should be added.These metadata records immediately help productivity as the interactive tools support their use in selection and bulk operations.Users rapidly appreciate the power of the encoded semantics as they reap the benefits.This improves the quality of provenance for users and management.Which in turn facilitates analysis of collections of runs,enabling users to manage results and validate procedures.It fosters reuse of data and methods and facilitates diagnostic investigations and optimisations.We present S-ProvFlow's use by scientists,research engineers and managers as part of the DARE hyper-platform as they create,validate and use their data-driven scientific workflows.
文摘The philosophy of building energy management is going through a paradigm change from traditional,often inefficient,user-controlled systems to one that is centrally automated with the aid of IoT-enabled technologies.In this context,occupants’perceived control and building automation may seem to be in conflict.The inquiry of this study is rooted in a proposition that while building automation and centralized control systems are assumed to provide indoor comfort and conserve energy use,limiting occupants’control over their work environment may result in dissatisfaction,and in turn decrease productivity.For assessing this hypothesis,data from the post-occupancy evaluation survey of a smart building in a university in Australia was used to analyze the relationships between perceived control,satisfaction,and perceived productivity.Using structural equation modeling,we have found a positive direct effect of occupants’perceived control on overall satisfaction with their working area.Meanwhile,perceived control exerts an influence on perceived productivity through satisfaction.Furthermore,a field experiment conducted in the same building revealed the potential impact that occupant controllability can have on energy saving.We changed the default light settings from automatic on-and-offto manual-on and automatic-off,letting occupants choose themselves whether to switch the light on or not.Interestingly,about half of the participants usually kept the lights off,preferring daylight in their rooms.This also resulted in a reduction in lighting electricity use by 17.8%without any upfront investment and major technical modification.These findings emphasize the important role of perceived control on occupant satisfaction and productivity,as well as on the energy-saving potential of the user-in-the-loop automation of buildings.
