Missile interception problem can be regarded as a two-person zero-sum differential games problem,which depends on the solution of Hamilton-Jacobi-Isaacs(HJI)equa-tion.It has been proved impossible to obtain a closed-f...Missile interception problem can be regarded as a two-person zero-sum differential games problem,which depends on the solution of Hamilton-Jacobi-Isaacs(HJI)equa-tion.It has been proved impossible to obtain a closed-form solu-tion due to the nonlinearity of HJI equation,and many iterative algorithms are proposed to solve the HJI equation.Simultane-ous policy updating algorithm(SPUA)is an effective algorithm for solving HJI equation,but it is an on-policy integral reinforce-ment learning(IRL).For online implementation of SPUA,the dis-turbance signals need to be adjustable,which is unrealistic.In this paper,an off-policy IRL algorithm based on SPUA is pro-posed without making use of any knowledge of the systems dynamics.Then,a neural-network based online adaptive critic implementation scheme of the off-policy IRL algorithm is pre-sented.Based on the online off-policy IRL method,a computa-tional intelligence interception guidance(CIIG)law is developed for intercepting high-maneuvering target.As a model-free method,intercepting targets can be achieved through measur-ing system data online.The effectiveness of the CIIG is verified through two missile and target engagement scenarios.展开更多
)The 2008 IEEE World Congress on Computational Intelligence (WCCI 2008) will be held at the HongKong Convention and Exhibition Centre during June 1-6, 2008. WCCI 2008 will be the fifth milestone inthis series with a g...)The 2008 IEEE World Congress on Computational Intelligence (WCCI 2008) will be held at the HongKong Convention and Exhibition Centre during June 1-6, 2008. WCCI 2008 will be the fifth milestone inthis series with a glorious history from WCCI 1994 in Orlando, WCCI 1998 in Anchorage, WCCI 2002in Honolulu, to WCCI 2006 in Vancouver. Sponsored by the IEEE Computational Intelligence Society,展开更多
In order to improve the performance of multi-objective workflow scheduling in cloud system, a multi-swarm multiobjective optimization algorithm(MSMOOA) is proposed to satisfy multiple conflicting objectives. Inspired ...In order to improve the performance of multi-objective workflow scheduling in cloud system, a multi-swarm multiobjective optimization algorithm(MSMOOA) is proposed to satisfy multiple conflicting objectives. Inspired by division of the same species into multiple swarms for different objectives and information sharing among these swarms in nature, each physical machine in the data center is considered a swarm and employs improved multi-objective particle swarm optimization to find out non-dominated solutions with one objective in MSMOOA. The particles in each swarm are divided into two classes and adopt different strategies to evolve cooperatively. One class of particles can communicate with several swarms simultaneously to promote the information sharing among swarms and the other class of particles can only exchange information with the particles located in the same swarm. Furthermore, in order to avoid the influence by the elastic available resources, a manager server is adopted in the cloud data center to collect the available resources for scheduling. The quality of the proposed method with other related approaches is evaluated by using hybrid and parallel workflow applications. The experiment results highlight the better performance of the MSMOOA than that of compared algorithms.展开更多
Objective To observe the value of self-supervised deep learning artificial intelligence(AI)noise reduction technology based on the nearest adjacent layer applicated in ultra-low dose CT(ULDCT)for urinary calculi.Metho...Objective To observe the value of self-supervised deep learning artificial intelligence(AI)noise reduction technology based on the nearest adjacent layer applicated in ultra-low dose CT(ULDCT)for urinary calculi.Methods Eighty-eight urinary calculi patients were prospectively enrolled.Low dose CT(LDCT)and ULDCT scanning were performed,and the effective dose(ED)of each scanning protocol were calculated.The patients were then randomly divided into training set(n=75)and test set(n=13),and a self-supervised deep learning AI noise reduction system based on the nearest adjacent layer constructed with ULDCT images in training set was used for reducing noise of ULDCT images in test set.In test set,the quality of ULDCT images before and after AI noise reduction were compared with LDCT images,i.e.Blind/Referenceless Image Spatial Quality Evaluator(BRISQUE)scores,image noise(SD ROI)and signal-to-noise ratio(SNR).Results The tube current,the volume CT dose index and the dose length product of abdominal ULDCT scanning protocol were all lower compared with those of LDCT scanning protocol(all P<0.05),with a decrease of ED for approximately 82.66%.For 13 patients with urinary calculi in test set,BRISQUE score showed that the quality level of ULDCT images before AI noise reduction reached 54.42%level but raised to 95.76%level of LDCT images after AI noise reduction.Both ULDCT images after AI noise reduction and LDCT images had lower SD ROI and higher SNR than ULDCT images before AI noise reduction(all adjusted P<0.05),whereas no significant difference was found between the former two(both adjusted P>0.05).Conclusion Self-supervised learning AI noise reduction technology based on the nearest adjacent layer could effectively reduce noise and improve image quality of urinary calculi ULDCT images,being conducive for clinical application of ULDCT.展开更多
Objective To observe the value of artificial intelligence(AI)models based on non-contrast chest CT for measuring bone mineral density(BMD).Methods Totally 380 subjects who underwent both non-contrast chest CT and quan...Objective To observe the value of artificial intelligence(AI)models based on non-contrast chest CT for measuring bone mineral density(BMD).Methods Totally 380 subjects who underwent both non-contrast chest CT and quantitative CT(QCT)BMD examination were retrospectively enrolled and divided into training set(n=304)and test set(n=76)at a ratio of 8∶2.The mean BMD of L1—L3 vertebrae were measured based on QCT.Spongy bones of T5—T10 vertebrae were segmented as ROI,radiomics(Rad)features were extracted,and machine learning(ML),Rad and deep learning(DL)models were constructed for classification of osteoporosis(OP)and evaluating BMD,respectively.Receiver operating characteristic curves were drawn,and area under the curves(AUC)were calculated to evaluate the efficacy of each model for classification of OP.Bland-Altman analysis and Pearson correlation analysis were performed to explore the consistency and correlation of each model with QCT for measuring BMD.Results Among ML and Rad models,ML Bagging-OP and Rad Bagging-OP had the best performances for classification of OP.In test set,AUC of ML Bagging-OP,Rad Bagging-OP and DL OP for classification of OP was 0.943,0.944 and 0.947,respectively,with no significant difference(all P>0.05).BMD obtained with all the above models had good consistency with those measured with QCT(most of the differences were within the range of Ax-G±1.96 s),which were highly positively correlated(r=0.910—0.974,all P<0.001).Conclusion AI models based on non-contrast chest CT had high efficacy for classification of OP,and good consistency of BMD measurements were found between AI models and QCT.展开更多
[Objective]Real-time monitoring of cow ruminant behavior is of paramount importance for promptly obtaining relevant information about cow health and predicting cow diseases.Currently,various strategies have been propo...[Objective]Real-time monitoring of cow ruminant behavior is of paramount importance for promptly obtaining relevant information about cow health and predicting cow diseases.Currently,various strategies have been proposed for monitoring cow ruminant behavior,including video surveillance,sound recognition,and sensor monitoring methods.How‐ever,the application of edge device gives rise to the issue of inadequate real-time performance.To reduce the volume of data transmission and cloud computing workload while achieving real-time monitoring of dairy cow rumination behavior,a real-time monitoring method was proposed for cow ruminant behavior based on edge computing.[Methods]Autono‐mously designed edge devices were utilized to collect and process six-axis acceleration signals from cows in real-time.Based on these six-axis data,two distinct strategies,federated edge intelligence and split edge intelligence,were investigat‐ed for the real-time recognition of cow ruminant behavior.Focused on the real-time recognition method for cow ruminant behavior leveraging federated edge intelligence,the CA-MobileNet v3 network was proposed by enhancing the MobileNet v3 network with a collaborative attention mechanism.Additionally,a federated edge intelligence model was designed uti‐lizing the CA-MobileNet v3 network and the FedAvg federated aggregation algorithm.In the study on split edge intelli‐gence,a split edge intelligence model named MobileNet-LSTM was designed by integrating the MobileNet v3 network with a fusion collaborative attention mechanism and the Bi-LSTM network.[Results and Discussions]Through compara‐tive experiments with MobileNet v3 and MobileNet-LSTM,the federated edge intelligence model based on CA-Mo‐bileNet v3 achieved an average Precision rate,Recall rate,F1-Score,Specificity,and Accuracy of 97.1%,97.9%,97.5%,98.3%,and 98.2%,respectively,yielding the best recognition performance.[Conclusions]It is provided a real-time and effective method for monitoring cow ruminant behavior,and the proposed federated edge intelligence model can be ap‐plied in practical settings.展开更多
深度学习是人工智能领域的热门研究方向之一,它通过构建多层人工神经网络模仿人脑对数据的处理机制。大型语言模型(large language model,LLM)基于深度学习的架构,在无需编程指令的情况下,能通过分析大量数据以获得理解和生成人类语言...深度学习是人工智能领域的热门研究方向之一,它通过构建多层人工神经网络模仿人脑对数据的处理机制。大型语言模型(large language model,LLM)基于深度学习的架构,在无需编程指令的情况下,能通过分析大量数据以获得理解和生成人类语言的能力,被广泛应用于自然语言处理、计算机视觉、智慧医疗、智慧交通等诸多领域。文章总结了LLM在医疗领域的应用,涵盖了LLM针对医疗任务的基本训练流程、特殊策略以及在具体医疗场景中的应用。同时,进一步讨论了LLM在应用中面临的挑战,包括决策过程缺乏透明度、输出准确性以及隐私、伦理问题等,随后列举了相应的改进策略。最后,文章展望了LLM在医疗领域的未来发展趋势,及其对人类健康事业发展的潜在影响。展开更多
文摘Missile interception problem can be regarded as a two-person zero-sum differential games problem,which depends on the solution of Hamilton-Jacobi-Isaacs(HJI)equa-tion.It has been proved impossible to obtain a closed-form solu-tion due to the nonlinearity of HJI equation,and many iterative algorithms are proposed to solve the HJI equation.Simultane-ous policy updating algorithm(SPUA)is an effective algorithm for solving HJI equation,but it is an on-policy integral reinforce-ment learning(IRL).For online implementation of SPUA,the dis-turbance signals need to be adjustable,which is unrealistic.In this paper,an off-policy IRL algorithm based on SPUA is pro-posed without making use of any knowledge of the systems dynamics.Then,a neural-network based online adaptive critic implementation scheme of the off-policy IRL algorithm is pre-sented.Based on the online off-policy IRL method,a computa-tional intelligence interception guidance(CIIG)law is developed for intercepting high-maneuvering target.As a model-free method,intercepting targets can be achieved through measur-ing system data online.The effectiveness of the CIIG is verified through two missile and target engagement scenarios.
文摘)The 2008 IEEE World Congress on Computational Intelligence (WCCI 2008) will be held at the HongKong Convention and Exhibition Centre during June 1-6, 2008. WCCI 2008 will be the fifth milestone inthis series with a glorious history from WCCI 1994 in Orlando, WCCI 1998 in Anchorage, WCCI 2002in Honolulu, to WCCI 2006 in Vancouver. Sponsored by the IEEE Computational Intelligence Society,
基金Project(61473078)supported by the National Natural Science Foundation of ChinaProject(2015-2019)supported by the Program for Changjiang Scholars from the Ministry of Education,China+1 种基金Project(16510711100)supported by International Collaborative Project of the Shanghai Committee of Science and Technology,ChinaProject(KJ2017A418)supported by Anhui University Science Research,China
文摘In order to improve the performance of multi-objective workflow scheduling in cloud system, a multi-swarm multiobjective optimization algorithm(MSMOOA) is proposed to satisfy multiple conflicting objectives. Inspired by division of the same species into multiple swarms for different objectives and information sharing among these swarms in nature, each physical machine in the data center is considered a swarm and employs improved multi-objective particle swarm optimization to find out non-dominated solutions with one objective in MSMOOA. The particles in each swarm are divided into two classes and adopt different strategies to evolve cooperatively. One class of particles can communicate with several swarms simultaneously to promote the information sharing among swarms and the other class of particles can only exchange information with the particles located in the same swarm. Furthermore, in order to avoid the influence by the elastic available resources, a manager server is adopted in the cloud data center to collect the available resources for scheduling. The quality of the proposed method with other related approaches is evaluated by using hybrid and parallel workflow applications. The experiment results highlight the better performance of the MSMOOA than that of compared algorithms.
文摘Objective To observe the value of self-supervised deep learning artificial intelligence(AI)noise reduction technology based on the nearest adjacent layer applicated in ultra-low dose CT(ULDCT)for urinary calculi.Methods Eighty-eight urinary calculi patients were prospectively enrolled.Low dose CT(LDCT)and ULDCT scanning were performed,and the effective dose(ED)of each scanning protocol were calculated.The patients were then randomly divided into training set(n=75)and test set(n=13),and a self-supervised deep learning AI noise reduction system based on the nearest adjacent layer constructed with ULDCT images in training set was used for reducing noise of ULDCT images in test set.In test set,the quality of ULDCT images before and after AI noise reduction were compared with LDCT images,i.e.Blind/Referenceless Image Spatial Quality Evaluator(BRISQUE)scores,image noise(SD ROI)and signal-to-noise ratio(SNR).Results The tube current,the volume CT dose index and the dose length product of abdominal ULDCT scanning protocol were all lower compared with those of LDCT scanning protocol(all P<0.05),with a decrease of ED for approximately 82.66%.For 13 patients with urinary calculi in test set,BRISQUE score showed that the quality level of ULDCT images before AI noise reduction reached 54.42%level but raised to 95.76%level of LDCT images after AI noise reduction.Both ULDCT images after AI noise reduction and LDCT images had lower SD ROI and higher SNR than ULDCT images before AI noise reduction(all adjusted P<0.05),whereas no significant difference was found between the former two(both adjusted P>0.05).Conclusion Self-supervised learning AI noise reduction technology based on the nearest adjacent layer could effectively reduce noise and improve image quality of urinary calculi ULDCT images,being conducive for clinical application of ULDCT.
文摘Objective To observe the value of artificial intelligence(AI)models based on non-contrast chest CT for measuring bone mineral density(BMD).Methods Totally 380 subjects who underwent both non-contrast chest CT and quantitative CT(QCT)BMD examination were retrospectively enrolled and divided into training set(n=304)and test set(n=76)at a ratio of 8∶2.The mean BMD of L1—L3 vertebrae were measured based on QCT.Spongy bones of T5—T10 vertebrae were segmented as ROI,radiomics(Rad)features were extracted,and machine learning(ML),Rad and deep learning(DL)models were constructed for classification of osteoporosis(OP)and evaluating BMD,respectively.Receiver operating characteristic curves were drawn,and area under the curves(AUC)were calculated to evaluate the efficacy of each model for classification of OP.Bland-Altman analysis and Pearson correlation analysis were performed to explore the consistency and correlation of each model with QCT for measuring BMD.Results Among ML and Rad models,ML Bagging-OP and Rad Bagging-OP had the best performances for classification of OP.In test set,AUC of ML Bagging-OP,Rad Bagging-OP and DL OP for classification of OP was 0.943,0.944 and 0.947,respectively,with no significant difference(all P>0.05).BMD obtained with all the above models had good consistency with those measured with QCT(most of the differences were within the range of Ax-G±1.96 s),which were highly positively correlated(r=0.910—0.974,all P<0.001).Conclusion AI models based on non-contrast chest CT had high efficacy for classification of OP,and good consistency of BMD measurements were found between AI models and QCT.
文摘[Objective]Real-time monitoring of cow ruminant behavior is of paramount importance for promptly obtaining relevant information about cow health and predicting cow diseases.Currently,various strategies have been proposed for monitoring cow ruminant behavior,including video surveillance,sound recognition,and sensor monitoring methods.How‐ever,the application of edge device gives rise to the issue of inadequate real-time performance.To reduce the volume of data transmission and cloud computing workload while achieving real-time monitoring of dairy cow rumination behavior,a real-time monitoring method was proposed for cow ruminant behavior based on edge computing.[Methods]Autono‐mously designed edge devices were utilized to collect and process six-axis acceleration signals from cows in real-time.Based on these six-axis data,two distinct strategies,federated edge intelligence and split edge intelligence,were investigat‐ed for the real-time recognition of cow ruminant behavior.Focused on the real-time recognition method for cow ruminant behavior leveraging federated edge intelligence,the CA-MobileNet v3 network was proposed by enhancing the MobileNet v3 network with a collaborative attention mechanism.Additionally,a federated edge intelligence model was designed uti‐lizing the CA-MobileNet v3 network and the FedAvg federated aggregation algorithm.In the study on split edge intelli‐gence,a split edge intelligence model named MobileNet-LSTM was designed by integrating the MobileNet v3 network with a fusion collaborative attention mechanism and the Bi-LSTM network.[Results and Discussions]Through compara‐tive experiments with MobileNet v3 and MobileNet-LSTM,the federated edge intelligence model based on CA-Mo‐bileNet v3 achieved an average Precision rate,Recall rate,F1-Score,Specificity,and Accuracy of 97.1%,97.9%,97.5%,98.3%,and 98.2%,respectively,yielding the best recognition performance.[Conclusions]It is provided a real-time and effective method for monitoring cow ruminant behavior,and the proposed federated edge intelligence model can be ap‐plied in practical settings.
文摘深度学习是人工智能领域的热门研究方向之一,它通过构建多层人工神经网络模仿人脑对数据的处理机制。大型语言模型(large language model,LLM)基于深度学习的架构,在无需编程指令的情况下,能通过分析大量数据以获得理解和生成人类语言的能力,被广泛应用于自然语言处理、计算机视觉、智慧医疗、智慧交通等诸多领域。文章总结了LLM在医疗领域的应用,涵盖了LLM针对医疗任务的基本训练流程、特殊策略以及在具体医疗场景中的应用。同时,进一步讨论了LLM在应用中面临的挑战,包括决策过程缺乏透明度、输出准确性以及隐私、伦理问题等,随后列举了相应的改进策略。最后,文章展望了LLM在医疗领域的未来发展趋势,及其对人类健康事业发展的潜在影响。
