In recent years,the problem of privacy leakage has attracted increasing attentions.Therefore,machine learning privacy protection becomes crucial research topic.In this paper,the Paillier homomorphic encryption algorit...In recent years,the problem of privacy leakage has attracted increasing attentions.Therefore,machine learning privacy protection becomes crucial research topic.In this paper,the Paillier homomorphic encryption algorithm is proposed to protect the privacy data.The original LeNet-5 convolutional neural network model was first improved.Then the activation function was modified and the C5 layer was removed to reduce the number of model parameters and improve the operation efficiency.Finally,by mapping the operation of each layer in the convolutional neural network from the plaintext domain to the ciphertext domain,an improved LeNet-5 model that can run on encrypted data was constructed.The purpose of using machine learning algorithmwas realized and privacywas ensured at the same time.The analysis shows that the model is feasible and the efficiency is improved.展开更多
Using a dynamic laser monitoring technique,the solubility of 3-nitro-1,2,4-triazole-5-one(NTO)was investigated in two different binary systems,namely hydroxylamine nitrate(HAN)-water and boric acid(HB)-water ranging f...Using a dynamic laser monitoring technique,the solubility of 3-nitro-1,2,4-triazole-5-one(NTO)was investigated in two different binary systems,namely hydroxylamine nitrate(HAN)-water and boric acid(HB)-water ranging from 278.15 K to 318.15 K.The solubility in each system was found to be positively correlated with temperature.Furthermore,solubility data were analyzed using four equations:the modified Apelblat equation,Van’t Hoff equation,λh equation and CNIBS/R-K equations,and they provided satisfactory results for both two systems.The average root-mean-square deviation(105RMSD)values for these models were less than 13.93.Calculations utilizing the Van’t Hoff equation and Gibbs equations facilitated the derivation of apparent thermodynamic properties of NTO dissolution in the two systems,including values for Gibbs free energy,enthalpy and entropy.The%ζ_(H)is larger than%ζ_(TS),and all the%ζ_(H)data are≥58.63%,indicating that the enthalpy make a greater contribution than entropy to theΔG_(soln)^(Θ).展开更多
The 5E model includes Engagement,Exploration,Explanation,Elaboration,and Evaluation,with“Evaluation”at the end,conflicting with teaching learning-evaluation consistency.Thus,formative levaluation is integrated into ...The 5E model includes Engagement,Exploration,Explanation,Elaboration,and Evaluation,with“Evaluation”at the end,conflicting with teaching learning-evaluation consistency.Thus,formative levaluation is integrated into the first four stages,and a summative evaluation table is designed for the fifth,enabling students to self-evaluate and reflect.Elementary school English picture book teaching is used as an example to demonstrate the optimized model's application.展开更多
As the commercialization of the fifth gen-eration communication(5G)is sped up,its system testing scheme is vital for the successful deployment of 5G.Especially,5G relies on the scale-increased multiple-input-multiple ...As the commercialization of the fifth gen-eration communication(5G)is sped up,its system testing scheme is vital for the successful deployment of 5G.Especially,5G relies on the scale-increased multiple-input-multiple output(MIMO)technique to improve its capacity and coverage.Thus,testing new functions of the 5G MIMO system accurately and ef-ficiently,including beamforming(beam-tracking with movement)and multiple-user(MU)multiplexing,is a challenging task.This paper tries to construct a lab-oratorial hardware and conduct equipment-controlled field testing.Firstly,the testing scheme is presented,which is composed of the framework,the channel models and the validation methods.Then,the channel model principles are explained in detail due to its di-rect influence on the testing accuracy.Specifically,we utilize the spatial consistency and the multi-link cor-relation properties to emulate the high-speed dynamic time-varying(HDT)and the multiple-cell(MC)-MU-MIMO channels.Finally,the above testing scheme is verified in a Shanghai 5G field experiment with the practical commercial equipment and the channel em-ulator.The results show that the 5G new functions are tested accurately and efficiently by switching the channel emulation configurations.展开更多
A“water”accelerated metal-free catalytic system has been discovered for the Mukaiyama-aldol reaction.The system involves the use of B(C_(6)F_(5))_(3) as a catalyst,which is water-tolerant and able to activate the ca...A“water”accelerated metal-free catalytic system has been discovered for the Mukaiyama-aldol reaction.The system involves the use of B(C_(6)F_(5))_(3) as a catalyst,which is water-tolerant and able to activate the carbonyl group through a hydrogen bonding network generated by the catalyst.This activation method allows the reactions to be performed with very low catalyst loading,as low as 0.5 mol%.The scope of substrates is broad and a wide variety of functional groups are well tolerated.Diverse aliphatic aldehydes,aromatic aldehydes,unsaturated aldehydes and aromatic ketones coupled with silyl enol ethers/silyl ketone acetals to generate their correspondingβ-hydroxy carbonyl compounds in moderate to good yields.This discovery represents a significant advancement in the field of organic synthesis,as it provides a new,practical and sustainable solution for carbon-carbon bond formation in water.展开更多
BACKGROUND:Sepsis survivors experience poor long-term quality of life post-discharge.The aim of this study was to analyze the factors that impact the long-term quality of life of sepsis survivors and develop a clinica...BACKGROUND:Sepsis survivors experience poor long-term quality of life post-discharge.The aim of this study was to analyze the factors that impact the long-term quality of life of sepsis survivors and develop a clinical prediction model.METHODS:A total of 442 sepsis patients from the Emergency Intensive Care Unit of a tertiary hospital in Wenzhou were included.These patients were assigned to the training set or the validation set at a ratio of 7:3.The European Quality of Life 5 Dimensions 5 Level Version(EQ-5D-5L) questionnaire was used to evaluate the quality of life in sepsis survivors one year after discharge.Multivariate logistic regression analysis was used to identify predictors,which were then used to develop the prediction model and subsequently derive a scoring system.The model's effectiveness was assessed using an area under the receiver operating characteristic curve,calibration curves,and clinical decision analysis.RESULTS:Of the 442 patients included,70 died one year after discharge,and 372 completed the questionnaire.A total of 46.6% of sepsis survivors have poor quality of life one year after discharge in the training set.Multivariate logistic regression revealed that age,platelet,serum albumin,serum urea,and C-reactive protein were independent risk factors for poor quality of life in sepsis survivors.The area under the curve of the scoring system was 0.777(95% CI:0.726–0.828).The calibration curves showed that it was well calibrated.Decision curve analysis indicated that the scoring system provided good clinical usefulness.The internal validation also demonstrated its effectiveness.CONCLUSION:The prediction model incorporating five risk factors may predict quality of life one year after discharge in sepsis survivors,which provides a measure to develop post-discharge rehabilitation and follow-up plans for this patient population.展开更多
该文旨在比较DeepSeek-V3.1与ChatGPT-5在结直肠癌肝转移(CRLM)多学科团队(MDT)决策中的应用表现,评估其与MDT专家意见的一致性,为大语言模型(LLMs)的临床实践提供循证依据与优化方向。该文基于真实世界数据与最新指南,设计了6例涵盖不...该文旨在比较DeepSeek-V3.1与ChatGPT-5在结直肠癌肝转移(CRLM)多学科团队(MDT)决策中的应用表现,评估其与MDT专家意见的一致性,为大语言模型(LLMs)的临床实践提供循证依据与优化方向。该文基于真实世界数据与最新指南,设计了6例涵盖不同肿瘤负荷、基因突变谱和体能状态的虚拟CRLM病例,通过结构化的提示策略,在DeepSeek-V3.1与ChatGPT-5模型中分别生成MDT治疗建议。由4名MDT专家采用7维度5级李克特量表对模型输出进行独立评审。并通过统计学分析对两款模型在各个病例、各项维度和各个学科的表现分别进行比较。2款大语言模型在所有病例中的综合得分均≥4.0分(满分5分),表明其在复杂的MDT决策场景下具备可接受的临床效能。在跨维度分析中,两者在清晰度、个体化程度、抗幻觉能力和伦理安全4项上得分较高,而在准确性、全面性和前沿性方面仍有一定提升空间。DeepSeek-V3.1在整体表现(4.27±0.77 vs 4.08±0.86)、前沿性(3.90±0.65 vs 3.24±0.72)与伦理安全(4.87±0.34 vs 4.58±0.65)方面显著优于ChatGPT-5(P<0.05);在放疗领域亦明显领先(4.55±0.67 vs 3.38±0.91,P<0.01)。ChatGPT-5则在胃肠外科领域表现优于DeepSeek-V3.1(4.48±0.67 vs 4.17±0.85,P=0.02)。DeepSeek-V3.1与ChatGPT-5均表现出为CRLM-MDT决策提供可靠建议的良好能力。其中,Deep Seek-V3.1在前沿知识整合、伦理安全性及放射肿瘤学领域展现出显著优势,而Chat GPT-5则在胃肠外科方面表现更优,二者形成优势互补。该文证实了大型语言模型作为“MDT协作者”的可行性,为缩小地域间诊疗水平差距、提升临床决策效率提供了一项便捷可靠的技术方案。展开更多
To clarify the high temperature flow stress behavior and microstructures evolution of a V-5Cr-5Ti （mass fraction, %） alloy, the isothermal hot compression tests were conducted in the temperature range of 1423-1573 K...To clarify the high temperature flow stress behavior and microstructures evolution of a V-5Cr-5Ti （mass fraction, %） alloy, the isothermal hot compression tests were conducted in the temperature range of 1423-1573 K with strain rates of 0.01, 0.1, and 1 s-1. The results show that the measured flow stress should be revised by friction and the calculated values of friction coefficient m are in the range of 0.45-0.56. Arrhenius-type constitutive equation was developed by regression analysis. The comparison between the experimental and predicted flow stress shows that the R~ and the average absolute relative error （AARE） are 0.948 and 5.44%, respectively. The measured apparent activation energy Qa is in the range of 540-890 kJ/mol. Both dis-continuous dynamic recrystallization （DDRX） and continuous dynamic recrystallization （CDRX） mechanisms are observed in the deformed alloy, but dynamic recovery （DRV） is the dominant softening mechanism up to a true strain of 1.5.展开更多
引入了Inception-SE卷积模块组来提升LeNet-5网络的广度与深度,运用SE模块增强了有用的特征并抑制了对当前任务用处不大的特征;使用BN层和Dropout优化网络,防止梯度弥散,提升精度;使用全局池化层(global average pooling,GAP)代替全连...引入了Inception-SE卷积模块组来提升LeNet-5网络的广度与深度,运用SE模块增强了有用的特征并抑制了对当前任务用处不大的特征;使用BN层和Dropout优化网络,防止梯度弥散,提升精度;使用全局池化层(global average pooling,GAP)代替全连接层来减少网络计算参数.研究结果表明:改进后网络的识别精度达到了99.88%,比传统的LeNet-5网络提高了1.71%.展开更多
基金The National Natural Science Foundation of China(No.61572521)Engineering University of PAP Innovation Team Science Foundation(No.KYTD201805)Natural Science Basic Research Plan in Shaanxi Province of China(2021JM252).
文摘In recent years,the problem of privacy leakage has attracted increasing attentions.Therefore,machine learning privacy protection becomes crucial research topic.In this paper,the Paillier homomorphic encryption algorithm is proposed to protect the privacy data.The original LeNet-5 convolutional neural network model was first improved.Then the activation function was modified and the C5 layer was removed to reduce the number of model parameters and improve the operation efficiency.Finally,by mapping the operation of each layer in the convolutional neural network from the plaintext domain to the ciphertext domain,an improved LeNet-5 model that can run on encrypted data was constructed.The purpose of using machine learning algorithmwas realized and privacywas ensured at the same time.The analysis shows that the model is feasible and the efficiency is improved.
文摘Using a dynamic laser monitoring technique,the solubility of 3-nitro-1,2,4-triazole-5-one(NTO)was investigated in two different binary systems,namely hydroxylamine nitrate(HAN)-water and boric acid(HB)-water ranging from 278.15 K to 318.15 K.The solubility in each system was found to be positively correlated with temperature.Furthermore,solubility data were analyzed using four equations:the modified Apelblat equation,Van’t Hoff equation,λh equation and CNIBS/R-K equations,and they provided satisfactory results for both two systems.The average root-mean-square deviation(105RMSD)values for these models were less than 13.93.Calculations utilizing the Van’t Hoff equation and Gibbs equations facilitated the derivation of apparent thermodynamic properties of NTO dissolution in the two systems,including values for Gibbs free energy,enthalpy and entropy.The%ζ_(H)is larger than%ζ_(TS),and all the%ζ_(H)data are≥58.63%,indicating that the enthalpy make a greater contribution than entropy to theΔG_(soln)^(Θ).
基金This paper is funded by Project Information:2023 Guangdong Undergraduate Colleges and Universities Teaching Quality and Teaching Reform Project Construction Project,Project Name:Action Research on Whole-area Nurturing of English Reading Teaching in Universities,Secondary and Primary Schools under the Perspective of Discipline Nurturing.Project serial number:895.
文摘The 5E model includes Engagement,Exploration,Explanation,Elaboration,and Evaluation,with“Evaluation”at the end,conflicting with teaching learning-evaluation consistency.Thus,formative levaluation is integrated into the first four stages,and a summative evaluation table is designed for the fifth,enabling students to self-evaluate and reflect.Elementary school English picture book teaching is used as an example to demonstrate the optimized model's application.
基金supported in part by National Natural Science Foundation of China under Grant 62201087,Grant 62525101,in part by the National Key R&D Program of China under Grant 2023YFB2904803in part by the Guangdong Major Project of Basic and Applied Basic Research under Grant 2023B0303000001+1 种基金in part by the Natural Science Foundation of Beijing-Xiaomi Innovation Joint Foundation under Grant L243002in part by the Beijing University of Posts and Telecommunications-China Mobile Research Institute Joint Institute.
文摘As the commercialization of the fifth gen-eration communication(5G)is sped up,its system testing scheme is vital for the successful deployment of 5G.Especially,5G relies on the scale-increased multiple-input-multiple output(MIMO)technique to improve its capacity and coverage.Thus,testing new functions of the 5G MIMO system accurately and ef-ficiently,including beamforming(beam-tracking with movement)and multiple-user(MU)multiplexing,is a challenging task.This paper tries to construct a lab-oratorial hardware and conduct equipment-controlled field testing.Firstly,the testing scheme is presented,which is composed of the framework,the channel models and the validation methods.Then,the channel model principles are explained in detail due to its di-rect influence on the testing accuracy.Specifically,we utilize the spatial consistency and the multi-link cor-relation properties to emulate the high-speed dynamic time-varying(HDT)and the multiple-cell(MC)-MU-MIMO channels.Finally,the above testing scheme is verified in a Shanghai 5G field experiment with the practical commercial equipment and the channel em-ulator.The results show that the 5G new functions are tested accurately and efficiently by switching the channel emulation configurations.
基金financial support from the Start-up Grant of Nanjing Tech University(Nos.38274017103,38037037)financial support from Distinguished University Professor grant(Nanyang Technological University)+1 种基金the Agency for Science,Technology and Research(A∗STAR)under its MTC Individual Research Grants(No.M21K2c0114)RIE2025 MTC Programmatic Fund(No.M22K9b0049).
文摘A“water”accelerated metal-free catalytic system has been discovered for the Mukaiyama-aldol reaction.The system involves the use of B(C_(6)F_(5))_(3) as a catalyst,which is water-tolerant and able to activate the carbonyl group through a hydrogen bonding network generated by the catalyst.This activation method allows the reactions to be performed with very low catalyst loading,as low as 0.5 mol%.The scope of substrates is broad and a wide variety of functional groups are well tolerated.Diverse aliphatic aldehydes,aromatic aldehydes,unsaturated aldehydes and aromatic ketones coupled with silyl enol ethers/silyl ketone acetals to generate their correspondingβ-hydroxy carbonyl compounds in moderate to good yields.This discovery represents a significant advancement in the field of organic synthesis,as it provides a new,practical and sustainable solution for carbon-carbon bond formation in water.
基金supported by the National Natural Science Foundation of China (82272202)the Wenzhou HighLevel Innovation Team (2024R3002)the Provincial Advantageous Characteristic Discipline of Wenzhou Medical University (Clinical Medicine)。
文摘BACKGROUND:Sepsis survivors experience poor long-term quality of life post-discharge.The aim of this study was to analyze the factors that impact the long-term quality of life of sepsis survivors and develop a clinical prediction model.METHODS:A total of 442 sepsis patients from the Emergency Intensive Care Unit of a tertiary hospital in Wenzhou were included.These patients were assigned to the training set or the validation set at a ratio of 7:3.The European Quality of Life 5 Dimensions 5 Level Version(EQ-5D-5L) questionnaire was used to evaluate the quality of life in sepsis survivors one year after discharge.Multivariate logistic regression analysis was used to identify predictors,which were then used to develop the prediction model and subsequently derive a scoring system.The model's effectiveness was assessed using an area under the receiver operating characteristic curve,calibration curves,and clinical decision analysis.RESULTS:Of the 442 patients included,70 died one year after discharge,and 372 completed the questionnaire.A total of 46.6% of sepsis survivors have poor quality of life one year after discharge in the training set.Multivariate logistic regression revealed that age,platelet,serum albumin,serum urea,and C-reactive protein were independent risk factors for poor quality of life in sepsis survivors.The area under the curve of the scoring system was 0.777(95% CI:0.726–0.828).The calibration curves showed that it was well calibrated.Decision curve analysis indicated that the scoring system provided good clinical usefulness.The internal validation also demonstrated its effectiveness.CONCLUSION:The prediction model incorporating five risk factors may predict quality of life one year after discharge in sepsis survivors,which provides a measure to develop post-discharge rehabilitation and follow-up plans for this patient population.
文摘该文旨在比较DeepSeek-V3.1与ChatGPT-5在结直肠癌肝转移(CRLM)多学科团队(MDT)决策中的应用表现,评估其与MDT专家意见的一致性,为大语言模型(LLMs)的临床实践提供循证依据与优化方向。该文基于真实世界数据与最新指南,设计了6例涵盖不同肿瘤负荷、基因突变谱和体能状态的虚拟CRLM病例,通过结构化的提示策略,在DeepSeek-V3.1与ChatGPT-5模型中分别生成MDT治疗建议。由4名MDT专家采用7维度5级李克特量表对模型输出进行独立评审。并通过统计学分析对两款模型在各个病例、各项维度和各个学科的表现分别进行比较。2款大语言模型在所有病例中的综合得分均≥4.0分(满分5分),表明其在复杂的MDT决策场景下具备可接受的临床效能。在跨维度分析中,两者在清晰度、个体化程度、抗幻觉能力和伦理安全4项上得分较高,而在准确性、全面性和前沿性方面仍有一定提升空间。DeepSeek-V3.1在整体表现(4.27±0.77 vs 4.08±0.86)、前沿性(3.90±0.65 vs 3.24±0.72)与伦理安全(4.87±0.34 vs 4.58±0.65)方面显著优于ChatGPT-5(P<0.05);在放疗领域亦明显领先(4.55±0.67 vs 3.38±0.91,P<0.01)。ChatGPT-5则在胃肠外科领域表现优于DeepSeek-V3.1(4.48±0.67 vs 4.17±0.85,P=0.02)。DeepSeek-V3.1与ChatGPT-5均表现出为CRLM-MDT决策提供可靠建议的良好能力。其中,Deep Seek-V3.1在前沿知识整合、伦理安全性及放射肿瘤学领域展现出显著优势,而Chat GPT-5则在胃肠外科方面表现更优,二者形成优势互补。该文证实了大型语言模型作为“MDT协作者”的可行性,为缩小地域间诊疗水平差距、提升临床决策效率提供了一项便捷可靠的技术方案。
基金Project(11105127) supported by the National Natural Science Foundation of China
文摘To clarify the high temperature flow stress behavior and microstructures evolution of a V-5Cr-5Ti （mass fraction, %） alloy, the isothermal hot compression tests were conducted in the temperature range of 1423-1573 K with strain rates of 0.01, 0.1, and 1 s-1. The results show that the measured flow stress should be revised by friction and the calculated values of friction coefficient m are in the range of 0.45-0.56. Arrhenius-type constitutive equation was developed by regression analysis. The comparison between the experimental and predicted flow stress shows that the R~ and the average absolute relative error （AARE） are 0.948 and 5.44%, respectively. The measured apparent activation energy Qa is in the range of 540-890 kJ/mol. Both dis-continuous dynamic recrystallization （DDRX） and continuous dynamic recrystallization （CDRX） mechanisms are observed in the deformed alloy, but dynamic recovery （DRV） is the dominant softening mechanism up to a true strain of 1.5.
文摘引入了Inception-SE卷积模块组来提升LeNet-5网络的广度与深度,运用SE模块增强了有用的特征并抑制了对当前任务用处不大的特征;使用BN层和Dropout优化网络,防止梯度弥散,提升精度;使用全局池化层(global average pooling,GAP)代替全连接层来减少网络计算参数.研究结果表明:改进后网络的识别精度达到了99.88%,比传统的LeNet-5网络提高了1.71%.
