鉴于ROC曲线下面积(Area Under the ROC Curve,AUC)对数据分布的不敏感特性,面向AUC的对抗训练(AdAUC)近来已成为机器学习领域中抵御长尾分布下对抗攻击的有效范式之一。当前主流方法大多遵循基于平方替代损失的AUC对抗训练框架,并将成...鉴于ROC曲线下面积(Area Under the ROC Curve,AUC)对数据分布的不敏感特性,面向AUC的对抗训练(AdAUC)近来已成为机器学习领域中抵御长尾分布下对抗攻击的有效范式之一。当前主流方法大多遵循基于平方替代损失的AUC对抗训练框架,并将成对比较形式的AUC对抗损失重构为一个逐样本的随机鞍点优化问题,克服端到端的计算瓶颈。然而,面向复杂的实际应用场景,基于平方损失设计的AUC对抗训练框架恐难以适应多样的下游任务需求。此外,与传统对抗训练范式类似,面向AUC的对抗训练方法在提高模型对抗鲁棒性的同时,也会降低模型在正常样本上的AUC性能,而目前鲜有针对该问题的有效解决方案。鉴于此,本文对如何构建一般化的高效AUC对抗机器学习范式展开系统研究。首先,提出了一种基于标准化分数扰动的通用AUC对抗训练框架(NSAdAUC),在相对温和的条件下,该框架可通过直接扰动模型对样本的预测得分实现对AUC指标的攻击,且不依赖于特定的AUC替代损失。在此基础上,本文进一步指出鲁棒AUC误差可分解为标准AUC误差和边界AUC误差两项之和,并据此设计了一种基于排序感知对抗正则化的AUC对抗训练框架(RARAdAUC),同时兼顾模型的标准AUC和鲁棒AUC性能。为验证所提框架的有效性,在5个长尾基准数据集上进行了大量实验,结果表明所提NSAdAUC和RARAdAUC框架在多种对抗攻击下的鲁棒性均优于现有方法,可在平均意义上分别产生0.94%、5.52%的标准AUC和5.69%、5.41%的鲁棒AUC性能提升。展开更多
A Receiver Operating Characteristic(ROC)analysis of a power is important and useful in clinical trials.A Classical Conditional Power(CCP)is a probability of a classical rejection region given values of true treatment ...A Receiver Operating Characteristic(ROC)analysis of a power is important and useful in clinical trials.A Classical Conditional Power(CCP)is a probability of a classical rejection region given values of true treatment effect and interim result.For hypotheses and reversed hypotheses under normal models,we obtain analytical expressions of the ROC curves of the CCP,find optimal ROC curves of the CCP,investigate the superiority of the ROC curves of the CCP,calculate critical values of the False Positive Rate(FPR),True Positive Rate(TPR),and cutoff of the optimal CCP,and give go/no go decisions at the interim of the optimal CCP.In addition,extensive numerical experiments are carried out to exemplify our theoretical results.Finally,a real data example is performed to illustrate the go/no go decisions of the optimal CCP.展开更多
This article examines the influence of annealing temperature on fracture toughness and forming limit curves of dissimilar aluminum/silver sheets.In the cold roll bonding process,after brushing and acid washing,the pre...This article examines the influence of annealing temperature on fracture toughness and forming limit curves of dissimilar aluminum/silver sheets.In the cold roll bonding process,after brushing and acid washing,the prepared surfaces are placed on top of each other and by rolling with reduction more than 50%,the bonding between layers is established.In this research,the roll bonding process was done at room temperature,without the use of lubricants and with a 70%thickness reduction.Then,the final thickness of the Ag/Al bilayer sheet reached 350μm by several stages of cold rolling.Before cold rolling,it should be noted that to decrease the hardness created due to plastic deformation,the roll-bonded samples were subjected to annealing heat treatment at 400℃for 90 min.Thus,the final samples were annealed at 200,300 and 400℃for 90 min and cooled in a furnace to examine the annealing temperature effects.The uniaxial tensile and microhardness tests measured mechanical properties.Also,to investigate the fracture mechanism,the fractography of the cross-section was examined by scanning electron microscope(SEM).To evaluate the formability of Ag/Al bilayer sheets,forming limit curves were obtained experimentally through the Nakazima test.The resistance of composites to failure due to cracking was also investigated by fracture toughness.The results showed that annealing increases the elongation and formability of the Ag/Al bilayer sheet while reduces the ultimate tensile strength and fracture toughness.However,the changing trend is not the same at different temperatures,and according to the results,the most significant effect is obtained at 300℃and aluminum layers.It was also determined that by increasing annealing temperature,the fracture mechanism from shear ductile with small and shallow dimples becomes ductile with deep cavities.展开更多
Objective Primary liver cancer,predominantly hepatocellular carcinoma(HCC),is a significant global health issue,ranking as the sixth most diagnosed cancer and the third leading cause of cancer-related mortality.Accura...Objective Primary liver cancer,predominantly hepatocellular carcinoma(HCC),is a significant global health issue,ranking as the sixth most diagnosed cancer and the third leading cause of cancer-related mortality.Accurate and early diagnosis of HCC is crucial for effective treatment,as HCC and non-HCC malignancies like intrahepatic cholangiocarcinoma(ICC)exhibit different prognoses and treatment responses.Traditional diagnostic methods,including liver biopsy and contrast-enhanced ultrasound(CEUS),face limitations in applicability and objectivity.The primary objective of this study was to develop an advanced,lightweighted classification network capable of distinguishing HCC from other non-HCC malignancies by leveraging the automatic analysis of brightness changes in CEUS images.The ultimate goal was to create a user-friendly and cost-efficient computer-aided diagnostic tool that could assist radiologists in making more accurate and efficient clinical decisions.Methods This retrospective study encompassed a total of 161 patients,comprising 131 diagnosed with HCC and 30 with non-HCC malignancies.To achieve accurate tumor detection,the YOLOX network was employed to identify the region of interest(ROI)on both B-mode ultrasound and CEUS images.A custom-developed algorithm was then utilized to extract brightness change curves from the tumor and adjacent liver parenchyma regions within the CEUS images.These curves provided critical data for the subsequent analysis and classification process.To analyze the extracted brightness change curves and classify the malignancies,we developed and compared several models.These included one-dimensional convolutional neural networks(1D-ResNet,1D-ConvNeXt,and 1D-CNN),as well as traditional machine-learning methods such as support vector machine(SVM),ensemble learning(EL),k-nearest neighbor(KNN),and decision tree(DT).The diagnostic performance of each method in distinguishing HCC from non-HCC malignancies was rigorously evaluated using four key metrics:area under the receiver operating characteristic(AUC),accuracy(ACC),sensitivity(SE),and specificity(SP).Results The evaluation of the machine-learning methods revealed AUC values of 0.70 for SVM,0.56 for ensemble learning,0.63 for KNN,and 0.72 for the decision tree.These results indicated moderate to fair performance in classifying the malignancies based on the brightness change curves.In contrast,the deep learning models demonstrated significantly higher AUCs,with 1D-ResNet achieving an AUC of 0.72,1D-ConvNeXt reaching 0.82,and 1D-CNN obtaining the highest AUC of 0.84.Moreover,under the five-fold cross-validation scheme,the 1D-CNN model outperformed other models in both accuracy and specificity.Specifically,it achieved accuracy improvements of 3.8%to 10.0%and specificity enhancements of 6.6%to 43.3%over competing approaches.The superior performance of the 1D-CNN model highlighted its potential as a powerful tool for accurate classification.Conclusion The 1D-CNN model proved to be the most effective in differentiating HCC from non-HCC malignancies,surpassing both traditional machine-learning methods and other deep learning models.This study successfully developed a user-friendly and cost-efficient computer-aided diagnostic solution that would significantly enhances radiologists’diagnostic capabilities.By improving the accuracy and efficiency of clinical decision-making,this tool has the potential to positively impact patient care and outcomes.Future work may focus on further refining the model and exploring its integration with multimodal ultrasound data to maximize its accuracy and applicability.展开更多
On a compact Riemann surface with finite punctures P_(1),…P_(k),we define toric curves as multivalued,totallyunramified holomorphic maps to P^(n)with monodromy in a maximal torus of PSU(n+1).Toric solutions to SU(n+1...On a compact Riemann surface with finite punctures P_(1),…P_(k),we define toric curves as multivalued,totallyunramified holomorphic maps to P^(n)with monodromy in a maximal torus of PSU(n+1).Toric solutions to SU(n+1)Todasystems on X\{P_(1);…;P_(k)}are recognized by the associated toric curves in.We introduce character n-ensembles as-tuples of meromorphic one-forms with simple poles and purely imaginary periods,generating toric curves on minus finitelymany points.On X,we establish a correspondence between character-ensembles and toric solutions to the SU(n+1)system with finitely many cone singularities.Our approach not only broadens seminal solutions with two conesingularities on the Riemann sphere,as classified by Jost-Wang(Int.Math.Res.Not.,2002,(6):277-290)andLin-Wei-Ye(Invent.Math.,2012,190(1):169-207),but also advances beyond the limits of Lin-Yang-Zhong’s existencetheorems(J.Differential Geom.,2020,114(2):337-391)by introducing a new solution class.展开更多
Traditional manufacturing processes for lightweight curved profiles are often associated with lengthy procedures,high costs,low efficiency,and high energy consumption.In order to solve this problem,a new staggered ext...Traditional manufacturing processes for lightweight curved profiles are often associated with lengthy procedures,high costs,low efficiency,and high energy consumption.In order to solve this problem,a new staggered extrusion(SE)process was used to form the curved profile of AZ31 magnesium alloy in this paper.The study investigates the mapping relationship between the curvature,microstructure,and mechanical properties of the formed profiles by using different eccentricities of the die.Scanning electron microscopy(SEM)and electron backscatter diffraction techniques are employed to examine the effects of different eccentricity values(e)on grain morphology,recrystallization mechanisms,texture,and Schmid factors of the products.The results demonstrate that the staggered extrusion method promotes the deep refinement of grain size in the extruded products,with an average grain size of only 15%of the original billet,reaching 12.28μm.The tensile strength and elongation of the curved profiles after extrusion under the eccentricity value of 10 mm,20 mm and 30 mm are significantly higher than those of the billet,with the tensile strength is increased to 250,270,235 MPa,and the engineering strain elongation increased to 10.5%,12.1%,15.9%.This indicates that staggered extrusion enables curvature control of the profiles while improving their strength.展开更多
This study presents a fragility curve to assess explosively induced damage to military vehicle tires based on shock tube experiments.To replicate lateral damage scenarios that may occur in real battlefield environment...This study presents a fragility curve to assess explosively induced damage to military vehicle tires based on shock tube experiments.To replicate lateral damage scenarios that may occur in real battlefield environments involving missile or bomb detonations,extreme overpressure conditions were generated using a shock tube.The influence of explosive charge mass on tire damage was quantitatively evaluated.Experimental results identified two critical failure thresholds:for loss of pressure,the threshold was 354 kPa peak overpressure and 3052 kPa·ms impulse;for rupture,the values were 485 kPa and 4237 kPa-ms,respectively.The same damage profile was reproduced through finite element analysis(FEA),verifying the reliability of the simulation.A Single Degree of Freedom(SDOF)model and Kingery-Bulmash(K-B)chart were employed to generate pressure-impulse data as a function of standoff distance.These data were applied to a finite element tire model using the BLAST ENHANCED keyword in LS-DYNA.The applied peak overpressures were identical to the experimental values with a 24%-27%difference in impulse.The simulation also captured recurring bead rim separation phenomenon,leading to internal pressure loss consistent with high-speed camera observations from the experiments.The resulting fragility curve clearly defines the threshold conditions for tire damage and provides a standardized damage assessment model applicable to various explosive charge masses and stand-off distances.The proposed model offers a quantitative basis for evaluating tire vulnerability,providing foundational reference data for defense applications.Specifically,the findings are expected to serve as a reliable source for weapon effects analysis and target vulnerability assessments involving wheeled military vehicles.展开更多
On-machine measurement(OMM)stands out as a pivotal technology in complex curved surface adaptive machining.However,the complex structure inherent in workpieces poses a significant challenge as the stylus orientation f...On-machine measurement(OMM)stands out as a pivotal technology in complex curved surface adaptive machining.However,the complex structure inherent in workpieces poses a significant challenge as the stylus orientation frequently shifts during the measurement process.Consequently,a substantial amount of time is allocated to calibrating pre-travel error and probe movement.Furthermore,the frequent movement of machine tools also increases the influence of machine errors.To enhance both accuracy and efficiency,an optimization strategy for the OMM process is proposed.Based on the kinematic chain of the machine tools,the relationship between the angle combination of rotary axes,the stylus orientation,and the calibration position of pre-travel error is disclosed.Additionally,an OMM efficiency optimization model for complex curved surfaces is developed.This model is solved to produce the optimal efficiency angle combinations for each to-be-measured point.Within each angle combination,the effects of positioning errors on measurement results are addressed by coordinate system offset and measurement result compensation method.Finally,the experiments on an impeller are used to demonstrate the practical utility of the proposed method.展开更多
接收者操作特性(Receiver operating characteristics,ROC)曲线下面积(Area under the ROC curve,AUC)常被用于度量分类器在整个类先验分布上的总体分类性能.原始Boosting算法优化分类精度,但在AUC度量下并非最优.提出了一种AUC优化Boos...接收者操作特性(Receiver operating characteristics,ROC)曲线下面积(Area under the ROC curve,AUC)常被用于度量分类器在整个类先验分布上的总体分类性能.原始Boosting算法优化分类精度,但在AUC度量下并非最优.提出了一种AUC优化Boosting改进算法,通过在原始Boosting迭代中引入数据重平衡操作,实现弱学习算法优化目标从精度向AUC的迁移.实验结果表明,较之原始Boosting算法,新算法在AUC度量下能获得更好性能.展开更多
准确率一直被作为分类器预测性能的主要评估标准，但是它存在着诸多的缺点和不足。本文将准确率与AUC（the area under the Receiver Operating Characteristic curve）进行了理论上的对比分析，并分别使用AUC和准确率对3种分类学习算法...准确率一直被作为分类器预测性能的主要评估标准，但是它存在着诸多的缺点和不足。本文将准确率与AUC（the area under the Receiver Operating Characteristic curve）进行了理论上的对比分析，并分别使用AUC和准确率对3种分类学习算法在15个两类数据集上进行了评估。综合理论和实验两个方面的结果，显示了AUC不但优于而且应该替代准确率，成为更好的分类器性能的评估度量。同时，用AUC对3种分类学习算法的重新评估，进一步证实了基于贝叶斯定理的Naive Bayes和TAN-CMI分类算法优于决策树分类算法C4．5。展开更多
AUC(area under the ROC curve)优化问题的损失函数由来自不同类别的样本对构成,这使得依赖于损失函数之和的目标函数与训练样本数二次相关,不能直接使用传统在线学习方法求解.当前的在线AUC优化算法聚焦于在求解过程中避免直接计算所...AUC(area under the ROC curve)优化问题的损失函数由来自不同类别的样本对构成,这使得依赖于损失函数之和的目标函数与训练样本数二次相关,不能直接使用传统在线学习方法求解.当前的在线AUC优化算法聚焦于在求解过程中避免直接计算所有的损失函数,以减小问题的规模,实现在线AUC优化.针对以上问题提出了一种AUC优化的新目标函数,该目标函数仅与训练样本数线性相关;理论分析表明:最小化该目标函数等价于最小化由L2正则化项和最小二乘损失函数组成的AUC优化的目标函数.基于新的目标函数,提出了在线AUC优化的线性方法(linear online AUC maximization,LOAM);根据不同的分类器更新策略,给出2种算法LOAMILSC和LOAMAda.实验表明:与原有方法相比,LOAMILSC算法获得了更优的AUC性能,而对于实时或高维学习任务,LOAMAda算法更加高效.展开更多
文摘鉴于ROC曲线下面积(Area Under the ROC Curve,AUC)对数据分布的不敏感特性,面向AUC的对抗训练(AdAUC)近来已成为机器学习领域中抵御长尾分布下对抗攻击的有效范式之一。当前主流方法大多遵循基于平方替代损失的AUC对抗训练框架,并将成对比较形式的AUC对抗损失重构为一个逐样本的随机鞍点优化问题,克服端到端的计算瓶颈。然而,面向复杂的实际应用场景,基于平方损失设计的AUC对抗训练框架恐难以适应多样的下游任务需求。此外,与传统对抗训练范式类似,面向AUC的对抗训练方法在提高模型对抗鲁棒性的同时,也会降低模型在正常样本上的AUC性能,而目前鲜有针对该问题的有效解决方案。鉴于此,本文对如何构建一般化的高效AUC对抗机器学习范式展开系统研究。首先,提出了一种基于标准化分数扰动的通用AUC对抗训练框架(NSAdAUC),在相对温和的条件下,该框架可通过直接扰动模型对样本的预测得分实现对AUC指标的攻击,且不依赖于特定的AUC替代损失。在此基础上,本文进一步指出鲁棒AUC误差可分解为标准AUC误差和边界AUC误差两项之和,并据此设计了一种基于排序感知对抗正则化的AUC对抗训练框架(RARAdAUC),同时兼顾模型的标准AUC和鲁棒AUC性能。为验证所提框架的有效性,在5个长尾基准数据集上进行了大量实验,结果表明所提NSAdAUC和RARAdAUC框架在多种对抗攻击下的鲁棒性均优于现有方法,可在平均意义上分别产生0.94%、5.52%的标准AUC和5.69%、5.41%的鲁棒AUC性能提升。
基金supported by the National Social Science Fund of China(Grand No.21XTJ001).
文摘A Receiver Operating Characteristic(ROC)analysis of a power is important and useful in clinical trials.A Classical Conditional Power(CCP)is a probability of a classical rejection region given values of true treatment effect and interim result.For hypotheses and reversed hypotheses under normal models,we obtain analytical expressions of the ROC curves of the CCP,find optimal ROC curves of the CCP,investigate the superiority of the ROC curves of the CCP,calculate critical values of the False Positive Rate(FPR),True Positive Rate(TPR),and cutoff of the optimal CCP,and give go/no go decisions at the interim of the optimal CCP.In addition,extensive numerical experiments are carried out to exemplify our theoretical results.Finally,a real data example is performed to illustrate the go/no go decisions of the optimal CCP.
基金Project(4013311)supported by the National Science Foundation of Iran(INSF)。
文摘This article examines the influence of annealing temperature on fracture toughness and forming limit curves of dissimilar aluminum/silver sheets.In the cold roll bonding process,after brushing and acid washing,the prepared surfaces are placed on top of each other and by rolling with reduction more than 50%,the bonding between layers is established.In this research,the roll bonding process was done at room temperature,without the use of lubricants and with a 70%thickness reduction.Then,the final thickness of the Ag/Al bilayer sheet reached 350μm by several stages of cold rolling.Before cold rolling,it should be noted that to decrease the hardness created due to plastic deformation,the roll-bonded samples were subjected to annealing heat treatment at 400℃for 90 min.Thus,the final samples were annealed at 200,300 and 400℃for 90 min and cooled in a furnace to examine the annealing temperature effects.The uniaxial tensile and microhardness tests measured mechanical properties.Also,to investigate the fracture mechanism,the fractography of the cross-section was examined by scanning electron microscope(SEM).To evaluate the formability of Ag/Al bilayer sheets,forming limit curves were obtained experimentally through the Nakazima test.The resistance of composites to failure due to cracking was also investigated by fracture toughness.The results showed that annealing increases the elongation and formability of the Ag/Al bilayer sheet while reduces the ultimate tensile strength and fracture toughness.However,the changing trend is not the same at different temperatures,and according to the results,the most significant effect is obtained at 300℃and aluminum layers.It was also determined that by increasing annealing temperature,the fracture mechanism from shear ductile with small and shallow dimples becomes ductile with deep cavities.
文摘Objective Primary liver cancer,predominantly hepatocellular carcinoma(HCC),is a significant global health issue,ranking as the sixth most diagnosed cancer and the third leading cause of cancer-related mortality.Accurate and early diagnosis of HCC is crucial for effective treatment,as HCC and non-HCC malignancies like intrahepatic cholangiocarcinoma(ICC)exhibit different prognoses and treatment responses.Traditional diagnostic methods,including liver biopsy and contrast-enhanced ultrasound(CEUS),face limitations in applicability and objectivity.The primary objective of this study was to develop an advanced,lightweighted classification network capable of distinguishing HCC from other non-HCC malignancies by leveraging the automatic analysis of brightness changes in CEUS images.The ultimate goal was to create a user-friendly and cost-efficient computer-aided diagnostic tool that could assist radiologists in making more accurate and efficient clinical decisions.Methods This retrospective study encompassed a total of 161 patients,comprising 131 diagnosed with HCC and 30 with non-HCC malignancies.To achieve accurate tumor detection,the YOLOX network was employed to identify the region of interest(ROI)on both B-mode ultrasound and CEUS images.A custom-developed algorithm was then utilized to extract brightness change curves from the tumor and adjacent liver parenchyma regions within the CEUS images.These curves provided critical data for the subsequent analysis and classification process.To analyze the extracted brightness change curves and classify the malignancies,we developed and compared several models.These included one-dimensional convolutional neural networks(1D-ResNet,1D-ConvNeXt,and 1D-CNN),as well as traditional machine-learning methods such as support vector machine(SVM),ensemble learning(EL),k-nearest neighbor(KNN),and decision tree(DT).The diagnostic performance of each method in distinguishing HCC from non-HCC malignancies was rigorously evaluated using four key metrics:area under the receiver operating characteristic(AUC),accuracy(ACC),sensitivity(SE),and specificity(SP).Results The evaluation of the machine-learning methods revealed AUC values of 0.70 for SVM,0.56 for ensemble learning,0.63 for KNN,and 0.72 for the decision tree.These results indicated moderate to fair performance in classifying the malignancies based on the brightness change curves.In contrast,the deep learning models demonstrated significantly higher AUCs,with 1D-ResNet achieving an AUC of 0.72,1D-ConvNeXt reaching 0.82,and 1D-CNN obtaining the highest AUC of 0.84.Moreover,under the five-fold cross-validation scheme,the 1D-CNN model outperformed other models in both accuracy and specificity.Specifically,it achieved accuracy improvements of 3.8%to 10.0%and specificity enhancements of 6.6%to 43.3%over competing approaches.The superior performance of the 1D-CNN model highlighted its potential as a powerful tool for accurate classification.Conclusion The 1D-CNN model proved to be the most effective in differentiating HCC from non-HCC malignancies,surpassing both traditional machine-learning methods and other deep learning models.This study successfully developed a user-friendly and cost-efficient computer-aided diagnostic solution that would significantly enhances radiologists’diagnostic capabilities.By improving the accuracy and efficiency of clinical decision-making,this tool has the potential to positively impact patient care and outcomes.Future work may focus on further refining the model and exploring its integration with multimodal ultrasound data to maximize its accuracy and applicability.
基金supported by the National Natural Science Foundation of China(11931009,12271495,11971450,and 12071449)Anhui Initiative in Quantum Information Technologies(AHY150200)the Project of Stable Support for Youth Team in Basic Research Field,Chinese Academy of Sciences(YSBR-001).
文摘On a compact Riemann surface with finite punctures P_(1),…P_(k),we define toric curves as multivalued,totallyunramified holomorphic maps to P^(n)with monodromy in a maximal torus of PSU(n+1).Toric solutions to SU(n+1)Todasystems on X\{P_(1);…;P_(k)}are recognized by the associated toric curves in.We introduce character n-ensembles as-tuples of meromorphic one-forms with simple poles and purely imaginary periods,generating toric curves on minus finitelymany points.On X,we establish a correspondence between character-ensembles and toric solutions to the SU(n+1)system with finitely many cone singularities.Our approach not only broadens seminal solutions with two conesingularities on the Riemann sphere,as classified by Jost-Wang(Int.Math.Res.Not.,2002,(6):277-290)andLin-Wei-Ye(Invent.Math.,2012,190(1):169-207),but also advances beyond the limits of Lin-Yang-Zhong’s existencetheorems(J.Differential Geom.,2020,114(2):337-391)by introducing a new solution class.
基金Project(JQ2022E004)supported by the Natural Science Foundation of Heilongjiang Province,China。
文摘Traditional manufacturing processes for lightweight curved profiles are often associated with lengthy procedures,high costs,low efficiency,and high energy consumption.In order to solve this problem,a new staggered extrusion(SE)process was used to form the curved profile of AZ31 magnesium alloy in this paper.The study investigates the mapping relationship between the curvature,microstructure,and mechanical properties of the formed profiles by using different eccentricities of the die.Scanning electron microscopy(SEM)and electron backscatter diffraction techniques are employed to examine the effects of different eccentricity values(e)on grain morphology,recrystallization mechanisms,texture,and Schmid factors of the products.The results demonstrate that the staggered extrusion method promotes the deep refinement of grain size in the extruded products,with an average grain size of only 15%of the original billet,reaching 12.28μm.The tensile strength and elongation of the curved profiles after extrusion under the eccentricity value of 10 mm,20 mm and 30 mm are significantly higher than those of the billet,with the tensile strength is increased to 250,270,235 MPa,and the engineering strain elongation increased to 10.5%,12.1%,15.9%.This indicates that staggered extrusion enables curvature control of the profiles while improving their strength.
基金part of the Agency for Defense Development(ADD)research project on Weapon lethality/effectiveness analysis technology for material targets and grant funded by the korean goverment(511225-912A03301)。
文摘This study presents a fragility curve to assess explosively induced damage to military vehicle tires based on shock tube experiments.To replicate lateral damage scenarios that may occur in real battlefield environments involving missile or bomb detonations,extreme overpressure conditions were generated using a shock tube.The influence of explosive charge mass on tire damage was quantitatively evaluated.Experimental results identified two critical failure thresholds:for loss of pressure,the threshold was 354 kPa peak overpressure and 3052 kPa·ms impulse;for rupture,the values were 485 kPa and 4237 kPa-ms,respectively.The same damage profile was reproduced through finite element analysis(FEA),verifying the reliability of the simulation.A Single Degree of Freedom(SDOF)model and Kingery-Bulmash(K-B)chart were employed to generate pressure-impulse data as a function of standoff distance.These data were applied to a finite element tire model using the BLAST ENHANCED keyword in LS-DYNA.The applied peak overpressures were identical to the experimental values with a 24%-27%difference in impulse.The simulation also captured recurring bead rim separation phenomenon,leading to internal pressure loss consistent with high-speed camera observations from the experiments.The resulting fragility curve clearly defines the threshold conditions for tire damage and provides a standardized damage assessment model applicable to various explosive charge masses and stand-off distances.The proposed model offers a quantitative basis for evaluating tire vulnerability,providing foundational reference data for defense applications.Specifically,the findings are expected to serve as a reliable source for weapon effects analysis and target vulnerability assessments involving wheeled military vehicles.
基金Projects(51775445,52175435)supported by the National Natural Science Foundation of ChinaProject(CX2023051)supported by the Innovation Foundation for Doctor Dissertation of Northwestern Polytechnical University,China。
文摘On-machine measurement(OMM)stands out as a pivotal technology in complex curved surface adaptive machining.However,the complex structure inherent in workpieces poses a significant challenge as the stylus orientation frequently shifts during the measurement process.Consequently,a substantial amount of time is allocated to calibrating pre-travel error and probe movement.Furthermore,the frequent movement of machine tools also increases the influence of machine errors.To enhance both accuracy and efficiency,an optimization strategy for the OMM process is proposed.Based on the kinematic chain of the machine tools,the relationship between the angle combination of rotary axes,the stylus orientation,and the calibration position of pre-travel error is disclosed.Additionally,an OMM efficiency optimization model for complex curved surfaces is developed.This model is solved to produce the optimal efficiency angle combinations for each to-be-measured point.Within each angle combination,the effects of positioning errors on measurement results are addressed by coordinate system offset and measurement result compensation method.Finally,the experiments on an impeller are used to demonstrate the practical utility of the proposed method.
文摘接收者操作特性(Receiver operating characteristics,ROC)曲线下面积(Area under the ROC curve,AUC)常被用于度量分类器在整个类先验分布上的总体分类性能.原始Boosting算法优化分类精度,但在AUC度量下并非最优.提出了一种AUC优化Boosting改进算法,通过在原始Boosting迭代中引入数据重平衡操作,实现弱学习算法优化目标从精度向AUC的迁移.实验结果表明,较之原始Boosting算法,新算法在AUC度量下能获得更好性能.
文摘准确率一直被作为分类器预测性能的主要评估标准，但是它存在着诸多的缺点和不足。本文将准确率与AUC（the area under the Receiver Operating Characteristic curve）进行了理论上的对比分析，并分别使用AUC和准确率对3种分类学习算法在15个两类数据集上进行了评估。综合理论和实验两个方面的结果，显示了AUC不但优于而且应该替代准确率，成为更好的分类器性能的评估度量。同时，用AUC对3种分类学习算法的重新评估，进一步证实了基于贝叶斯定理的Naive Bayes和TAN-CMI分类算法优于决策树分类算法C4．5。
文摘AUC(area under the ROC curve)优化问题的损失函数由来自不同类别的样本对构成,这使得依赖于损失函数之和的目标函数与训练样本数二次相关,不能直接使用传统在线学习方法求解.当前的在线AUC优化算法聚焦于在求解过程中避免直接计算所有的损失函数,以减小问题的规模,实现在线AUC优化.针对以上问题提出了一种AUC优化的新目标函数,该目标函数仅与训练样本数线性相关;理论分析表明:最小化该目标函数等价于最小化由L2正则化项和最小二乘损失函数组成的AUC优化的目标函数.基于新的目标函数,提出了在线AUC优化的线性方法(linear online AUC maximization,LOAM);根据不同的分类器更新策略,给出2种算法LOAMILSC和LOAMAda.实验表明:与原有方法相比,LOAMILSC算法获得了更优的AUC性能,而对于实时或高维学习任务,LOAMAda算法更加高效.
