Objective:The perioperative period of major orthopedic surgery is associated with a high risk of thrombosis,but the best chemopreventive agent for thrombosis prophylaxis is still inconclusive.For this reason,this pape...Objective:The perioperative period of major orthopedic surgery is associated with a high risk of thrombosis,but the best chemopreventive agent for thrombosis prophylaxis is still inconclusive.For this reason,this paper evaluated the efficacy and safety of aspirin versus low-molecular heparin using a Meta-analysis.Methods:Ten randomized controlled studies on the application of aspirin and low-molecular heparin for the prevention of deep vein thrombosis in orthopedic major surgery were retrieved by computer searches of PubMed,CochraneLibrary,WebofScience,China Knowledge Network,Wanfang,and Vipul databases according to inclusion and exclusion criteria,and the literature was managed using Endnote software,and the data were analyzed using Revman 5.3 software was used to perform Meta-analysis of the extracted data,focusing on the effects of these two drugs on pulmonary embolism,deep vein thrombosis,major bleeding events,minor bleeding events,wound complications,mortality and blood loss within 90 days after major orthopedic surgery.Results:(1)Ten randomized controlled trials of high quality were included,with a total of 12,974 patients,7,026 in the aspirin group and 5,948 in the low-molecular heparin group;(2)Meta-analysis showed that aspirin had a higher incidence of pulmonary embolism(OR=1.59,95%CI:1.02 to 2.49,P=0.04)and deep vein thrombosis(OR=1.60,95%CI:1.26 to 2.02,P=0.0001)than low molecular heparin;(3)The incidence of major bleeding events(OR=0.85,95%CI:0.47 to 1.55,P=0.60),minor bleeding events(OR=0.79,95%CI:0.55 to 1.12,P=0.18),adverse wound reactions(OR=0.79,95%CI:0.48 to 1.31,P=0.36),mortality within 90 days(OR=0.69,95%CI:0.20 to 2.31,P=0.55)and perioperative blood loss(MD=0.69,95%CI:0.20 to 2.31,P=0.55)in both drug groups,mortality within 90 days(OR=0.69,95%CI:0.20 to 2.31,P=0.55)and perioperative blood loss(MD=0.69,95%CI:0.20 to 2.31,P=0.55)were not statistically significant.Conclusion:Low-molecular heparin was superior to aspirin in the prevention of pulmonary embolism and lower extremity deep vein thrombosis after major orthopedic surgery,but the safety and adverse drug reactions of both groups were basically similar.Based on this,the authors recommend that low-molecular heparin should be preferred for the prevention of deep vein thrombosis in major orthopaedic surgery;however,the inclusion of randomized controlled trials remains limited,necessitating high-quality,large-sample,long-term follow-up clinical studies.展开更多
密码专用处理器常采用分簇式超长指令字(Very Long Instruction Word,VLIW)架构,其性能的发挥依赖于编译器的实现.当前对于通用VLIW架构的编译后端优化方案,在密码专用处理器上都有一定的不适应性.为此,本文提出了一种面向密码专用处理...密码专用处理器常采用分簇式超长指令字(Very Long Instruction Word,VLIW)架构,其性能的发挥依赖于编译器的实现.当前对于通用VLIW架构的编译后端优化方案,在密码专用处理器上都有一定的不适应性.为此,本文提出了一种面向密码专用处理器的、同时进行簇指派、指令调度和寄存器分配的编译器后端优化方法.构造“定值-引用”链,求解变量的候选寄存器类型集合交集,确定其寄存器类型;实时评估可用资源,进行基于优先级的指令选择和基于平衡寄存器压力的簇指派;改进线性扫描算法,基于变量的“待引用次数”列表进行实时的寄存器分配.实验结果表明,本方法能够提升生成代码的性能,且算法是非启发式的,减小了编译所需的时间.展开更多
高光谱遥感图像虽然具有较高的光谱分辨率,但只能提供二维光谱信息,而激光雷达(Light Detection and Ranging, LiDAR)可以提供可靠的三维数据和森林的冠层特征,二者结合能够优势互补,协同提高地物分类能力。基于此,本文提出了一种利用...高光谱遥感图像虽然具有较高的光谱分辨率,但只能提供二维光谱信息,而激光雷达(Light Detection and Ranging, LiDAR)可以提供可靠的三维数据和森林的冠层特征,二者结合能够优势互补,协同提高地物分类能力。基于此,本文提出了一种利用高光谱和机载激光雷达数据进行树种识别的方法,探讨了二者协同对树种精细分类的影响,最后通过公开数据的实验来验证方法的有效性。结果表明,高光谱和激光雷达数据结合后,树种分类精度得到了显著提高。通过本文提出的方法,即使使用少量特征也能大大提高分类精度。整体分类准确率最高达到93.21%,比单个高光谱数据的分类准确率提高9.31%。展开更多
Quantized training has been proven to be a prominent method to achieve deep neural network training under limited computational resources.It uses low bit-width arithmetics with a proper scaling factor to achieve negli...Quantized training has been proven to be a prominent method to achieve deep neural network training under limited computational resources.It uses low bit-width arithmetics with a proper scaling factor to achieve negligible accuracy loss.Cambricon-Q is the ASIC design proposed to efficiently support quantized training,and achieves significant performance improvement.However,there are still two caveats in the design.First,Cambricon-Q with different hardware specifications may lead to different numerical errors,resulting in non-reproducible behaviors which may become a major concern in critical applications.Second,Cambricon-Q cannot leverage data sparsity,where considerable cycles could still be squeezed out.To address the caveats,the acceleration core of Cambricon-Q is redesigned to support fine-grained irregular data processing.The new design not only enables acceleration on sparse data,but also enables performing local dynamic quantization by contiguous value ranges(which is hardware independent),instead of contiguous addresses(which is dependent on hardware factors).Experimental results show that the accuracy loss of the method still keeps negligible,and the accelerator achieves 1.61×performance improvement over Cambricon-Q,with about 10%energy increase.展开更多
基金Natural Science Foundation of Xinjiang Uygur Autonomous Region(No.2019D01C244)Special Research Project for Health and Young Medical Science and Technology Talents in Xinjiang Uygur Autonomous Region(No.WJWY-202142)。
文摘Objective:The perioperative period of major orthopedic surgery is associated with a high risk of thrombosis,but the best chemopreventive agent for thrombosis prophylaxis is still inconclusive.For this reason,this paper evaluated the efficacy and safety of aspirin versus low-molecular heparin using a Meta-analysis.Methods:Ten randomized controlled studies on the application of aspirin and low-molecular heparin for the prevention of deep vein thrombosis in orthopedic major surgery were retrieved by computer searches of PubMed,CochraneLibrary,WebofScience,China Knowledge Network,Wanfang,and Vipul databases according to inclusion and exclusion criteria,and the literature was managed using Endnote software,and the data were analyzed using Revman 5.3 software was used to perform Meta-analysis of the extracted data,focusing on the effects of these two drugs on pulmonary embolism,deep vein thrombosis,major bleeding events,minor bleeding events,wound complications,mortality and blood loss within 90 days after major orthopedic surgery.Results:(1)Ten randomized controlled trials of high quality were included,with a total of 12,974 patients,7,026 in the aspirin group and 5,948 in the low-molecular heparin group;(2)Meta-analysis showed that aspirin had a higher incidence of pulmonary embolism(OR=1.59,95%CI:1.02 to 2.49,P=0.04)and deep vein thrombosis(OR=1.60,95%CI:1.26 to 2.02,P=0.0001)than low molecular heparin;(3)The incidence of major bleeding events(OR=0.85,95%CI:0.47 to 1.55,P=0.60),minor bleeding events(OR=0.79,95%CI:0.55 to 1.12,P=0.18),adverse wound reactions(OR=0.79,95%CI:0.48 to 1.31,P=0.36),mortality within 90 days(OR=0.69,95%CI:0.20 to 2.31,P=0.55)and perioperative blood loss(MD=0.69,95%CI:0.20 to 2.31,P=0.55)in both drug groups,mortality within 90 days(OR=0.69,95%CI:0.20 to 2.31,P=0.55)and perioperative blood loss(MD=0.69,95%CI:0.20 to 2.31,P=0.55)were not statistically significant.Conclusion:Low-molecular heparin was superior to aspirin in the prevention of pulmonary embolism and lower extremity deep vein thrombosis after major orthopedic surgery,but the safety and adverse drug reactions of both groups were basically similar.Based on this,the authors recommend that low-molecular heparin should be preferred for the prevention of deep vein thrombosis in major orthopaedic surgery;however,the inclusion of randomized controlled trials remains limited,necessitating high-quality,large-sample,long-term follow-up clinical studies.
文摘密码专用处理器常采用分簇式超长指令字(Very Long Instruction Word,VLIW)架构,其性能的发挥依赖于编译器的实现.当前对于通用VLIW架构的编译后端优化方案,在密码专用处理器上都有一定的不适应性.为此,本文提出了一种面向密码专用处理器的、同时进行簇指派、指令调度和寄存器分配的编译器后端优化方法.构造“定值-引用”链,求解变量的候选寄存器类型集合交集,确定其寄存器类型;实时评估可用资源,进行基于优先级的指令选择和基于平衡寄存器压力的簇指派;改进线性扫描算法,基于变量的“待引用次数”列表进行实时的寄存器分配.实验结果表明,本方法能够提升生成代码的性能,且算法是非启发式的,减小了编译所需的时间.
文摘高光谱遥感图像虽然具有较高的光谱分辨率,但只能提供二维光谱信息,而激光雷达(Light Detection and Ranging, LiDAR)可以提供可靠的三维数据和森林的冠层特征,二者结合能够优势互补,协同提高地物分类能力。基于此,本文提出了一种利用高光谱和机载激光雷达数据进行树种识别的方法,探讨了二者协同对树种精细分类的影响,最后通过公开数据的实验来验证方法的有效性。结果表明,高光谱和激光雷达数据结合后,树种分类精度得到了显著提高。通过本文提出的方法,即使使用少量特征也能大大提高分类精度。整体分类准确率最高达到93.21%,比单个高光谱数据的分类准确率提高9.31%。
基金the National Key Research and Devecopment Program of China(No.2022YFB4501601)the National Natural Science Foundation of China(No.62102398,U20A20227,62222214,62002338,U22A2028,U19B2019)+1 种基金the Chinese Academy of Sciences Project for Young Scientists in Basic Research(YSBR-029)Youth Innovation Promotion Association Chinese Academy of Sciences。
文摘Quantized training has been proven to be a prominent method to achieve deep neural network training under limited computational resources.It uses low bit-width arithmetics with a proper scaling factor to achieve negligible accuracy loss.Cambricon-Q is the ASIC design proposed to efficiently support quantized training,and achieves significant performance improvement.However,there are still two caveats in the design.First,Cambricon-Q with different hardware specifications may lead to different numerical errors,resulting in non-reproducible behaviors which may become a major concern in critical applications.Second,Cambricon-Q cannot leverage data sparsity,where considerable cycles could still be squeezed out.To address the caveats,the acceleration core of Cambricon-Q is redesigned to support fine-grained irregular data processing.The new design not only enables acceleration on sparse data,but also enables performing local dynamic quantization by contiguous value ranges(which is hardware independent),instead of contiguous addresses(which is dependent on hardware factors).Experimental results show that the accuracy loss of the method still keeps negligible,and the accelerator achieves 1.61×performance improvement over Cambricon-Q,with about 10%energy increase.
