The attention mechanism is one of the most important priori knowledge to enhance convolutional neural networks.Most attention mechanisms are bound to the convolutional layer and use local or global contextual informat...The attention mechanism is one of the most important priori knowledge to enhance convolutional neural networks.Most attention mechanisms are bound to the convolutional layer and use local or global contextual information to recalibrate the input.This is a popular attention strategy design method.Global contextual information helps the network to consider the overall distribution,while local contextual information is more general.The contextual information makes the network pay attention to the mean or maximum value of a particular receptive field.Different from the most attention mechanism,this article proposes a novel attention mechanism with the heuristic difference attention module(HDAM).HDAM’s input recalibration is based on the difference between the local and global contextual information instead of the mean and maximum values.At the same time,to make different layers have amore suitable local receptive field sizes and increase the flexibility of the local receptive field design,we use genetic algorithm to heuristically produce local receptive fields.First,HDAM extracts the mean value of the global and local receptive fields as the corresponding contextual information.Then the difference between the global and local contextual information is calculated.Finally,HDAM uses this difference to recalibrate the input.In addition,we use the heuristic ability of genetic algorithm to search for the local receptive field size of each layer.Our experiments on CIFAR-10 and CIFAR-100 show that HDAM can use fewer parameters than other attention mechanisms to achieve higher accuracy.We implement HDAM with the Python library,Pytorch,and the code and models will be publicly available.展开更多
The xanthid genus Medaeops Guinot,1967 currently contains seven species reported from the Indo-West Pacific region.However,only one species,Medaeops granulosus(Haswell,1882),has been recorded from China seas and was o...The xanthid genus Medaeops Guinot,1967 currently contains seven species reported from the Indo-West Pacific region.However,only one species,Medaeops granulosus(Haswell,1882),has been recorded from China seas and was once thought to be widespread in the Indo-West Pacific region.In this study,two species of the Medaeops granulosus species-group were identified during the analysis on Medaeops crabs from China seas.Medaeops japonicus(Rathbun,1898),which had previously been considered a junior synonym of M.granulosus,is shown to be a distinct species that can be distinguished from M.granulosus by the lower first anterolateral tooth,the larger and concentrated granules on carapace dorsally,and the different shape of the male’s first gonopod.In addition,a new recorded species,Medaeops edwardsi Guinot,1967,was reported from China seas for the first time.A molecular analysis was conducted to determine the status of each species,and an updated key for species of the genus Medaeops was provided.展开更多
Five new records of the family Xanthidae are reported from China:Hypocolpus haanii Rathbun,1909,Neoliomera variolosa(A.Milne-Edwards,1873),Leptodius australis Ward,1936,L.davaoensis Ward,1941,and Xanthias joanneae Men...Five new records of the family Xanthidae are reported from China:Hypocolpus haanii Rathbun,1909,Neoliomera variolosa(A.Milne-Edwards,1873),Leptodius australis Ward,1936,L.davaoensis Ward,1941,and Xanthias joanneae Mendoza,2013.The specimens were collected from Hainan Island and deposited in the Marine Biological Museum,Chinese Academy of Sciences.Diagnosis based on observation,photographs,line drawings,and some taxonomic discussions are provided.展开更多
目的比较机器人无辅助切口NOSES组(R-NOSES)和机器人常规辅助切口组(R-LA)在结直肠癌治疗中的近期疗效差异。方法纳入203例于哈尔滨医科大学附属第二医院肿瘤中心行机器人辅助下结直肠癌根治术的患者的病例资料,包括患者的术前一般资料...目的比较机器人无辅助切口NOSES组(R-NOSES)和机器人常规辅助切口组(R-LA)在结直肠癌治疗中的近期疗效差异。方法纳入203例于哈尔滨医科大学附属第二医院肿瘤中心行机器人辅助下结直肠癌根治术的患者的病例资料,包括患者的术前一般资料、手术相关资料、病理资料以及术后相关资料。对比分析R-NOSES和R-LA的近期疗效差异。结果本研究共纳入R-NOSES患者36例,R-LA患者167例。通过倾向性评分后,R-NOSES患者36例,R-LA患者25例,R-NOSES组的平均手术时间长于R-LA组(267.64±57.23 min vs. 239.80±33.68 min,t=-2.181,P=0.033)。R-NOSES组的平均术后住院时间(t=2.860,P=0.006)、平均术后进食流质食物时间(t=2.929,P=0.005)和平均术后首次排气时间(t=3.046,P=0.003)均短于R-LA组,R-NOSES组并发症比例较R-LA组少(χ^(2)=4.164,P=0.041),术后第一日疼痛评分低于R-LA组(t=2.994,P=0.005)。R-NOSES组术后躯体功能(t=-7.530,P<0.001)、角色功能(t=-6.359,P<0.001)、情绪功能(t=-4.812,P<0.001)恢复更好,整体健康情况优于R-LA组(t=-4.288,P<0.001)。结论与R-LA组相比,R-NOSES组虽然平均手术时间略有延长,但术后首次排气时间、术后首次进流食时间、术后住院时间均更短,术后并发症少,功能恢复更快,术后疼痛更少。展开更多
基金partially supported by the National Natural Science Foundation of China(61876089,61403206,61876185,61902281)the Opening Project of Jiangsu Key Laboratory of Data Science and Smart Software(No.2019DS302)+4 种基金the Natural Science Foundation of Jiangsu Province(BK20141005)the Natural Science Foundation of the Jiangsu Higher Education Institutions of China(14KJB520025)the Science and technology program of Ministry of Housing and Urban-Rural Development(2019-K-141)the Entrepreneurial team of sponge City(2017R02002)the Priority Academic Program Development of Jiangsu Higher Education Institutions.
文摘The attention mechanism is one of the most important priori knowledge to enhance convolutional neural networks.Most attention mechanisms are bound to the convolutional layer and use local or global contextual information to recalibrate the input.This is a popular attention strategy design method.Global contextual information helps the network to consider the overall distribution,while local contextual information is more general.The contextual information makes the network pay attention to the mean or maximum value of a particular receptive field.Different from the most attention mechanism,this article proposes a novel attention mechanism with the heuristic difference attention module(HDAM).HDAM’s input recalibration is based on the difference between the local and global contextual information instead of the mean and maximum values.At the same time,to make different layers have amore suitable local receptive field sizes and increase the flexibility of the local receptive field design,we use genetic algorithm to heuristically produce local receptive fields.First,HDAM extracts the mean value of the global and local receptive fields as the corresponding contextual information.Then the difference between the global and local contextual information is calculated.Finally,HDAM uses this difference to recalibrate the input.In addition,we use the heuristic ability of genetic algorithm to search for the local receptive field size of each layer.Our experiments on CIFAR-10 and CIFAR-100 show that HDAM can use fewer parameters than other attention mechanisms to achieve higher accuracy.We implement HDAM with the Python library,Pytorch,and the code and models will be publicly available.
基金Supported by the National Natural Science Foundation of China(No.42176138)the Ministry of Science and Technology of China(No.2021YFF0502801)+1 种基金the Strategic Priority Research Program of Chinese Academy of Sciences(No.XDB42000000)the National Key R&D Program of China(No.2018YFD0900804)。
文摘The xanthid genus Medaeops Guinot,1967 currently contains seven species reported from the Indo-West Pacific region.However,only one species,Medaeops granulosus(Haswell,1882),has been recorded from China seas and was once thought to be widespread in the Indo-West Pacific region.In this study,two species of the Medaeops granulosus species-group were identified during the analysis on Medaeops crabs from China seas.Medaeops japonicus(Rathbun,1898),which had previously been considered a junior synonym of M.granulosus,is shown to be a distinct species that can be distinguished from M.granulosus by the lower first anterolateral tooth,the larger and concentrated granules on carapace dorsally,and the different shape of the male’s first gonopod.In addition,a new recorded species,Medaeops edwardsi Guinot,1967,was reported from China seas for the first time.A molecular analysis was conducted to determine the status of each species,and an updated key for species of the genus Medaeops was provided.
基金Supported by the National Key R&D Program of China(No.2018YFD0900804)the National Science&Technology Fundamental Resources Investigation Program of China(No.2018FY10010006)。
文摘Five new records of the family Xanthidae are reported from China:Hypocolpus haanii Rathbun,1909,Neoliomera variolosa(A.Milne-Edwards,1873),Leptodius australis Ward,1936,L.davaoensis Ward,1941,and Xanthias joanneae Mendoza,2013.The specimens were collected from Hainan Island and deposited in the Marine Biological Museum,Chinese Academy of Sciences.Diagnosis based on observation,photographs,line drawings,and some taxonomic discussions are provided.
文摘目的比较机器人无辅助切口NOSES组(R-NOSES)和机器人常规辅助切口组(R-LA)在结直肠癌治疗中的近期疗效差异。方法纳入203例于哈尔滨医科大学附属第二医院肿瘤中心行机器人辅助下结直肠癌根治术的患者的病例资料,包括患者的术前一般资料、手术相关资料、病理资料以及术后相关资料。对比分析R-NOSES和R-LA的近期疗效差异。结果本研究共纳入R-NOSES患者36例,R-LA患者167例。通过倾向性评分后,R-NOSES患者36例,R-LA患者25例,R-NOSES组的平均手术时间长于R-LA组(267.64±57.23 min vs. 239.80±33.68 min,t=-2.181,P=0.033)。R-NOSES组的平均术后住院时间(t=2.860,P=0.006)、平均术后进食流质食物时间(t=2.929,P=0.005)和平均术后首次排气时间(t=3.046,P=0.003)均短于R-LA组,R-NOSES组并发症比例较R-LA组少(χ^(2)=4.164,P=0.041),术后第一日疼痛评分低于R-LA组(t=2.994,P=0.005)。R-NOSES组术后躯体功能(t=-7.530,P<0.001)、角色功能(t=-6.359,P<0.001)、情绪功能(t=-4.812,P<0.001)恢复更好,整体健康情况优于R-LA组(t=-4.288,P<0.001)。结论与R-LA组相比,R-NOSES组虽然平均手术时间略有延长,但术后首次排气时间、术后首次进流食时间、术后住院时间均更短,术后并发症少,功能恢复更快,术后疼痛更少。