期刊文献+
共找到385篇文章
< 1 2 20 >
每页显示 20 50 100
BP模型中的激励函数和改进的网络训练法 被引量:28
1
作者 向国全 董道珍 《计算机研究与发展》 EI CSCD 北大核心 1997年第2期113-117,共5页
本文研究BP算法中激励函数f对收敛速度的影响,得出了陡峭函数收敛快的结论.其次。
关键词 神经网络 BP算法 激励函数 梯度下降法
下载PDF
基于ArcReLU函数的神经网络激活函数优化研究 被引量:18
2
作者 许赟杰 徐菲菲 《数据采集与处理》 CSCD 北大核心 2019年第3期517-529,共13页
近年来深度学习发展迅猛。由于深度学习的概念源于神经网络,而激活函数更是神经网络模型在学习理解非线性函数时不可或缺的部分,因此本文对常用的激活函数进行了研究比较。针对常用的激活函数在反向传播神经网络中具有收敛速度较慢、存... 近年来深度学习发展迅猛。由于深度学习的概念源于神经网络,而激活函数更是神经网络模型在学习理解非线性函数时不可或缺的部分,因此本文对常用的激活函数进行了研究比较。针对常用的激活函数在反向传播神经网络中具有收敛速度较慢、存在局部极小或梯度消失的问题,将Sigmoid系和ReLU系激活函数进行了对比,分别讨论了其性能,详细分析了几类常用激活函数的优点及不足,并通过研究Arctan函数在神经网络中应用的可能性,结合ReLU函数,提出了一种新型的激活函数ArcReLU。实验证明,该函数既能显著加快反向传播神经网络的训练速度,又能有效降低训练误差并避免梯度消失的问题。 展开更多
关键词 神经网络 激活函数 反正切函数 ArcReLU
下载PDF
Mechanism of Kinect-based virtual reality training for motor functional recovery of upper limbs after subacute stroke 被引量:15
3
作者 Xiao Bao Yurong Mao +6 位作者 Qiang Lin Yunhai Qiu Shaozhen Chen Le Li Ryan S.Cates Shufeng Zhou Dongfeng Huang 《Neural Regeneration Research》 SCIE CAS CSCD 2013年第31期2904-2913,共10页
The Kinect-based virtual reality system for the Xbox 360 enables users to control and interact with the game console without the need to touch a game controller, and provides rehabilitation training for stroke patient... The Kinect-based virtual reality system for the Xbox 360 enables users to control and interact with the game console without the need to touch a game controller, and provides rehabilitation training for stroke patients with lower limb dysfunctions. However, the underlying mechanism remains un- clear. In this study, 18 healthy subjects and five patients after subacute stroke were included. The five patients were scanned using functional MRI prior to training, 3 weeks after training and at a 12-week follow-up, and then compared with healthy subjects. The FugI-Meyer Assessment and Wolf Motor Function Test scores of the hemiplegic upper limbs of stroke patients were significantly increased 3 weeks after training and at the 12-week follow-up. Functional MRI results showed that contralateral primary sensorimotor cortex was activated after Kinect-based virtual reality training in the stroke patients compared with the healthy subjects. Contralateral primary sensorimotor cortex, the bilateral supplementary motor area and the ipsilateral cerebellum were also activated during hand-clenching in all 18 healthy subjects. Our findings indicate that Kinect-based virtual reality training could promote the recovery of upper limb motor function in subacute stroke patients, and brain reorganization by Kinect-based virtual reality training may be linked to the contralateral sen- sorimotor cortex. 展开更多
关键词 neural regeneration neurological rehabilitation rehabilitation training neural plasticity virtual reality functional MRI stroke Kinect-based virtual reality training upper limb cerebral cortex brain activation region of interest grants-supported paper NEUROREGENERATION
下载PDF
BP网络激励函数的研究 被引量:6
4
作者 宋卫荣 翁桂荣 《苏州大学学报(自然科学版)》 CAS 2002年第4期57-61,共5页
研究BP算法中激励函数f对收敛速度的影响 ,并将其用于传感器静态误差的综合修正 ,实验证明取得了较好的效果 。
关键词 BP网络 人工神经网络 BP算法 激励函数 局部极小 梯度下降法
下载PDF
关于非线性激活函数的深度学习分类方法研究 被引量:13
5
作者 杨国亮 许楠 +1 位作者 李放 龚曼 《江西理工大学学报》 CAS 2018年第3期76-83,共8页
提出了一种非线性指数函数来作为深度神经网络的激活层,通过调节函数的参数,能够有效提高深层神经网络在分类任务中的测试精度和其他性能.在Cifar10/100数据库中使用最先进的架构进行训练,只通过更换激活函数并调节该函数中的参数能够... 提出了一种非线性指数函数来作为深度神经网络的激活层,通过调节函数的参数,能够有效提高深层神经网络在分类任务中的测试精度和其他性能.在Cifar10/100数据库中使用最先进的架构进行训练,只通过更换激活函数并调节该函数中的参数能够实现更优的结果.实验结果表明所提出的非线性函数在深度学习中能够带来更好的泛化性能. 展开更多
关键词 神经网络 激活 指数函数 深度学习
下载PDF
芪仙通络方对脑缺血大鼠内源性神经干细胞再生的影响及机制 被引量:13
6
作者 林秀慧 周春吉 +2 位作者 马珂 刘芳 刘祖贻 《中国实验方剂学杂志》 CAS CSCD 北大核心 2017年第24期141-147,共7页
目的:探讨芪仙通络方对脑缺血大鼠内源性神经干细胞再生的影响及可能机制。方法:采用线栓法复制大脑中动脉阻塞(middle cerebral artery occlusion,MCAO)局灶性脑缺血模型;将160只大鼠随机分为正常组,假手术组,模型组,吡拉西坦组(吡拉... 目的:探讨芪仙通络方对脑缺血大鼠内源性神经干细胞再生的影响及可能机制。方法:采用线栓法复制大脑中动脉阻塞(middle cerebral artery occlusion,MCAO)局灶性脑缺血模型;将160只大鼠随机分为正常组,假手术组,模型组,吡拉西坦组(吡拉西坦片灌胃),芪仙通络方组(芪仙通络方灌胃);5-溴脱氧尿嘧啶核苷(5-bromo-2-deoxy uridine,Brd U)腹腔注射标记脑内增殖细胞,造模后3,7,14,28 d,分别采用免疫荧光Brd U/神经元核抗原(neuronal nuclei antigen,Neu N),Brd U/胶质纤维酸性蛋白(glial fibrillary acidic protein,GFAP)双标法检测缺血海马区新生的神经元及星形胶质细胞,并计算其双标阳性细胞率,免疫组化法检测缺血侧脑内脑源性神经营养因子(brain-derived neurotrophic factor,BDNF)的表达。结果:术后7,14,28 d,芪仙通络方组、吡拉西坦组Brd U/Neu N,Brd U/GFAP双标阳性细胞率均高于模型组(P<0.01),且各时间点芪仙通络方组Brd U/Neu N阳性细胞均高于吡拉西坦组(P<0.05),而仅在术后7 d,芪仙通络方组Brd U/GFAP双标阳性细胞率高于吡拉西坦组(P<0.05),随后均呈下降趋势;与模型组比较,术后各时间点芪仙通络方组、吡拉西坦组缺血脑内BDNF表达均升高(P<0.01),芪仙通络方组BDNF表达于术后3 d达到高峰,之后水平虽有所下降,但仍高于吡拉西坦组(P<0.05)。结论:芪仙通络方可促进脑缺血大鼠内源性神经干细胞再生,以恢复期和后遗症期明显,而早期活化星形胶质细胞及上调BDNF的表达可能是其作用机制之一。 展开更多
关键词 脑缺血 芪仙通络方 神经干细胞再生 脑源性神经营养因子 星形胶质细胞 大鼠
原文传递
Optogenetics and its application in neural degeneration and regeneration 被引量:10
7
作者 Josue D.Ordaz Wei Wu Xiao-Ming Xu 《Neural Regeneration Research》 SCIE CAS CSCD 2017年第8期1197-1209,共13页
Neural degeneration and regeneration are important topics in neurological diseases. There are limited options for therapeutic interventions in neurological diseases that provide simultaneous spatial and temporal contr... Neural degeneration and regeneration are important topics in neurological diseases. There are limited options for therapeutic interventions in neurological diseases that provide simultaneous spatial and temporal control of neurons. This drawback increases side effects due to non-specific targeting. Optogenetics is a technology that allows precise spatial and temporal control of cells. Therefore, this technique has high potential as a therapeutic strategy for neurological diseases. Even though the application of optogenetics in understanding brain functional organization and complex behaviour states have been elaborated, reviews of its therapeutic potential especially in neurodegeneration and regeneration are still limited. This short review presents representative work in optogenetics in disease models such as spinal cord injury, multiple sclerosis, epilepsy, Alzheimer's disease and Parkinson's disease. It is aimed to provide a broader perspective on optogenetic therapeutic potential in neurodegeneration and neural regeneration. 展开更多
关键词 light-activated proteins neural plasticity spinal cord injury EPILEPSY Parkinson's disease Alzheimer'sdisease multiple sclerosis neural engineering memory retrieval neuron inhibition neuron activation neuralregeneration
下载PDF
基于激活函数四参可调的BP神经网络改进算法 被引量:11
8
作者 李恩玉 杨平先 孙兴波 《微电子学与计算机》 CSCD 北大核心 2008年第11期89-93,共5页
为改善BP神经网络的性能,以标准Sigmoidal函数为基础,提出了一种四参数可调的激活函数模型.在学习过程中,它能同时对激活函数的陡度、位置及映射范围进行调节,具有更强的非线性映射能力.并推导出其在BP神经网络中的学习算法.仿真结果显... 为改善BP神经网络的性能,以标准Sigmoidal函数为基础,提出了一种四参数可调的激活函数模型.在学习过程中,它能同时对激活函数的陡度、位置及映射范围进行调节,具有更强的非线性映射能力.并推导出其在BP神经网络中的学习算法.仿真结果显示,改进后的激活函数与传统的标准Sigmoidal函数相比,收敛速度能提高10倍以上,收敛精度误差可减小到传统误差的0.4%以下,而且可以有效地减少隐层的结点数,学习能力可得到较大的提高. 展开更多
关键词 神经网络 BP算法 激活函数 可调参数
下载PDF
激励函数可调的混沌神经网络短路负荷预测 被引量:9
9
作者 张晓星 程其云 +2 位作者 周湶 刘凡 孙才新 《高电压技术》 EI CAS CSCD 北大核心 2005年第7期51-54,共4页
为提高BP神经网络的收敛速度并避免陷入局部极小,提出了一种多参数可调激活函数的人工神经网络,推导出相应的BP学习算法并将改进的BP算法和混沌优化算法相结合,使混合算法的学习向全局最优方向进行以不至于陷入局部极小。在非线性函数... 为提高BP神经网络的收敛速度并避免陷入局部极小,提出了一种多参数可调激活函数的人工神经网络,推导出相应的BP学习算法并将改进的BP算法和混沌优化算法相结合,使混合算法的学习向全局最优方向进行以不至于陷入局部极小。在非线性函数仿真和短期负荷预测的研究中,该算法和传统BP算法的对比试验显示,改进算法的收敛速度更快、预测精度更高。 展开更多
关键词 神经网络 激励函数 短期负荷预测 混沌
下载PDF
权值初始化与激励函数调整相结合的学习算法 被引量:6
10
作者 武妍 王守觉 《计算机工程与应用》 CSCD 北大核心 2004年第30期23-25,44,共4页
提出了一种基于独立元分析(ICA)方法的权值初始化方法和动态调整S型激励函数的斜率相结合的神经网络学习算法。该方法利用ICA从输入数据中提取显著的特征信息来初始化输入层到隐含层权值。而且通过使神经网络的输出位于激励函数的活动区... 提出了一种基于独立元分析(ICA)方法的权值初始化方法和动态调整S型激励函数的斜率相结合的神经网络学习算法。该方法利用ICA从输入数据中提取显著的特征信息来初始化输入层到隐含层权值。而且通过使神经网络的输出位于激励函数的活动区域,对隐含层到输出层的权值进行初始化。在学习过程中,再对每个隐单元和输出单元的激励函数的斜率进行自动调整。最后通过计算机仿真实际的基准问题,验证了论文提出的方法的有效性。实验结果表明,所提出的方法能有效地加快多层前向神经网络的训练过程。 展开更多
关键词 前向神经网络 权值初始化 独立元分析 激励函数
下载PDF
Pretreated Oenan the Javanica extract increases anti-inflammatory cytokines, attenuates gliosis, and protects hippocampal neurons following transient global cerebral ischemia in gerbils 被引量:7
11
作者 Joon Ha Park In Hye Kim +12 位作者 Ji Hyeon Ahn YooHun Noh Sung-Su Kim Tae-Kyeong Lee Jae-Chul Lee Bich-Na Shin Tae Heung Sim Hyun Sam Lee Jeong Hwi Cho In Koo Hwang Il Jun Kang Jong Dai Kim Moo-Ho Won 《Neural Regeneration Research》 SCIE CAS CSCD 2019年第9期1536-1543,共8页
Recently,we have reported that Oenanthe javanica extract(OJE)displays strong neuroprotective effect against ischemic damage after transient global cerebral ischemia.However,neuroprotective mechanisms of OJE have not b... Recently,we have reported that Oenanthe javanica extract(OJE)displays strong neuroprotective effect against ischemic damage after transient global cerebral ischemia.However,neuroprotective mechanisms of OJE have not been fully identified.Thus,this study investigated the neuroprotection of OJE in the hippocampal CA1 area and its anti-inflammatory activity in gerbils subjected to 5 minutes of transient global cerebral ischemia.We treated the animals by intragastrical injection of OJE(100 and 200 mg/kg)once daily for 1 week prior to transient global cerebral ischemia.Neuroprotection of OJE was observed by immunohistochemistry for neuronal nuclear antigen and histofluorescence staining for Fluoro-Jade B.Immunohistochemistry of glial fibrillary acidic protein and ionized calcium-binding adapter molecule 1 was done for astrocytosis and microgliosis,respectively.To investigate the neuroprotective mechanisms of OJE,we performed immunohistochemistry of tumor necrosis factor-alpha and interleukin-2 for pro-inflammatory function and interleukin-4 and interleukin-13 for anti-inflammatory function.When we treated the animals by intragastrical administration of 200 mg/kg of OJE,hippocampal CA1 pyramidal neurons were protected from transient global cerebral ischemia and cerebral ischemia-induced gliosis was inhibited in the ischemic hippocampal CA1 area.We also found that interleukin-4 and-13 immunoreactivities were significantly increased in pyramidal neurons of the ischemic CA1 area after OJE pretreatment,and the increased immunoreactivities were sustained in the CA1 pyramidal neurons after transient global cerebral ischemia.However,OJE pretreatment did not increase interleukin-2 and tumor necrosis factor-alpha immunoreactivities in the CA1 pyramidal neurons.Our findings suggest that pretreatment with OJE can protect neurons and attenuate gliosis from transient global cerebral ischemia via increasing expressions of interleukin-4 and-13.The experimental plan of this study was reviewed and approved by the Institutional Animal C 展开更多
关键词 Oenanthe JAVANICA EXTRACT TRANSIENT global CEREBRAL ISCHEMIA hippocampus ischemic damage CEREBRAL ISCHEMIA neuroprotection glial activation pro-inflammatory CYTOKINES anti-inflammatory CYTOKINES inflammation neural regeneration
下载PDF
基于神经网络的强化学习算法实现倒立摆控制 被引量:7
12
作者 张涛 吴汉生 《计算机仿真》 CSCD 2006年第4期298-300,325,共4页
运用强化学习的方法来对连续的倒立摆系统实现平衡控制是一直以来有待解决的问题。该文将Q学习与神经网络中的BP网络、S激活函数相结合,利用神经网络的泛化性能,设计出一种新的学习控制策略,通过迭代和学习过程,不但能够解决倒立摆系统... 运用强化学习的方法来对连续的倒立摆系统实现平衡控制是一直以来有待解决的问题。该文将Q学习与神经网络中的BP网络、S激活函数相结合,利用神经网络的泛化性能,设计出一种新的学习控制策略,通过迭代和学习过程,不但能够解决倒立摆系统连续状态空间的输入问题,还成功解决了输出连续动作空间的问题。将此方法运用于连续倒立摆系统的平衡控制中,经过基于实际控制模型的Matlab软件仿真实验,结果显示了这个方法的可行性。该方法进一步提高了强化学习理论在实际控制系统中的应用价值。 展开更多
关键词 强化学习 神经网络 激活函数 泛化性能 连续动作空间
下载PDF
基于深度学习下的卷积神经网络参数学习 被引量:8
13
作者 吕响 张书玉 +3 位作者 宋英楠 王长忠 李雯琦 牛嘉瑞 《渤海大学学报(自然科学版)》 CAS 2021年第4期369-375,共7页
近年来,深度学习(deep learning)在图像目标识别、语音识别和自然语言处理等诸多方面都得到了广泛的应用.目前,深度学习的主要形式是深层神经网络,而深度卷积神经网络(Convolutional neural networks,CNN)则是其中一种经典而广泛应用的... 近年来,深度学习(deep learning)在图像目标识别、语音识别和自然语言处理等诸多方面都得到了广泛的应用.目前,深度学习的主要形式是深层神经网络,而深度卷积神经网络(Convolutional neural networks,CNN)则是其中一种经典而广泛应用的结构.随着深度学习方法在诸多领域的不断深入应用,深层卷积神经网络的学习研究在特征学习、目标分类、边框回归等方面表现出的优势已愈发突出.当前关于卷积神经网络的大多参考书目仅仅是说明了原理,却缺少中间参数迭代计算的推导过程,使得数学基础较弱的初学者理解其原理较为困难.本文详细地给出了CNN算法各层的参数的计算公式及其反向传播下的参数的计算公式的推导过程.这有助于加强对网络结构的了解和对网络参数的学习有一个深刻的认识,对以后的学习及建立模型有很大的帮助. 展开更多
关键词 深度学习 神经网络 激活函数 反向传播 参数学习
下载PDF
采用变参数激励函数的人工神经网络 被引量:6
14
作者 龙欣海 肖田元 陈晓峰 《计算机工程》 CAS CSCD 北大核心 2001年第12期71-73,共3页
隐层神经元采用相同的激励函数会限制网络的非线性表达能力,因此,提出变参数激励函数,在学习中同时调整网络权值和激励函数的参数,增强网络的表达能力。并使用遗传算法与MBP结合的学习算法训练网络。此方法具有全局收敛能力和很高的精度。
关键词 人工神经网络 遗传算法 变参数激励函数 学习算法
下载PDF
Changes in brain activation in stroke patients after mental practice and physical exercise: a functional MRI study 被引量:5
15
作者 Hua Liu Luping Song Tong Zhang 《Neural Regeneration Research》 SCIE CAS CSCD 2014年第15期1474-1484,共11页
Mental practice is a new rehabilitation method that reters to the mental rehearsal ot motor imagery content with the goal of improving motor performance. However, the relationship between activated regions and motor r... Mental practice is a new rehabilitation method that reters to the mental rehearsal ot motor imagery content with the goal of improving motor performance. However, the relationship between activated regions and motor recovery after mental practice training is not well understood. In this study, 15 patients who suffered a firstever subcortical stroke with neurological deficits affecting the right hand, but no significant cognitive impairment were recruited. 10 patients underwent mental practice combined with physical practice training, and 5 patients only underwent physical practice training. We observed brain activation regions after 4 weeks of training, and explored the correlation of activation changes with functional recovery of the affected hands. The results showed that, after 4 weeks of mental practice combined with physical training, the Fugl-Meyer assessment score for the affected right hand was significantly increased than that after 4 weeks of practice training alone. Functional MRI showed enhanced activation in the left primary somatosensory cortex, attenuated activation intensity in the right primary motor cortex, and enhanced right cerebellar activation observed during the motor imagery task using the affected right hand after mental practice training. The changes in brain cortical activity were related to functional recovery of the hand. Experimental findings indicate that cortical and cerebellar functional reorganization following mental practice contributed to the improvement of hand function. 展开更多
关键词 nerve regeneration brain activation cortical activation somatosensory cortex corticalreorganization STROKE mental practice functional recovery motor imagery neural regeneration
下载PDF
Emotion processing in Parkinson's disease: a blood oxygenation level-dependent functional magnetic resonance imaging study 被引量:6
16
作者 Mohammed Benzagmout Sa?d Boujraf +8 位作者 Badreeddine Alami Hassane Ali Amadou Halima El Hamdaoui Amine Bennani Mounir Jaafari Ismail Rammouz Mustapha Maaroufi Rabia Magoul Driss Boussaoud 《Neural Regeneration Research》 SCIE CAS CSCD 2019年第4期666-672,共7页
Parkinson's disease is a neurodegenerative disorder caused by loss of dopamine neurons in the substantia nigra pars compacta. Tremor, rigidity, and bradykinesia are the major symptoms of the disease. These motor i... Parkinson's disease is a neurodegenerative disorder caused by loss of dopamine neurons in the substantia nigra pars compacta. Tremor, rigidity, and bradykinesia are the major symptoms of the disease. These motor impairments are often accompanied by affective and emotional dysfunctions which have been largely studied over the last decade. The aim of this study was to investigate emotional processing organization in the brain of patients with Parkinson's disease and to explore whether there are differences between recognition of different types of emotions in Parkinson's disease. We examined 18 patients with Parkinson's disease(8 men, 10 women) with no history of neurological or psychiatric comorbidities. All these patients underwent identical brain blood oxygenation level-dependent functional magnetic resonance imaging for emotion evaluation. Blood oxygenation level-dependent functional magnetic resonance imaging results revealed that the occipito-temporal cortices, insula, orbitofrontal cortex, basal ganglia, and parietal cortex which are involved in emotion processing, were activated during the functional control. Additionally, positive emotions activate larger volumes of the same anatomical entities than neutral and negative emotions. Results also revealed that Parkinson's disease associated with emotional disorders are increasingly recognized as disabling as classic motor symptoms. These findings help clinical physicians to recognize the emotional dysfunction of patients with Parkinson's disease. 展开更多
关键词 Parkinson’s disease EMOTION processing BLOOD OXYGENATION level-dependent functional magnetic resonance imaging brain activation DOPAMINE neural REGENERATION
下载PDF
Uncertainty-Aware Physical Simulation of Neural Radiance Fields for Fluids
17
作者 Haojie Lian Jiaqi Wang +4 位作者 Leilei Chen Shengze Li Ruochen Cao Qingyuan Hu Peiyun Zhao 《Computer Modeling in Engineering & Sciences》 SCIE EI 2024年第7期1143-1163,共21页
This paper presents a novel framework aimed at quantifying uncertainties associated with the 3D reconstruction of smoke from2Dimages.This approach reconstructs color and density fields from 2D images using Neural Radi... This paper presents a novel framework aimed at quantifying uncertainties associated with the 3D reconstruction of smoke from2Dimages.This approach reconstructs color and density fields from 2D images using Neural Radiance Field(NeRF)and improves image quality using frequency regularization.The NeRF model is obtained via joint training ofmultiple artificial neural networks,whereby the expectation and standard deviation of density fields and RGB values can be evaluated for each pixel.In addition,customized physics-informed neural network(PINN)with residual blocks and two-layer activation functions are utilized to input the density fields of the NeRF into Navier-Stokes equations and convection-diffusion equations to reconstruct the velocity field.The velocity uncertainties are also evaluated through ensemble learning.The effectiveness of the proposed algorithm is demonstrated through numerical examples.The presentmethod is an important step towards downstream tasks such as reliability analysis and robust optimization in engineering design. 展开更多
关键词 Uncertainty quantification neural radiance field physics-informed neural network frequency regularization twolayer activation function ensemble learning
下载PDF
MicroRNAs of microglia: wrestling with central nervous system disease 被引量:5
18
作者 Xiao-Hua Wang Tian-Long Wang 《Neural Regeneration Research》 SCIE CAS CSCD 2018年第12期2067-2072,共6页
Microglia serve as brain-resident myeloid cells that affect cerebral development, ischemia, neurodegeneration, and neuro-viral infection. MicroRNAs play a key role in central nervous system disease through post-transc... Microglia serve as brain-resident myeloid cells that affect cerebral development, ischemia, neurodegeneration, and neuro-viral infection. MicroRNAs play a key role in central nervous system disease through post-transcriptional regulation. Indeed, evidence shows that microRNAs are one of the most important regulators mediating microglial activation, polarization, and autophagy, and subsequently affecting neuroinflammation and the outcome of central nervous system disease. In this review, we provide insight into the function of microRNAs, which may be an attractive strategy and influential treatment for microglia-related central nervous system dysfunction. Moreover, we comprehensively describe how microglia fight against central nervous system disease via multiple functional microRNAs. 展开更多
关键词 MICROGLIA NEURODEGENERATION central nervous system disease MICRORNAS activation polarization AUTOPHAGY neural regeneration
下载PDF
用于手写体数字识别的视觉与神经网络混合算法 被引量:3
19
作者 盛积德 常胜江 +1 位作者 陈戍 张延炘 《光电子.激光》 EI CAS CSCD 北大核心 2001年第12期1280-1283,共4页
本文提出了用于无约束手写体数字特征提取和识别的视觉与神经网络混合算法。为了提高不变性特征的稳定性及网络收敛速度 ,我们引入周期包函数来取代传统的 sigmoid激活函数 ,计算机模拟结果显示该算法及激活函数能有效地提取手写体不变... 本文提出了用于无约束手写体数字特征提取和识别的视觉与神经网络混合算法。为了提高不变性特征的稳定性及网络收敛速度 ,我们引入周期包函数来取代传统的 sigmoid激活函数 ,计算机模拟结果显示该算法及激活函数能有效地提取手写体不变性特征 。 展开更多
关键词 神经网络 模式识别 激活函数 视觉 手写体数字识别
原文传递
SOM人工神经网络与阻抗谱方法结合研究涂层的失效过程 被引量:5
20
作者 赵霞 王佳 +3 位作者 孔涛 张伟 王燕华 潘国运 《腐蚀科学与防护技术》 CAS CSCD 北大核心 2008年第4期275-278,共4页
分析了干湿循环条件下有机涂层的电化学阻抗谱特征后选取了阻抗变化率参数作为一维自组织特征映射(self-organizing feature map,SOM)神经网络的训练样本,每个样本对输出层神经元都会产生一定的激发水平,根据该神经网络的特性,输出神经... 分析了干湿循环条件下有机涂层的电化学阻抗谱特征后选取了阻抗变化率参数作为一维自组织特征映射(self-organizing feature map,SOM)神经网络的训练样本,每个样本对输出层神经元都会产生一定的激发水平,根据该神经网络的特性,输出神经元激发水平的变化趋势便可以反映涂层的降解过程.结合涂层阻抗谱的变化特征,可以把干湿循环条件下涂层的失效过程分为三个阶段,即介质渗入涂层过程,介质到达金属表面诱发金属腐蚀过程和腐蚀扩展涂层剥离过程. 展开更多
关键词 电化学阻抗谱 阻抗变化率 神经网络 激发水平 降解
下载PDF
上一页 1 2 20 下一页 到第
使用帮助 返回顶部