The Long-Term Effective Mechanism of Rural Poverty Alleviation in China from the Perspective of Ecological Management

Based on the ecological environmental situation of poverty-stricken areas in China and the domestic and foreign research results,the long-term effective mechanism of sustainable poverty alleviation in China is established(the ideological premise is ecological culture,the material base is ecological economy and the basic guarantee is ecological system) from the perspective of ecological management.To be specific,ecological culture,the ideological premise of rural sustainable poverty alleviation in China,includes two aspects:the first one is upholding the Marxism,passing on the Chinese traditional ecological wisdom and fostering ecological consciousness;the second one is mirroring the thought of western modern ecological ethics,emphasizing ecological criticism and redoubling the ecological education.As for ecological economy,the material guarantee for sustainable poverty alleviation in Chinese rural poverty-stricken areas,also contents two aspects.The first one is promoting the way of "clean production" and developing ecological industry;the second one is building ecological concept of consumption and establishing the ecological lifestyle.In addition,ecological system,the basic guarantee of sustainable poverty alleviation in Chinese rural poverty-stricken areas,covers three terms.The first one is implementing the ecological policies and stipulating ecological plan;the second one is establishing perfect ecological legislation and ecological system and intensifying their implementation;the third one is enforcing ecological management in the process of sustainable poverty alleviation.Through the establishment of the above mechanism,we look forward to realizing sustainability of economic development and poverty alleviation effects in the process of poverty alleviation in Chinese rural areas,as well as ecological management of the poverty-stricken areas.

Countermeasures for Establishing a Long-term Security Mechanism for Landless Farmers

In the process of urbanization,farmers will inevitably lose their land,which is resulted from the large-scale requisition of rural collective land.The imperfect social security system in rural areas has resulted in a large number of landless farmers being transformed into urban poor people.Therefore,proper resettlement of landless farmers and the establishment of a long-term security mechanism for ensuring the basic livelihood of landless farmers are important issues that need to be resolved in the process of China’s social and economic development.

Building Indoor Dangerous Behavior Recognition Based on LSTM-GCN with Attention Mechanism

Building indoor dangerous behavior recognition is a specific application in the field of abnormal human recognition.A human dangerous behavior recognition method based on LSTM-GCN with attention mechanism(GLA)model was proposed aiming at the problem that the existing human skeleton-based action recognition methods cannot fully extract the temporal and spatial features.The network connects GCN and LSTMnetwork in series,and inputs the skeleton sequence extracted by GCN that contains spatial information into the LSTM layer for time sequence feature extraction,which fully excavates the temporal and spatial features of the skeleton sequence.Finally,an attention layer is designed to enhance the features of key bone points,and Softmax is used to classify and identify dangerous behaviors.The dangerous behavior datasets are derived from NTU-RGB+D and Kinetics data sets.Experimental results show that the proposed method can effectively identify some dangerous behaviors in the building,and its accuracy is higher than those of other similar methods.

Could Long-Term Stability Last Forever?

The subject of the present paper is to prove that the recently introduced conjecture of boundedness puts a ban over the view of stability as asymptotic property. This result comes in sharp contrast with the prescription of the traditional thermodynamics and statistical physics which consider the existence of equilibrium as asymptotic property of all systems. The difference commences from the use of infinitesimal calculus as the basic implement for modelling by the latter while the primary premise of the conjecture of boundedness is sustaining the energy/matter/information permanently bounded and finite. The latter property overrules the infinitesimal calculus as the major implement of modelling because, among all, it is proven that the traditional one suffers unsoluble difficulties.

MICROSTRUCTURE AND MECHANICAL PROPERTIES OF Fe-Ni-Co-BASE SUPERALLOY AFTER LONGTERM AGING AT 650℃

The microstructure and mechanical properties of a Ni-Fe-Co base superalloy with low thermal expansion coefficient after aging at 650℃ for 200,500,1000 and 2000 h have been investigated.It was found that they are stable at 650℃.This alloy is believed to.fill the re- quirements for long time service to aircraft engine.

基于自注意力机制和改进的K-BiLSTM的水产养殖水体溶解氧含量预测模型

为精确预测水产养殖水体溶解氧含量,本研究提出一种基于自注意力机制(ATTN)和改进的K-means聚类-基于残差和批标准化(BN)的双向长短期记忆网络(BiLSTM)的水产养殖水体溶解氧含量预测模型。首先,根据环境数据的相似性,使用改进的K-means算法将数据划分成若干个类别;然后,在BiLSTM基础上构建残差连接和加入BN完成高层次特征提取,利用BiLSTM的长期记忆能力保存特征信息;最后,引入自注意力机制突出不同时间节点数据特征的重要性,进一步提升模型的性能。试验结果表明,本研究提出的基于自注意力机制和改进的K-BiLSTM模型的平均绝对误差为0.238、均方根误差为0.322、平均绝对百分比误差为0.035,与单一的BP模型、CNN-LSTM模型、传统的K-means-基于残差和BN的BiLSTM-ATTN等模型相比具有更优的预测性能和泛化能力。

基于注意力机制与LSTM-CCN的月降水量预测

针对现有月降水量预测方法预测准确性不高的问题,提出一种基于注意力机制与LSTM-CCN的月降水量预测方法。首先,利用长短时记忆神经网络(long short-term memory neural network,LSTM)提取气象数据在时间维度的特征分布,从时间相关性方面捕获相邻时间段或长距离气象数据段中的统计分布;其次,利用因果卷积神经网络(causal convolutional network,CCN)将气象数据映射到空间维度,深层次地从空间维度捕获气象数据在空间中的特征统计分布;再次,以并联的方式将时间和空间特征作为交叉注意力网络的输入,构造融合的时空特征;最后,以长短时记忆神经网络构造解码器,并将融合的时空特征作为解码器的输入,预测的月降水量作为输出。选取河南省新乡市2001~2017年数据集进行测试,结果表明:所提出方法的均方根误差仅为13.08 mm,相比主流方法具有更低的预测误差。研究成果可为提高气象预测的准确性和实用性提供参考。

CEEMDAN-WPE-CLSA超短期风电功率预测方法研究

提出了一种结合自适应噪声完全集合经验模态分解、加权排列熵、卷积神经网络、长短期记忆网络和自注意力机制的超短期风电功率预测方法。首先,利用自适应噪声完全集合经验模态分解将原始风电功率时间序列自适应分解为一系列的模态分量,降低原始序列的非线性和波动性;其次,根据加权排列熵计算各模态分量间的相似性并对相似的分量进行重组,以修正自适应噪声完全集合经验模态分解的过度分解问题,使得修正后的模态分量更具规律性;最后,将重组后的分量输入卷积长短期记忆网络进行时序建模,并利用自注意力机制对卷积长短期记忆网络的神经元权重进行重新分配,提高了卷积长短期记忆网络对输入特征不确定性的适应能力。在此基础上,明确了自注意力机制和自适应噪声完全集合经验模态分解、加权排列熵在风电功率预测中的作用机制,以及风电功率信号包含的重要物理信息,证明了自适应噪声完全集合经验模态分解、加权排列熵以及自注意力机制在风电功率信号模态分解和长短期记忆网络隐层输出权重分配中的有效性。

基于AM-LSTM的飞行区航空器滑行轨迹预测与冲突识别

为解决航空器点源定位难以有效预测而引发冲突风险愈来愈多的问题,构建基于注意力机制(AM)和长短期记忆网络(LSTM)的时间序列轨迹预测模型AM-LSTM,预测未来短时间内飞行区航空器的瞬时点源位置;在此基础上,根据航空器型号和滑行航向对其进行轮廓扩展,以航空器速度作为安全距离权重,通过射线法实现轮廓冲突的判定;并以乌鲁木齐地窝堡机场为例进行验证,利用训练完成的轨迹预测模型预测飞行区航空器滑行轨迹,以识别航空器轮廓间的滑行冲突。结果表明:AM-LSTM预测模型能够准确预测飞行区航空器运动轨迹。未来3 s内轨迹位置预测的平均位移误差为1.05 m,轨迹点位置预测精准性可达94.37%,故能在轨迹预测的基础上精确识别滑行冲突风险,有利于保障飞行区的安全运行。

基于Transformer-LSTM的闽南语唇语识别

针对端到端句子级闽南语唇语识别的问题,提出一种基于Transformer和长短时记忆网络(LSTM)的编解码模型.编码器采用时空卷积神经网络及Transformer编码器用于提取唇读序列时空特征,解码器采用长短时记忆网络并结合交叉注意力机制用于文本序列预测.最后,在自建闽南语唇语数据集上进行实验.实验结果表明:模型能有效地提高唇语识别的准确率.

基于注意力机制的CNN-BiLSTM的IGBT剩余使用寿命预测

针对绝缘栅双极型晶体管(IGBT)可靠性问题,提出了一种融合卷积神经网络(CNN)、双向长短期记忆(BiLSTM)网络和注意力机制的剩余使用寿命(RUL)预测模型,可用于IGBT的寿命预测。模型中使用CNN提取特征参数,BiLSTM提取时序信息,注意力机制加权处理特征参数。使用IGBT加速老化数据集对提出的模型进行验证。结果表明,对比自回归差分移动平均(ARIMA)、长短期记忆(LSTM)、多层LSTM(Multi-LSTM)、 BiLSTM预测模型,在均方根误差和决定系数等评价指标方面该模型的性能最优。验证了提出的寿命预测模型对IGBT失效预测是有效的。

基于CBAM-LSTM的风电集群功率短期预测方法

风电功率的精准预测对我国实现“碳达峰”、“碳中和”的目标具有重要意义。传统的风电功率预测方法往往忽视了时间序列数据中的长期依赖关系和空间相关性,导致预测结果不准确。为了解决这个问题,文中提出了了卷积块注意力机制(Convolutional Block Attention Module, CBAM)和长短时记忆网络(Long Short-Term Memory, LSTM)相结合的模型。首先,使用CBAM对风电功率时间序列数据特征和数值天气预报中蕴含的空间特性进行提取,该模块能够自适应地学习时间和空间上的重要特征;然后,将提取的特征输入到LSTM层结构中进行功率预测。为了验证所提方法的有效性,使用中国吉林省某风电场的数据集进行验证,实验结果表明,与其他功率预测方法相比,文中所提方法平均绝对误差(Mean Absolute Error, MAE)平均降低2.67%;决定系数(R-Square, R2)平均提高23%;均方根误差(Root Mean Square Error, RMSE)平均降低2.69%。

基于改进LSTM网络的无人机MEMS-IMU零偏在线标定方法

针对在卫星信号拒止、视觉系统退化场景中无人机MEMS-IMU零偏无法准确估计并补偿导致导航误差迅速发散的问题,提出一种基于改进长短时记忆(LSTM)网络的零偏在线标定方法。首先,为解决MEMS-IMU零偏数据非线性强、传统循环时间网络训练效果差的问题,设计序列到序列的LSTM神经网络结构,引入教师强迫机制,提高了网络特征学习能力。然后,在导航过程中使用训练后的网络对MEMS-IMU零偏在线标定,补偿后的IMU量测与视觉信息联合优化,保证了导航定位精度。实验结果表明,在纯惯性导航实验中,所提方法的绝对位置误差比传统LSTM方法减小了6.5%;在EUROC数据集下进行的视觉惯性组合导航实验中,所提方法的平均绝对位置误差比传统LSTM方法减小了15%。

基于LSTM-SAFCN模型的生物质锅炉NO_(x)排放浓度预测

针对生物质锅炉燃烧过程的动态特性,提出一种改进的长短期记忆-自注意力机制全卷积神经网络(LSTM-SAFCN)模型用于预测NO_(x)排放浓度。首先利用完全自适应噪声集合经验模态分解法(CEEMDAN)对数据进行预处理,消除数据噪声对NO_(x)排放浓度预测的影响;其次融合自注意力机制与长短时记忆-全卷积神经网络(LSTM-FCN)进行特征提取与预测建模,该拓展方法能够同时兼顾时间序列数据的局部细节与长期趋势特征;最后,利用生物质热电联产系统的实际运行数据验证了所提算法的有效性。

Alkali-Silica Reaction Inhibited by LiOH and Its Mechanism

A high alkali reactive aggregate zeolitization perlite was used to test the long term effectiveness of LiOH in inhibiting alkali silica reaction.In this paper,the rigorous conditions were designed that the mortar bars had been cured at 80℃ for 3 years after autoclaved 24 hours at 150℃.Under this condition,LiOH was able to inhibit the alkali silica reaction long term effectiveness.Not only the relationship between the molar ratio of n(Li)/(Na) and the alkali contents in systems was established, but also the governing mechanism of such effects was also studied by SEM.

基于改进U-Net网络的相位解包裹技术研究

提出了一种结合深度学习的空间相位解包裹方法,采用基于改进U-Net网络的编码器-解码器架构,同时加入包含双向长短期记忆网络(BILSTM)的CBiLSTM模块,并且结合注意力机制,避免了典型卷积神经网络学习全局空间依赖关系的固有缺陷的同时增强了深度学习模型对相位解包裹任务中的关键信息的关注能力。通过大量的模拟数据,验证了文中方法在严重噪声(SNR=0)、不连续条件和混叠条件下的鲁棒性,在以上三种情况下,同其他深度学习网络模型进行对比,文中所提出的网络模型的归一化均方根误差(NRMSE)分别为0.75%、1.81%和1.68%;结构相似性指数(SSIM)分别为0.98、0.92和0.94;峰值信噪比(PSNR)分别为40.87、32.56、37.38;同时计算时间显著减少,适合应用到需要快速准确的空间相位解包裹任务中去。通过实际测量数据,验证了文中提出网络模型的可行性。该研究将双向长短期记忆网络(BILSTM)和注意力机制同时引入光学相位解包裹问题中,为解决复杂相位场的解包裹提供了新的思路和方案。

基于BiLSTM-Attention的F_(10.7)指数预测模型与中国自主数据集的应用

F_(10.7)指数是太阳活动的重要指标,准确预测F_(10.7)指数有助于预防和缓解太阳活动对无线电通信、导航和卫星通信等领域的影响.基于F_(10.7)射电流量的特性,在双向长短时记忆网络(Bidirectional Long Short-Term Memory Network,BiLSTM)基础上融入注意力机制(Attention),提出了一种基于BiLSTM-Attention的F_(10.7)预报模型.在加拿大DRAO数据集上其平均绝对误差(MAE)为5.38,平均绝对百分比误差(MAPE)控制在5%以内,相关系数(R)高达0.987,与其他RNN模型相比拥有优越的预测性能.针对中国廊坊L&S望远镜观测的F_(10.7)数据集,提出了一种转换平均校准(Conversion Average Calibration,CAC)方法进行数据预处理,处理后的数据与DRAO数据集具有较高的相关性.基于该数据集对比分析了RNN系列模型的预报效果,实验结果表明,BiLSTM-Attention和BiLSTM两种模型在预测F_(10.7)指数方面具有较好的优势,表现出较好的预测性能和稳定性.

CNS-Net:一种循环多注意力特征聚合架构

基于卷积神经网络(convolutional neural network,CNN)方法在高分辨率遥感图像分类时,提取全局特征信息和局部特征信息方面有一定优势,但不能有效区别关键信息和干扰信息。由此,提出一种端到端的CNS-Net网络来提取图像显著性特征。首先,设计了一个全局增强性模块(global enhancement module,GEM)通过显示建模通道之间的相互依赖关系,使网络选择性提取关键区域;其次,提出了一个多级循环注意模块(multi-stage cyclic attention module,MCAM)来捕捉特征信息的长期依赖关系和上下文感知信息;最后,在4个公共数据集上进行实验,结果表明所提方法达到了最优的分类性能。

基于注意力的VAE-ConvLSTM模型的剩余寿命预测研究

为解决传统有监督算法难以达到较高预测精度的问题,提出了一种新的半监督模型。首先,将不同传感器采集的监测数据直接作为网络输入,减少了可能出现的信息损失;接着,在无监督部分采用变分自动编码器(variationalautoencoder,VAE),以自动提取输入数据的深层表达;然后,在有监督部分使用卷积长短期记忆(convolutionallongshort-term memory, ConvLSTM)网络进一步提取时序数据的时空特征,并引入注意力机制,提高重要特征因子的权重;最后,在NASA提供的C-MAPSS数据集上进行对比实验,以均方根误差和数据集自定义的Score作为评价指标。实验结果表明,所提出的模型在复杂预测场景中取得了最好的结果,证明了该模型的有效性。