Feedback LSTM Network Based on Attention for Image Description Generator

Images are complex multimedia data which contain rich semantic information.Most of current image description generator algorithms only generate plain description,with the lack of distinction between primary and secondary object,leading to insufficient high-level semantic and accuracy under public evaluation criteria.The major issue is the lack of effective network on high-level semantic sentences generation,which contains detailed description for motion and state of the principal object.To address the issue,this paper proposes the Attention-based Feedback Long Short-Term Memory Network(AFLN).Based on existing codec framework,there are two independent sub tasks in our method:attention-based feedback LSTM network during decoding and the Convolutional Block Attention Module(CBAM)in the coding phase.First,we propose an attentionbased network to feedback the features corresponding to the generated word from the previous LSTM decoding unit.We implement feedback guidance through the related field mapping algorithm,which quantifies the correlation between previous word and latter word,so that the main object can be tracked with highlighted detailed description.Second,we exploit the attention idea and apply a lightweight and general module called CBAM after the last layer of VGG 16 pretraining network,which can enhance the expression of image coding features by combining channel and spatial dimension attention maps with negligible overheads.Extensive experiments on COCO dataset validate the superiority of our network over the state-of-the-art algorithms.Both scores and actual effects are proved.The BLEU 4 score increases from 0.291 to 0.301 while the CIDEr score rising from 0.912 to 0.952.

Behavior recognition based on the fusion of 3D-BN-VGG and LSTM network

In order to effectively solve the problems of low accuracy,large amount of computation and complex logic of deep learning algorithms in behavior recognition,a kind of behavior recognition based on the fusion of 3 dimensional batch normalization visual geometry group(3D-BN-VGG)and long short-term memory(LSTM)network is designed.In this network,3D convolutional layer is used to extract the spatial domain features and time domain features of video sequence at the same time,multiple small convolution kernels are stacked to replace large convolution kernels,thus the depth of neural network is deepened and the number of network parameters is reduced.In addition,the latest batch normalization algorithm is added to the 3-dimensional convolutional network to improve the training speed.Then the output of the full connection layer is sent to LSTM network as the feature vectors to extract the sequence information.This method,which directly uses the output of the whole base level without passing through the full connection layer,reduces the parameters of the whole fusion network to 15324485,nearly twice as much as those of 3D-BN-VGG.Finally,it reveals that the proposed network achieves 96.5%and 74.9%accuracy in the UCF-101 and HMDB-51 respectively,and the algorithm has a calculation speed of 1066 fps and an acceleration ratio of 1,which has a significant predominance in velocity.

Prediction of surface subsidence in Changchun City based on LSTM network

Monitoring and predicting of urban surface subsidence are important for urban disaster prevention and mitigation.In this paper,the Long Short-Term Memory(LSTM)network was used to predict the surface subsidence process of Changchun City from 2018 to 2020 based on PS-InSAR monitoring data.The results show that the prediction error of 57.89% of PS points in the LSTM network was less than 1mm with the average error of 1.8 mm and the standard deviation of 2.8 mm.The accuracy and reliability of the prediction were better than regression analysis,time series analysis and grey model.

LSTM Based Neural Network Model for Anomaly Event Detection in Care-Independent Smart Homes

This study introduces a long-short-term memory(LSTM)-based neural network model developed for detecting anomaly events in care-independent smart homes,focusing on the critical application of elderly fall detection.It balances the dataset using the Synthetic Minority Over-sampling Technique(SMOTE),effectively neutralizing bias to address the challenge of unbalanced datasets prevalent in time-series classification tasks.The proposed LSTM model is trained on the enriched dataset,capturing the temporal dependencies essential for anomaly recognition.The model demonstrated a significant improvement in anomaly detection,with an accuracy of 84%.The results,detailed in the comprehensive classification and confusion matrices,showed the model’s proficiency in distinguishing between normal activities and falls.This study contributes to the advancement of smart home safety,presenting a robust framework for real-time anomaly monitoring.

LSTM Network-Based Adaptation Approach for Dynamic Integration in Intelligent End-Edge-Cloud Systems

Edge computing, which migrates compute-intensive tasks to run on the storage resources of edge devices, efficiently reduces data transmission loss and protects data privacy. However, due to limited computing resources and storage capacity, edge devices fail to support real-time streaming data query and processing. To address this challenge, first, we propose a Long Short-Term Memory (LSTM) network-based adaptive approach in the intelligent end-edge-cloud system. Specifically, we maximize the Quality of Experience (QoE) of users by automatically adapting their resource requirements to the storage capacity of edge devices through an event mechanism. Second, to reduce the uncertainty and non-complete adaption of the edge device towards the user’s requirements, we use the LSTM network to analyze the storage capacity of the edge device in real time. Finally, the storage features of the edge devices are aggregated to the cloud to re-evaluate the comprehensive capability of the edge devices and ensure the fast response of the user devices during the dynamic adaptation matching process. A series of experimental results show that the proposed approach has superior performance compared with traditional centralized and matrix decomposition based approaches.

Prediction and Analysis of Elevator Traffic Flow under the LSTM Neural Network

Elevators are essential components of contemporary buildings, enabling efficient vertical mobility for occupants. However, the proliferation of tall buildings has exacerbated challenges such as traffic congestion within elevator systems. Many passengers experience dissatisfaction with prolonged wait times, leading to impatience and frustration among building occupants. The widespread adoption of neural networks and deep learning technologies across various fields and industries represents a significant paradigm shift, and unlocking new avenues for innovation and advancement. These cutting-edge technologies offer unprecedented opportunities to address complex challenges and optimize processes in diverse domains. In this study, LSTM (Long Short-Term Memory) network technology is leveraged to analyze elevator traffic flow within a typical office building. By harnessing the predictive capabilities of LSTM, the research aims to contribute to advancements in elevator group control design, ultimately enhancing the functionality and efficiency of vertical transportation systems in built environments. The findings of this research have the potential to reference the development of intelligent elevator management systems, capable of dynamically adapting to fluctuating passenger demand and optimizing elevator usage in real-time. By enhancing the efficiency and functionality of vertical transportation systems, the research contributes to creating more sustainable, accessible, and user-friendly living environments for individuals across diverse demographics.

基于LSTM网络的单台仪器地震烈度预测模型

烈度是地震预警系统的关键产出.如何实现快速预测目标场址的地震烈度是地震预警方法技术研究中的核心问题.本文提出了一种基于长短时记忆神经网络(Long Short-Term Memory,LSTM)的单台仪器地震烈度的预测模型(LSTM-Ⅰ).该模型以一个台站观测到地震动参数的时间序列特征为输入,实现动态预测该台站可能遭受的最大烈度.选取了日本K-NET台网记录的102次地震的5103条强震加速度记录训练了神经网络,利用89次地震的3781条数据检验了模型的泛化能力.利用准确率、漏报率以及误报率三个评价指标评价了LSTM-Ⅰ模型的性能.结果表明,当采用P波触发后3 s的序列进行预测时,模型出现漏报的概率为46.78%,出现误报的概率为1.25%;当采用P波触发后10 s的序列进行预测时,模型出现漏报的概率大幅降低到17.6%,出现误报的概率降低到1.14%.结果表明LSTM-Ⅰ模型很好把握住了时间序列中蕴含的特征.进一步基于LSTM-Ⅰ模型评估了Ⅵ度下台站所能提供的预警时间.本文模型能够提供的预警时间与P-S波到时差接近,说明LSTM-Ⅰ模型具有较高的时效性.

基于CNN-LSTM的大坝变形组合预测模型研究

为了提高大坝变形预测模型精度和泛化能力,建立了一种基于卷积神经网络(Convolutional neural networks,CNN)与深度学习长短期记忆(Long short-term memory,LSTM)神经网络的组合预测模型CNN-LSTM。该模型先利用CNN提取大坝变形监测时间序列的特征,再利用LSTM生成特征描述,该模型精度高、泛化能力强。以柏叶口水库混凝土面板堆石坝为例,经过CNN-LSTM模型计算,将模型变形预测值与原型监测资料进行对比,再与LSTM模型及CNN模型的预测结果进行对比。结果表明,CNN-LSTM模型预测值最接近监测资料实测结果。

基于CNN-LSTM混合神经网络的高速铁路地震响应预测

为了更好地挖掘高速铁路在地震时的响应信息,提高光纤光栅监测的效率及预测精度,该文针对地震响应数据的时序性及非线性的特点,提出卷积神经网络(CNN)和长短期记忆(LSTM)网络的混合神经网络模型预测方法.通过在高速铁路简支梁桥上布设准分布式光纤光栅采集地震时轨道板、钢轨、底座板、箱梁的响应数据,在每根光纤上布置7个光栅,利用两边光栅的响应数据预测中间点的光栅响应,将采集位置、历史数据及地震波形等信息作为特征图输入.利用CNN提取特征,再将提前提取出来的特征数据以时序方式作为LSTM网络的输入数据,最后LSTM网络进行地震应变响应预测.实验结果表明,LSTM网络在3层时效果最好,CNN-LSTM方法具有较高的预测精度,根均平方误差(R_(RMSE))、平均绝对误差(R_(MAE))、决定系数(R^(2))分别达到了0.3753、0.2968、0.9371.

适用于区域建筑群实时震害模拟的LSTM-FC组合深度网络模型研究

建筑物破坏在地震灾害中往往会导致巨大损失,对城市建筑群进行灾前灾时的震害预测具有重要意义。传统BP(back propagation)网络和CNN(convolutional neural networks)网络等人工智能方法在进行震害预测时多集中于提取建筑物信息。然而,这些方法在处理地震波的时序数据方面有所不足,导致其在整合和分析对地震灾害预测至关重要的时序相关因素时效果有限。因此,本文提出一种耦合LSTM(long short-term memory)和FC(fully connected)神经网络的震害预测方法。LSTM网络擅长处理具有时间序列特性的地震波信息,能够捕捉和分析随时间变化的地震波动模式。同时,全连接网络可用于综合分析所有相关的震害因子。通过对云浮地区265栋典型钢混建筑进行指标量化并确定输入指标(震害影响因子)和输出指标(震害指数),利用LSTM-FC组合深度网络、CNN网络和BP网络模型对数据进行训练并优化。通过将LSTM-FC网络模型的预测结果与弹塑性时程分析比较,发现该模型在拟合效果和精度方面优于传统的BP和CNN模型。拟合效果提升了36.8%和10.6%,精度分别提升了77.6%和91.7%,表明LSTM-FC网络在地震损害预测上更为有效。同时,将该方法应用于广东省云浮市钢混结构群震害预测,构建的易损性矩阵与华南地区的易损性矩阵均值进行了对比,显示误差相对较小,说明该模型不仅理论上可行,在实际应用中也能表现出较高的准确性和有效性。

基于改进INFO-Bi-LSTM模型的SO_(2)排放质量浓度预测

针对火电机组SO_(2)排放质量浓度的影响因素众多,难以准确预测的问题,提出一种改进向量加权平均(weighted mean of vectors,INFO)算法与双向长短期记忆(bi-directional long short term memory,Bi-LSTM)神经网络相结合的预测模型(改进INFO-Bi-LSTM模型)。采用Circle混沌映射和反向学习产生高质量初始化种群,引入自适应t分布提升INFO算法跳出局部最优解和全局搜索的能力。选取改进INFO-Bi-LSTM模型和多种预测模型对炉内外联合脱硫过程中4种典型工况下的SO_(2)排放质量浓度进行预测,将预测结果进行验证对比。结果表明:改进INFO算法的寻优能力得到提升,并且改进INFO-Bi-LSTM模型精度更高,更加适用于SO_(2)排放质量浓度的预测,可为变工况下的脱硫控制提供控制理论支撑。

基于CNN-LSTM的水泥熟料f-CaO预测模型

水泥熟料中游离氧化钙(f-CaO)含量的传统人工离线检测缺乏时效性,不利于生产指导。针对离线检测的滞后问题和软测量模型中f-CaO含量与辅助变量的时序匹配问题,提出了一种基于卷积神经网络(convolutional neural network,CNN)和长短时记忆(long short-term memory,LSTM)神经网络的f-CaO含量预测模型。首先,利用滑动窗口截取辅助变量的区间数据;然后,采用CNN提取区间数据的时序特征;之后,构建LSTM神经网络模型;最后,控制截取辅助变量的延迟时间和间隔时间,根据模型预测拟合度提取辅助变量的最优时序特征。仿真结果表明,所提模型提高了水泥熟料中f-CaO含量的预测精度。

基于深度学习的LSTM-GRU复合模型矿井涌水量预测方法研究

为了解决矿井涌水预测问题,引入深度学习理论,将长短期记忆网络(LSTM)和门控循环单元(GRU)进行结合,选取矿井涌水量为研究对象,建立一种LSTM-GRU的矿井涌水预测模型。以陕西某矿的矿井涌水量为样本数据,采用7∶3的比例将数据集划分为训练集和测试集,选择模型训练效果较好的梯度下降算法确定网络模型参数和正则化参数,为了证明LSTM-GRU模型的预测精度,同时将结果分别与传统的ARIMA模型和LSTM模型预测矿井涌水所得到的预测结果进行对比。结果表明:LSTM-GRU复合模型的平均绝对百分比误差(RMSE)为70.51,均方根误差(MAE)为53.4,平均绝对误差(MAPE)为2.80%,可决系数(R^(2))为0.86,具有较高的预测精度和可靠性,预测效果优于传统的ARIMA模型和LSTM模型。

基于PSO-LSTM的重载铁路车轨桥系统随机振动响应预测方法

在车桥系统随机振动分析中,随机动力响应是评价行车安全性的关键因素之一,而现有的响应计算方法存在耗时长、成本高的问题。能够快速准确预测车-轨-桥系统的动力响应对重载铁路桥梁的状态评估和运维养维具有重要意义。本文提出了一种基于粒子群优化(Particle Swarm Optimization,PSO)长短期记忆(Long Short-term Memory,LSTM)神经网络模型的重载车桥系统随机振动响应预测方法。该方法以车桥随机参数与轨道随机不平顺激励为输入,以桥梁动力响应为输出构造代理模型。首先,基于商业软件MATLAB平台构建PSO-LSTM网络模型;其次,通过建立的车-轨-桥系统随机振动分析模型计算初始样本集对应的随机动态响应,并进行模型训练,同时利用PSO算法优化LSTM结构参数;最后,使用训练好的PSO-LSTM模型对桥梁动态响应进行预测。为了验证本算法的优越性和鲁棒性,以朔黄重载铁路实测数据为例,对比本算法与BP(Back Propagation)神经网络、GRU(Gated Recurrent Unit)神经网络和LSTM神经网络的预测效率,并讨论不同车速下的预测情况,开展本模型与实测数据及有限元分析数据的对比分析。研究结果表明:在PSO优化下,LSTM模型预测结果得到一定的改善,PSO-LSTM模型拟合相关性系数可以达到0.97,其他评价误差值也均小于BP神经网络、GRU神经网络模型,本文模型可更高效准确地预测桥梁随机动力响应,可为进一步发展车-轨-桥系统随机振动响应预测理论提供技术支持。

一种融合GA和LSTM的边坡变形预测优化网络模型及其应用

考虑到BP神经网络模型忽略边坡监测数据存在的时间相关性,以及LSTM模型由于超参数选择存在主观性而易陷入局部最优等问题,提出一种基于遗传算法和长短期记忆网络(GA-LSTM)相结合的边坡变形预测模型,以发挥遗传算法全局搜索能力和LSTM预测时序数据的优势。以海明矿业露天采场边坡为研究对象,分别采用BP神经网络模型、LSTM网络模型以及GA-LSTM网络模型对边坡监测点GNSS49变形进行预测分析,并对比各模型达到收敛条件的时间。结果表明,GA-LSTM模型与其他模型达到同一收敛条件的时间差异不大,GA-LSTM模型的拟合准确度在0.1~0.2 mm,是LSTM神经网络模型的5~7倍,是BP神经网络模型的10~20倍,具有较高的精度和稳定性,其预测值与实际监测数据基本一致,可为矿山边坡的安全生产、管理以及决策控制提供科学依据。

基于GMM聚类的AM-BiLSTM机场安检旅客流量预测

针对现有安检旅客流量预测研究大多为正常情况下的预测,未考虑异常突发情况下安检旅客流量的变化趋势,提出一种基于高斯混合模型(GMM)聚类的融合注意力机制的多变量双向长短期记忆(AM-BiLSTM)机场安检旅客流量预测模型。首先,利用GMM聚类算法对原始数据集使用日期特征和延误特征分别进行聚类分析,根据聚类所得的不同日安检旅客流量场景构建不同的AM-BiLSTM旅客流量预测模型。实验结果表明:与现有多种预测方法相比,该方法在不同场景下均能准确预测各时段的安检旅客流量。

Short-Term Prediction of Photovoltaic Power Based on DBSCAN-SVM Data Cleaning and PSO-LSTM Model

Accurate short-termphotovoltaic(PV)power prediction helps to improve the economic efficiency of power stations and is of great significance to the arrangement of grid scheduling plans.In order to improve the accuracy of PV power prediction further,this paper proposes a data cleaning method combining density clustering and support vector machine.It constructs a short-termPVpower predictionmodel based on particle swarmoptimization(PSO)optimized Long Short-Term Memory(LSTM)network.Firstly,the input features are determined using Pearson’s correlation coefficient.The feature information is clustered using density-based spatial clustering of applications withnoise(DBSCAN),and then,the data in each cluster is cleanedusing support vectormachines(SVM).Secondly,the PSO is used to optimize the hyperparameters of the LSTM network to obtain the optimal network structure.Finally,different power prediction models are established,and the PV power generation prediction results are obtained.The results show that the data methods used are effective and that the PSO-LSTM power prediction model based on DBSCAN-SVM data cleaning outperforms existing typical methods,especially under non-sunny days,and that the model effectively improves the accuracy of short-term PV power prediction.

基于SO-LSTM的立柱液压系统故障诊断方法研究

针对目前无法快速、准确地诊断矿用立柱液压系统故障等问题,在建立仿真模型分析单一故障机制的基础上,基于优化算法提出多种故障诊断方法。将立柱物理模块与立柱液压系统模块相结合,建立立柱液压系统仿真模型;基于Simulink分析单一故障的影响,基于蛇优化LSTM神经网络建立诊断模型;最后,根据实际数据进行模型的实例验证。结果表明:蛇优化LSTM模型对液压立柱故障仿真数据识别率达到99.5%,对液压立柱故障真实数据识别率达到97%,与模型仿真数据的预测精度仅相差2.5%,预测精度较高,达到了预期目标。

基于K均值聚类算法和LSTM神经网络的管道腐蚀阶段预测方法

针对声发射检测获得的管道腐蚀信号,提出了一种基于K均值(K-means)聚类算法和长短期记忆(LSTM)神经网络的管道腐蚀阶段预测方法。首先,利用K-means聚类算法将腐蚀信号分类,再构建LSTM神经网络模型,并采取了无监督学习的方式,以声发射波形为出发点,对模型进行参数优化,最后进行管道腐蚀阶段预测,并根据评价指标对模型进行评价。研究表明:对LSTM神经网络模型适当增加隐藏层,可以使得模型更加稳定,鲁棒性更好;与现有故障诊断模型相比,LSTM神经网络模型的精度更高。

基于TDCSO优化CNN-Bi-LSTM网络的井底钻压预测方法

为了准确预测井底钻压,提高钻井效率、降低钻井成本,建立了融合双向长短期记忆网络(Bi-LSTM)和卷积神经网络(CNN)的混合模型。采用三角函数驱动的粒子群优化(TDCSO)方法对模型进行超参数优化,以提高预测钻压的精度;采用美国犹他州FORGE 58-32井和FORGE 58-62井的2个公开数据集对建立的模型进行验证,并采用平均绝对误差、均方根误差、决定系数和均方误差等指标进行模型性能评估。研究结果表明,所提出TDCSO-CNN-Bi-LSTM模型平均绝对误差、均方误差和均方根误差等3个关键性能指标较好,其中决定系数大于0.980,明显优于现有的LSTM、GRU、CNN-LSTM、CNN-Bi-LSTM等方法。研究表明,所提出的TDCSO-CNN-Bi-LSTM模型在井底钻压预测方面具有出色的准确性,能够实现实时监测,并与自动钻进系统集成,实现对钻压的精准控制,不仅提高了钻井效率,还降低了钻井成本,对未来的钻井作业具有重要的实际应用价值。