耦合Encoder-Decoder与RFR的径流预报模型研究

针对传统径流预报模型存在可靠性不高的缺陷,提出耦合Encoder-Decoder与RFR的径流预报模型,即通过Encoder-Decoder架构深度学习模块对径流-气象资料进行编码、解码处理以提取得到新的语义特征,进而将其作为输入变量用以随机森林回归(RFR)拟合。在阜阳市径流量预报实证中表明,Encoder-Decoder与RFR模型的R2=0.75,MAE、RMSE分别为3.75、4.26亿m3;较之于RFR模型的R2提升了12.67%,而MAE和RMSE依次减小了17.40%、16.63%。

基于encoder-decoder框架的城镇污水厂出水水质预测

由于污水厂的出水水质指标繁多、污水处理过程中反应复杂、时序非线性程度高,基于机理模型的预测方法无法取得理想效果。针对此问题,提出基于深度学习的污水厂出水水质预测方法,并以吉林省某污水厂监测水质为来源数据,利用多种结合encoder-decoder结构的神经网络预测水质。结果显示,所提结构对LSTM和GRU网络预测能力都有一定提升,对长期预测能力提升更加显著,ED-GRU模型效果最佳,短期预测中的4个出水水质指标均方根误差(RMSE)为0.7551、0.2197、0.0734、0.3146,拟合优度(R2)为0.9013、0.9332、0.9167、0.9532,可以预测出水质局部变化,而长期预测中的4个指标RMSE为1.7204、1.7689、0.4478、0.8316,R2为0.4849、0.5507、0.4502、0.7595,可以预测出水质变化趋势,与顺序结构相比,短期预测RMSE降低10%以上,R2增加2%以上,长期预测RMSE降低25%以上,R2增加15%以上。研究结果表明,基于encoder-decoder结构的神经网络可以对污水厂出水水质进行准确预测,为污水处理工艺改进提供技术支撑。

基于时空特征融合的Encoder-Decoder多步4D短期航迹预测

航迹预测在确保空中交通安全、高效运行中扮演着至关重要的角色。所预测的航迹信息是航迹优化、冲突告警等决策工具的输入,而预测准确性取决于模型对航迹序列特征的提取能力。航迹序列数据是具有丰富时空特征的多维时间序列,其中每个变量都呈现出长短期的时间变化模式,并且这些变量之间还存在着相互依赖的空间信息。为了充分提取这种时空特征,本文提出了基于融合时空特征的编码器-解码器(Spatio-Temporal EncoderDecoder,STED)航迹预测模型。在Encoder中使用门控循环单元(Gated Recurrent Unit,GRU)、卷积神经网络(Convolutional Neural Network,CNN)和注意力机制(Attention,AT)构成的双通道网络来分别提取航迹时空特征,Decoder对时空特征进行拼接融合,并利用GRU对融合特征进行学习和递归输出,实现对未来多步航迹信息的预测。利用真实的航迹数据对算法性能进行验证,实验结果表明,所提STED网络模型能够在未来10 min预测范围内进行高精度的短期航迹预测,相比于LSTM、CNN-LSTM和AT-LSTM等数据驱动航迹预测模型具有更高的精度。此外,STED网络模型预测一个航迹点平均耗时为0.002 s,具有良好的实时性。

利用Encoder-Decoder框架的深度学习网络实现绕射波分离及成像

利用单纯绕射波场实现地下地质异常体的识别具有坚实的理论基础,对应的实施方法得到了广泛研究,且有效地应用于实际勘探。但现有技术在微小尺度异常体成像方面收效甚微,相关研究多数以射线传播理论为基础,对于影响绕射波分离成像精度的因素分析并不完备。相较于反射波,由于存在不连续构造而产生的绕射波能量微弱并且相互干涉,同时环境干扰使得绕射波进一步湮没。因此,更高精度的波场分离及单独成像是现阶段基于绕射波超高分辨率处理、解释的重点研究方向。为此,首先针对地球物理勘探中地质异常体的准确定位,以携带高分辨率信息的绕射波为研究对象,系统分析在不同尺度、不同物性参数的异常体情况下绕射波的能量大小及形态特征,掌握绕射波与其他类型波叠加的具体形式;然后根据相应特征性质提出基于深度学习技术的绕射波分离成像方法,即利用Encoder-Decoder框架的空洞卷积网络捕获绕射波场特征,从而实现绕射波分离,基于速度连续性原则构建单纯绕射波场的偏移速度模型并完成最终成像。数据测试表明,该方法最终可满足微小地质异常体高精度识别的需求。

Underwater Acoustic Signal Noise Reduction Based on a Fully Convolutional Encoder-Decoder Neural Network

Noise reduction analysis of signals is essential for modern underwater acoustic detection systems.The traditional noise reduction techniques gradually lose efficacy because the target signal is masked by biological and natural noise in the marine environ-ment.The feature extraction method combining time-frequency spectrograms and deep learning can effectively achieve the separation of noise and target signals.A fully convolutional encoder-decoder neural network(FCEDN)is proposed to address the issue of noise reduc-tion in underwater acoustic signals.The time-domain waveform map of underwater acoustic signals is converted into a wavelet low-frequency analysis recording spectrogram during the denoising process to preserve as many underwater acoustic signal characteristics as possible.The FCEDN is built to learn the spectrogram mapping between noise and target signals that can be learned at each time level.The transposed convolution transforms are introduced,which can transform the spectrogram features of the signals into listenable audio files.After evaluating the systems on the ShipsEar Dataset,the proposed method can increase SNR and SI-SNR by 10.02 and 9.5dB,re-spectively.

基于Encoder-Decoder-ILSTM模型的瓦斯浓度预测研究

近年来,神经网络在各领域均发挥了巨大作用,同样在煤矿瓦斯浓度预测当中也有应用。为了提高模型的预测精度和实时性,结合Encoder-Decoder结构、长短期记忆形成、蛇优化算法提出了一种新的神经网络,为促进煤矿安全生产提供了技术支持。

基于注意力机制的Encoder-Decoder光伏发电预测模型

影响光伏发电系统出力的天气因素具有很大的波动性和不连续性,因此需要创建合适的预测模型来对光伏出力特性进行精准预测,从而保证电网系统的有效运行。本文通过最大信息系数选择合适的历史光伏发电数据,将其作为特征之一进行输入数据重构,并在由LSTM神经元构建的Encoder-Decoder模型上引入注意力机制,最终得到结合注意力机制的Encoder-Decoder光伏发电预测模型。经实际光伏电厂算例分析,验证了所提模型在光伏发电预测方面的准确性和适用性。

A Road Extraction Method for Remote Sensing Image Based on Encoder-Decoder Network

According to the characteristics of the road features,an Encoder-Decoder deep semantic segmentation network is designed for the road extraction of remote sensing images.Firstly,as the features of the road target are rich in local details and simple in semantic features,an Encoder-Decoder network with shallow layers and high resolution is designed to improve the ability to represent detail information.Secondly,as the road area is a small proportion in remote sensing images,the cross-entropy loss function is improved,which solves the imbalance between positive and negative samples in the training process.Experiments on large road extraction datasets show that the proposed method gets the recall rate 83.9%,precision 82.5%and F1-score 82.9%,which can extract the road targets in remote sensing images completely and accurately.The Encoder-Decoder network designed in this paper performs well in the road extraction task and needs less artificial participation,so it has a good application prospect.

Encoder-Decoder Based LSTM Model to Advance User QoE in 360-Degree Video

The development of multimedia content has resulted in a massiveincrease in network traffic for video streaming. It demands such types ofsolutions that can be addressed to obtain the user’s Quality-of-Experience(QoE). 360-degree videos have already taken up the user’s behavior by storm.However, the users only focus on the part of 360-degree videos, known as aviewport. Despite the immense hype, 360-degree videos convey a loathsomeside effect about viewport prediction, making viewers feel uncomfortablebecause user viewport needs to be pre-fetched in advance. Ideally, we canminimize the bandwidth consumption if we know what the user motionin advance. Looking into the problem definition, we propose an EncoderDecoder based Long-Short Term Memory (LSTM) model to more accuratelycapture the non-linear relationship between past and future viewport positions. This model takes the transforming data instead of taking the direct inputto predict the future user movement. Then, this prediction model is combinedwith a rate adaptation approach that assigns the bitrates to various tiles for360-degree video frames under a given network capacity. Hence, our proposedwork aims to facilitate improved system performance when QoE parametersare jointly optimized. Some experiments were carried out and compared withexisting work to prove the performance of the proposed model. Last but notleast, the experiments implementation of our proposed work provides highuser’s QoE than its competitors.

Classification of Arrhythmia Based on Convolutional Neural Networks and Encoder-Decoder Model

As a common and high-risk type of disease,heart disease seriously threatens people’s health.At the same time,in the era of the Internet of Thing(IoT),smart medical device has strong practical significance for medical workers and patients because of its ability to assist in the diagnosis of diseases.Therefore,the research of real-time diagnosis and classification algorithms for arrhythmia can help to improve the diagnostic efficiency of diseases.In this paper,we design an automatic arrhythmia classification algorithm model based on Convolutional Neural Network(CNN)and Encoder-Decoder model.The model uses Long Short-Term Memory(LSTM)to consider the influence of time series features on classification results.Simultaneously,it is trained and tested by the MIT-BIH arrhythmia database.Besides,Generative Adversarial Networks(GAN)is adopted as a method of data equalization for solving data imbalance problem.The simulation results show that for the inter-patient arrhythmia classification,the hybrid model combining CNN and Encoder-Decoder model has the best classification accuracy,of which the accuracy can reach 94.05%.Especially,it has a better advantage for the classification effect of supraventricular ectopic beats(class S)and fusion beats(class F).

Robust Cultivated Land Extraction Using Encoder-Decoder

Cultivated land extraction is essential for sustainable development and agriculture.In this paper,the network we propose is based on the encoder-decoder structure,which extracts the semantic segmentation neural network of cultivated land from satellite images and uses it for agricultural automation solutions.The encoder consists of two part:the first is the modified Xception,it can used as the feature extraction network,and the second is the atrous convolution,it can used to expand the receptive field and the context information to extract richer feature information.The decoder part uses the conventional upsampling operation to restore the original resolution.In addition,we use the combination of BCE and Loves-hinge as a loss function to optimize the Intersection over Union(IoU).Experimental results show that the proposed network structure can solve the problem of cultivated land extraction in Yinchuan City.

Rethinking the Encoder-decoder Structure in Medical Image Segmentation from Releasing Decoder Structure

Medical image segmentation has witnessed rapid advancements with the emergence of encoder-decoder based methods.In the encoder-decoder structure,the primary goal of the decoding phase is not only to restore feature map resolution,but also to mitigate the loss of feature information incurred during the encoding phase.However,this approach gives rise to a challenge:multiple up-sampling operations in the decoder segment result in the loss of feature information.To address this challenge,we propose a novel network that removes the decoding structure to reduce feature information loss(CBL-Net).In particular,we introduce a Parallel Pooling Module(PPM)to counteract the feature information loss stemming from conventional and pooling operations during the encoding stage.Furthermore,we incorporate a Multiplexed Dilation Convolution(MDC)module to expand the network's receptive field.Also,although we have removed the decoding stage,we still need to recover the feature map resolution.Therefore,we introduced the Global Feature Recovery(GFR)module.It uses attention mechanism for the image feature map resolution recovery,which can effectively reduce the loss of feature information.We conduct extensive experimental evaluations on three publicly available medical image segmentation datasets:DRIVE,CHASEDB and MoNuSeg datasets.Experimental results show that our proposed network outperforms state-of-the-art methods in medical image segmentation.In addition,it achieves higher efficiency than the current network of coding and decoding structures by eliminating the decoding component.

Longs short-term based semi-supervised memory encoder-decoder for early prediction of failures in self-lubricating bearings

The existing knowledge regarding the interfacial forces,lubrication,and wear of bearings in real-world operation has significantly improved their designs over time,allowing for prolonged service life.As a result,self-lubricating bearings have become a viable alternative to traditional bearing designs in industrial machines.However,wear mechanisms are still inevitable and occur progressively in self-lubricating bearings,as characterized by the loss of the lubrication film and seizure.Therefore,monitoring the stages of the wear states in these components will help to impart the necessary countermeasures to reduce the machine maintenance downtime.This article proposes a methodology for using a long short-term memory(LSTM)-based encoder-decoder architecture on interfacial force signatures to detect abnormal regimes,aiming to provide early predictions of failure in self-lubricating sliding contacts even before they occur.Reciprocating sliding experiments were performed using a self-lubricating bronze bushing and steel shaft journal in a custom-built transversally oscillating tribometer setup.The force signatures corresponding to each cycle of the reciprocating sliding motion in the normal regime were used as inputs to train the encoder-decoder architecture,so as to reconstruct any new signal of the normal regime with the minimum error.With this semi-supervised training exercise,the force signatures corresponding to the abnormal regime could be differentiated from the normal regime,as their reconstruction errors would be very high.During the validation procedure for the proposed LSTM-based encoder-decoder model,the model predicted the force signals corresponding to the normal and abnormal regimes with an accuracy of 97%.In addition,a visualization of the reconstruction error across the entire force signature showed noticeable patterns in the reconstruction error when temporally decoded before the actual critical failure point,making it possible to be used for early predictions of failure.

Action-Aware Encoder-Decoder Network for Pedestrian Trajectory Prediction

Accurate pedestrian trajectory predictions are critical in self-driving systems,as they are fundamental to the response-and decision-making of ego vehicles.In this study,we focus on the problem of predicting the future trajectory of pedestrians from a first-person perspective.Most existing trajectory prediction methods from the first-person view copy the bird’s-eye view,neglecting the differences between the two.To this end,we clarify the differences between the two views and highlight the importance of action-aware trajectory prediction in the first-person view.We propose a new action-aware network based on an encoder-decoder framework with an action prediction and a goal estimation branch at the end of the encoder.In the decoder part,bidirectional long short-term memory(Bi-LSTM)blocks are adopted to generate the ultimate prediction of pedestrians’future trajectories.Our method was evaluated on a public dataset and achieved a competitive performance,compared with other approaches.An ablation study demonstrates the effectiveness of the action prediction branch.

基于Encoder-Decoder网络的列车轮对激光曲线提取

提出了一种基于Encoder-Decoder网络的列车轮对激光曲线精确分割的算法。针对列车轮对激光曲线数据集局部特征丰富、语义信息简单的特点,设计了具有深度较浅、分辨率较高、细节表现良好的网络。设计的网络很好的利用了密集链接机制和上采样模块,加强了特征复用以及特征传播,具有较少参数的同时,能多尺度提取上下文语义信息。实验证明,Encoder-Decoder网络相比于其他网络在列车轮对激光曲线提取上表现出优异的性能。基于Encoder-Decoder的网络在列车轮对激光曲线数据集上交并比、召回率、准确率和F1score指标分别达到了86.5%、89.2%、99.9%、85.0%,能够比较精准提取列车轮对激光条纹。同时Encoder-Decoder网络在进行列车轮对激光条纹分割时能在一定程度上改善噪声对条纹提取的影响。因此在铁路安全方面具有良好的应用前景。

耦合Encoder-Decoder的LSTM径流预报模型研究

将长短期记忆神经网络(long short-term memory neural network,LSTM)与Encoder-Decoder结构耦合应用为LSTM-ED模型,并与LSTM人工智能径流预报模型进行比较。通过在闽江建溪流域进行应用,结果表明,相较于LSTM,LSTM-ED在检验期整体和各预见期具有更高的精度和稳定性,且对于典型洪水的预报洪峰误差更小,其独有的语义向量可以保持水文信息的连续性,预报径流过程不易受降雨波动干扰。2个模型的预报能力都与流域最大汇流时间密切相关,当预见期小于流域最大汇流时间时,2个模型都有很好的预报能力;当预见期大于流域最大汇流时间时,模型预报能力显著变差;当预见期远大于流域最大汇流时间时,2个模型都失去预报可靠性。

Anomaly detection in smart grid based on encoder-decoder framework with recurrent neural network

Anomaly detection in smart grid is critical to enhance the reliability of power systems. Excessive manpower has to be involved in analyzing the measurement data collected from intelligent motoring devices while performance of anomaly detection is still not satisfactory. This is mainly because the inherent spatio-temporality and multi-dimensionality of the measurement data cannot be easily captured. In this paper, we propose an anomaly detection model based on encoder-decoder framework with recurrent neural network （RNN）. In the model, an input time series is reconstructed and an anomaly can be detected by an unexpected high reconstruction error. Both Manhattan distance and the edit distance are used to evaluate the difference between an input time series and its reconstructed one. Finally, we validate the proposed model by using power demand data from University of California, Riverside （UCR） time series classification archive and IEEE 39 bus system simulation data. Results from the analysis demonstrate that the proposed encoder-decoder framework is able to successfully capture anomalies with a precision higher than 95%.

基于贝叶斯优化的GRU网络轴承剩余使用寿命预测方法

传统的滚动轴承剩余使用寿命预测模型存在参数优化的困难。针对这一问题,笔者提出了一种基于贝叶斯优化的GRU网络滚动轴承剩余使用寿命预测方法,并进行了实验验证,即以PHM2012数据集为例,结合贝叶斯优化算法对基于Encoder-Decoder结构的门控循环单元(GRU)预测模型的多个超参数进行了优化。首先,对包含噪声的原始数据进行了小波包处理,从滚动轴承的振动机理和故障特征出发提取了时域特征,针对该时域特征进行了优化、筛选,并将其输入到模型中的编码器部分,进一步提取了更深层次的时序特征;其次,结合注意力机制与Encoder-Decoder结构,构造了双向GRU神经网络模型,在模型的高维超参数空间中采用贝叶斯优化方法搜索超参数,得到了最优的超参数组合,并在解码器中融入了线性变换,得到了滚动轴承的寿命预测值;最后,封装了全部模型构建、训练与使用过程,建立了基于贝叶斯优化的GRU网络滚动轴承寿命预测流程,并对方法的有效性进行了对比实验验证。研究结果表明:采用基于贝叶斯优化的GRU网络可以有效预测滚动轴承的剩余使用寿命,相比于其他3种方法的最优结果,基于贝叶斯优化的GRU网络的平均预测得分提高了8.01%;基于贝叶斯优化的GRU网络对于真实寿命较短的轴承预测结果较为准确,而对于真实寿命较长的轴承则没有出现预测值大于真实值的情况,可以作为轴承临近失效阶段剩余使用寿命估计的参考。

A Video Captioning Method by Semantic Topic-Guided Generation

In the video captioning methods based on an encoder-decoder,limited visual features are extracted by an encoder,and a natural sentence of the video content is generated using a decoder.However,this kind ofmethod is dependent on a single video input source and few visual labels,and there is a problem with semantic alignment between video contents and generated natural sentences,which are not suitable for accurately comprehending and describing the video contents.To address this issue,this paper proposes a video captioning method by semantic topic-guided generation.First,a 3D convolutional neural network is utilized to extract the spatiotemporal features of videos during the encoding.Then,the semantic topics of video data are extracted using the visual labels retrieved from similar video data.In the decoding,a decoder is constructed by combining a novel Enhance-TopK sampling algorithm with a Generative Pre-trained Transformer-2 deep neural network,which decreases the influence of“deviation”in the semantic mapping process between videos and texts by jointly decoding a baseline and semantic topics of video contents.During this process,the designed Enhance-TopK sampling algorithm can alleviate a long-tail problem by dynamically adjusting the probability distribution of the predicted words.Finally,the experiments are conducted on two publicly used Microsoft Research Video Description andMicrosoft Research-Video to Text datasets.The experimental results demonstrate that the proposed method outperforms several state-of-art approaches.Specifically,the performance indicators Bilingual Evaluation Understudy,Metric for Evaluation of Translation with Explicit Ordering,Recall Oriented Understudy for Gisting Evaluation-longest common subsequence,and Consensus-based Image Description Evaluation of the proposed method are improved by 1.2%,0.1%,0.3%,and 2.4% on the Microsoft Research Video Description dataset,and 0.1%,1.0%,0.1%,and 2.8% on the Microsoft Research-Video to Text dataset,respectively,compared with the existing video captioning methods.As a result,the proposed method can generate video captioning that is more closely aligned with human natural language expression habits.

Triple-Branch Asymmetric Network for Real-time Semantic Segmentation of Road Scenes

As the field of autonomous driving evolves, real-time semantic segmentation has become a crucial part of computer vision tasks. However, most existing methods use lightweight convolution to reduce the computational effort, resulting in lower accuracy. To address this problem, we construct TBANet, a network with an encoder-decoder structure for efficient feature extraction. In the encoder part, the TBA module is designed to extract details and the ETBA module is used to learn semantic representations in a high-dimensional space. In the decoder part, we design a combination of multiple upsampling methods to aggregate features with less computational overhead. We validate the efficiency of TBANet on the Cityscapes dataset. It achieves 75.1% mean Intersection over Union(mIoU) with only 2.07 million parameters and can reach 90.3 Frames Per Second(FPS).