Selection of a non-localized landing region for an unmanned helicopter based on an attention model

A new landing region selection algorithm for an unmanned helicopter is proposed based on an attention model.Different from the original attention model,some properties of the possible safe landing regions（e.g.,depth,regional color and motion features）are included in the selection algorithm.Furthermore,regional color and motion features are fused directly into the saliency map because these features do not have the ＂central-peripheral＂property.Experimental results validate the feasibility and efficiency of this approach.

A Dual Attention Encoder-Decoder Text Summarization Model

A worthy text summarization should represent the fundamental content of the document.Recent studies on computerized text summarization tried to present solutions to this challenging problem.Attention models are employed extensively in text summarization process.Classical attention techniques are utilized to acquire the context data in the decoding phase.Nevertheless,without real and efficient feature extraction,the produced summary may diverge from the core topic.In this article,we present an encoder-decoder attention system employing dual attention mechanism.In the dual attention mechanism,the attention algorithm gathers main data from the encoder side.In the dual attentionmodel,the system can capture and producemore rational main content.The merging of the two attention phases produces precise and rational text summaries.The enhanced attention mechanism gives high score to text repetition to increase phrase score.It also captures the relationship between phrases and the title giving them higher score.We assessed our proposed model with or without significance optimization using ablation procedure.Our model with significance optimization achieved the highest performance of 96.7%precision and the least CPU time among other models in both training and sentence extraction.

Recommendation Algorithm Integrating CNN and Attention System in Data Extraction

With the rapid development of the Internet globally since the 21st century,the amount of data information has increased exponentially.Data helps improve people’s livelihood and working conditions,as well as learning efficiency.Therefore,data extraction,analysis,and processing have become a hot issue for people from all walks of life.Traditional recommendation algorithm still has some problems,such as inaccuracy,less diversity,and low performance.To solve these problems and improve the accuracy and variety of the recommendation algorithms,the research combines the convolutional neural networks(CNN)and the attention model to design a recommendation algorithm based on the neural network framework.Through the text convolutional network,the input layer in CNN has transformed into two channels:static ones and non-static ones.Meanwhile,the self-attention system focuses on the system so that data can be better processed and the accuracy of feature extraction becomes higher.The recommendation algorithm combines CNN and attention system and divides the embedding layer into user information feature embedding and data name feature extraction embedding.It obtains data name features through a convolution kernel.Finally,the top pooling layer obtains the length vector.The attention system layer obtains the characteristics of the data type.Experimental results show that the proposed recommendation algorithm that combines CNN and the attention system can perform better in data extraction than the traditional CNN algorithm and other recommendation algorithms that are popular at the present stage.The proposed algorithm shows excellent accuracy and robustness.

An Efficient Attention-Based Strategy for Anomaly Detection in Surveillance Video

In the present technological world,surveillance cameras generate an immense amount of video data from various sources,making its scrutiny tough for computer vision specialists.It is difficult to search for anomalous events manually in thesemassive video records since they happen infrequently and with a low probability in real-world monitoring systems.Therefore,intelligent surveillance is a requirement of the modern day,as it enables the automatic identification of normal and aberrant behavior using artificial intelligence and computer vision technologies.In this article,we introduce an efficient Attention-based deep-learning approach for anomaly detection in surveillance video(ADSV).At the input of the ADSV,a shots boundary detection technique is used to segment prominent frames.Next,The Lightweight ConvolutionNeuralNetwork(LWCNN)model receives the segmented frames to extract spatial and temporal information from the intermediate layer.Following that,spatial and temporal features are learned using Long Short-Term Memory(LSTM)cells and Attention Network from a series of frames for each anomalous activity in a sample.To detect motion and action,the LWCNN received chronologically sorted frames.Finally,the anomaly activity in the video is identified using the proposed trained ADSV model.Extensive experiments are conducted on complex and challenging benchmark datasets.In addition,the experimental results have been compared to state-ofthe-artmethodologies,and a significant improvement is attained,demonstrating the efficiency of our ADSV method.

Modelling an Efficient URL Phishing Detection Approach Based on a Dense Network Model

The social engineering cyber-attack is where culprits mislead the users by getting the login details which provides the information to the evil server called phishing.The deep learning approaches and the machine learning are compared in the proposed system for presenting the methodology that can detect phishing websites via Uniform Resource Locator(URLs)analysis.The legal class is composed of the home pages with no inclusion of login forms in most of the present modern solutions,which deals with the detection of phishing.Contrarily,the URLs in both classes from the login page due,considering the representation of a real case scenario and the demonstration for obtaining the rate of false-positive with the existing approaches during the legal login pages provides the test having URLs.In addition,some model reduces the accuracy rather than training the base model and testing the latest URLs.In addition,a feature analysis is performed on the present phishing domains to identify various approaches to using the phishers in the campaign.A new dataset called the MUPD dataset is used for evaluation.Lastly,a prediction model,the Dense forward-backwards Long Short Term Memory(LSTM)model(d−FBLSTM),is presented for combining the forward and backward propagation of LSMT to obtain the accuracy of 98.5%on the initiated login URL dataset.

Non-human primate models in drug addiction deserve more attention

Dear Editor, The process of relapse involves firm or aberrant memories of environmental cues associated with drug craving or addiction. To date, it is not known where these memories are stored in the brain, what kinds of regulatory biological factors or molecules are involved, nor why it is so difficult to stop addiction psychologically. Currently, rodent animal models, such as the self-administration and conditioning place preference / aversion paradigm are still widely used in the studies of drug withdrawal syndromes or drug-associate memories. However, the differences between humans and rodents--particularly in terms of genetics, and pathology and pharmacology--have significantly limited the application of further studies on this topic. Essentially, rodents lack the longterm or life-time memories humans possess and lose their drug-associated memory only after a few weeks of withdrawal.

A Recurrent Attention and Interaction Model for Pedestrian Trajectory Prediction

The movement of pedestrians involves temporal continuity,spatial interactivity,and random diversity.As a result,pedestrian trajectory prediction is rather challenging.Most existing trajectory prediction methods tend to focus on just one aspect of these challenges,ignoring the temporal information of the trajectory and making too many assumptions.In this paper,we propose a recurrent attention and interaction(RAI)model to predict pedestrian trajectories.The RAI model consists of a temporal attention module,spatial pooling module,and randomness modeling module.The temporal attention module is proposed to assign different weights to the input sequence of a target,and reduce the speed deviation of different pedestrians.The spatial pooling module is proposed to model not only the social information of neighbors in historical frames,but also the intention of neighbors in the current time.The randomness modeling module is proposed to model the uncertainty and diversity of trajectories by introducing random noise.We conduct extensive experiments on several public datasets.The results demonstrate that our method outperforms many that are state-ofthe-art.

Recent Progresses in Deep Learning Based Acoustic Models

In this paper,we summarize recent progresses made in deep learning based acoustic models and the motivation and insights behind the surveyed techniques.We first discuss models such as recurrent neural networks(RNNs) and convolutional neural networks(CNNs) that can effectively exploit variablelength contextual information,and their various combination with other models.We then describe models that are optimized end-to-end and emphasize on feature representations learned jointly with the rest of the system,the connectionist temporal classification(CTC) criterion,and the attention-based sequenceto-sequence translation model.We further illustrate robustness issues in speech recognition systems,and discuss acoustic model adaptation,speech enhancement and separation,and robust training strategies.We also cover modeling techniques that lead to more efficient decoding and discuss possible future directions in acoustic model research.

Effective Video Summarization Approach Based on Visual Attention

Video summarization is applied to reduce redundancy and developa concise representation of key frames in the video, more recently, video summaries have been used through visual attention modeling. In these schemes,the frames that stand out visually are extracted as key frames based on humanattention modeling theories. The schemes for modeling visual attention haveproven to be effective for video summaries. Nevertheless, the high cost ofcomputing in such techniques restricts their usability in everyday situations.In this context, we propose a method based on KFE (key frame extraction)technique, which is recommended based on an efficient and accurate visualattention model. The calculation effort is minimized by utilizing dynamicvisual highlighting based on the temporal gradient instead of the traditionaloptical flow techniques. In addition, an efficient technique using a discretecosine transformation is utilized for the static visual salience. The dynamic andstatic visual attention metrics are merged by means of a non-linear weightedfusion technique. Results of the system are compared with some existing stateof-the-art techniques for the betterment of accuracy. The experimental resultsof our proposed model indicate the efficiency and high standard in terms ofthe key frames extraction as output.

Multi Head Deep Neural Network Prediction Methodology for High-Risk Cardiovascular Disease on Diabetes Mellitus

Major chronic diseases such as Cardiovascular Disease(CVD),diabetes,and cancer impose a significant burden on people and healthcare systems around the globe.Recently,Deep Learning(DL)has shown great potential for the development of intelligentmobile Health(mHealth)interventions for chronic diseases that could revolutionize the delivery of health care anytime,anywhere.The aimof this study is to present a systematic review of studies that have used DL based on mHealth data for the diagnosis,prognosis,management,and treatment of major chronic diseases and advance our understanding of the progress made in this rapidly developing field.Type 2 Diabetes Mellitus(T2DMs)is a regular chronic disorder that is caused by the secretion of insulin,which leads to serious death-related issues and the most complicated ones.Coronary Heart Disease(CHD)is the most frequent issue related to T2DM patients.The major concern is recognizing the high possibility of CHD complications,yet the model is not available to identify it.This work introduces a deep learning technique that can predict heart disease effectively using a hybrid model,which integrates DNNs(Deep Neural Networks)with a Multi-Head Attention Model called MADNN.The scheme canbedesignedtoautomatically learnthe best-quality features fromElectronic Health Records(EHRs),and effectively combine heterogeneous and time-sequencedmedical data for predicting the risk of CVD.The analysis is done using the Kaggle dataset.The outcomes prove that the MADNN has improved accuracy by about 95%and indicates the precise accuracy is higher for the disease compared with SVM,CNN and ANN.

A life-prediction method for lithium-ion batteries based on a fusion model and an attention mechanism

The current life-prediction models for lithium-ion batteries have several problems, such as the construction of complex feature structures, a high number of feature dimensions, and inaccurate prediction results. To overcome these problems, this paper proposes a deep-learning model combining an autoencoder network and a long short-term memory network. First, this model applies the characteristics of the autoencoder to reduce the dimensionality of the high-dimensional features extracted from the battery data set and realize the fusion of complex time-domain features, which overcomes the problems of redundant model information and low computational efficiency. This model then uses a long short-term memory network that is sensitive to time-series data to solve the long-path dependence problem in the prediction of battery life. Lastly, the attention mechanism is used to give greater weight to features that have a greater impact on the target value, which enhances the learning effect of the model on the long input sequence. To verify the efficacy of the proposed model, this paper uses NASA's lithium-ion battery cycle life data set.

Deep global-attention based convolutional network with dense connections for text classification

Text classification is a classic task innatural language process(NLP).Convolutional neural networks(CNNs)have demonstrated its effectiveness in sentence and document modeling.However,most of existing CNN models are applied to the fixed-size convolution filters,thereby unable to adapt different local interdependency.To address this problem,a deep global-attention based convolutional network with dense connections(DGA-CCN)is proposed.In the framework,dense connections are applied to connect each convolution layer to each of the other layers which can accept information from all previous layers and get multiple sizes of local information.Then the local information extracted by the convolution layer is reweighted by deep global-attention to obtain a sequence representation with more valuable information of the whole sequence.A series of experiments are conducted on five text classification benchmarks,and the experimental results show that the proposed model improves upon the state of-the-art baselines on four of five datasets,which can show the effectiveness of our model for text classification.

Multi-Lead ECG Classification via an Information-Based Attention Convolutional Neural Network

A novel structure based on channel-wise attention mechanism is presented in this paper.With the proposed structure embedded,an efficient classification model that accepts multi-lead electrocardiogram(ECG)as input is constructed.One-dimensional convolutional neural networks(CNNs)have proven to be effective in pervasive classification tasks,enabling the automatic extraction of features while classifying targets.We implement the residual connection and design a structure which can learn the weights from the information contained in different channels in the input feature map during the training process.An indicator named mean square deviation is introduced to monitor the performance of a particular model segment in the classification task on the two out of five ECG classes.The data in the MIT-BIH arrhythmia database is used and a series of control experiments is conducted.Utilizing both leads of the ECG signals as input to the neural network classifier can achieve better classification results than those from using single channel inputs in different application scenarios.Models embedded with the channel-wise attention structure always achieve better scores on sensitivity and precision than the plain Resnet models.The proposed model exceeds most of the state-of-the-art models in ventricular ectopic beats(VEB)classification performance and achieves competitive scores for supraventricular ectopic beats(SVEB).Adopting more lead ECG signals as input can increase the dimensions of the input feature maps,helping to improve both the performance and generalization of the network model.Due to its end-to-end characteristics,and the extensible intrinsic for multi-lead heart diseases diagnosing,the proposed model can be used for the realtime ECG tracking of ECG waveforms for Holter or wearable devices.

Multivariate time series imputation for energy data using neural networks

Multivariate time series with missing values are common in a wide range of applications,including energy data.Existing imputation methods often fail to focus on the temporal dynamics and the cross-dimensional correlation simultaneously.In this paper we propose a two-step method based on an attention model to impute missing values in multivariate energy time series.First,the underlying distribution of the missing values in the data is learned.This information is then further used to train an attention based imputation model.By learning the distribution prior to the imputation process,the model can respond flexibly to the specific characteristics of the underlying data.The developed model is applied to European energy data,obtained from the European Network of Transmission System Operators for Electricity.Using different evaluation metrics and benchmarks,the conducted experiments show that the proposed model is preferable to the benchmarks and is able to accurately impute missing values.

基于双支路特征融合的MRI颅脑肿瘤图像分割研究

针对磁共振成像(magnetic resonance imaging, MRI)颅脑肿瘤区域误识别与分割网络空间信息丢失问题,提出一种基于双支路特征融合的MRI脑肿瘤图像分割方法。首先通过主支路的重构VGG与注意力模型(re-parameterization visual geometry group and attention model, RVAM)提取网络的上下文信息,然后使用可变形卷积与金字塔池化模型(deformable convolution and pyramid pooling model, DCPM)在副支路获取丰富的空间信息,之后使用特征融合模块对两支路的特征信息进行融合。最后引入注意力模型,在上采样过程中加强分割目标在解码时的权重。提出的方法在Kaggle_3m数据集和BraTS2019数据集上进行了实验验证,实验结果表明该方法具有良好的脑肿瘤分割性能,其中在Kaggle_3m上,Dice相似系数、杰卡德系数分别达到了91.45%和85.19%。