Long Short-Term Memory Recurrent Neural Network-Based Acoustic Model Using Connectionist Temporal Classification on a Large-Scale Training Corpus

A Long Short-Term Memory(LSTM) Recurrent Neural Network(RNN) has driven tremendous improvements on an acoustic model based on Gaussian Mixture Model(GMM). However, these models based on a hybrid method require a forced aligned Hidden Markov Model(HMM) state sequence obtained from the GMM-based acoustic model. Therefore, it requires a long computation time for training both the GMM-based acoustic model and a deep learning-based acoustic model. In order to solve this problem, an acoustic model using CTC algorithm is proposed. CTC algorithm does not require the GMM-based acoustic model because it does not use the forced aligned HMM state sequence. However, previous works on a LSTM RNN-based acoustic model using CTC used a small-scale training corpus. In this paper, the LSTM RNN-based acoustic model using CTC is trained on a large-scale training corpus and its performance is evaluated. The implemented acoustic model has a performance of 6.18% and 15.01% in terms of Word Error Rate(WER) for clean speech and noisy speech, respectively. This is similar to a performance of the acoustic model based on the hybrid method.

Conditional Random Field Tracking Model Based on a Visual Long Short Term Memory Network

In dense pedestrian tracking,frequent object occlusions and close distances between objects cause difficulty when accurately estimating object trajectories.In this study,a conditional random field tracking model is established by using a visual long short term memory network in the three-dimensional(3D)space and the motion estimations jointly performed on object trajectory segments.Object visual field information is added to the long short term memory network to improve the accuracy of the motion related object pair selection and motion estimation.To address the uncertainty of the length and interval of trajectory segments,a multimode long short term memory network is proposed for the object motion estimation.The tracking performance is evaluated using the PETS2009 dataset.The experimental results show that the proposed method achieves better performance than the tracking methods based on the independent motion estimation.

Binaural Speech Separation Algorithm Based on Long and Short Time Memory Networks

Speaker separation in complex acoustic environment is one of challenging tasks in speech separation.In practice,speakers are very often unmoving or moving slowly in normal communication.In this case,the spatial features among the consecutive speech frames become highly correlated such that it is helpful for speaker separation by providing additional spatial information.To fully exploit this information,we design a separation system on Recurrent Neural Network(RNN)with long short-term memory(LSTM)which effectively learns the temporal dynamics of spatial features.In detail,a LSTM-based speaker separation algorithm is proposed to extract the spatial features in each time-frequency(TF)unit and form the corresponding feature vector.Then,we treat speaker separation as a supervised learning problem,where a modified ideal ratio mask(IRM)is defined as the training function during LSTM learning.Simulations show that the proposed system achieves attractive separation performance in noisy and reverberant environments.Specifically,during the untrained acoustic test with limited priors,e.g.,unmatched signal to noise ratio(SNR)and reverberation,the proposed LSTM based algorithm can still outperforms the existing DNN based method in the measures of PESQ and STOI.It indicates our method is more robust in untrained conditions.

一种基于long short-term memory的唇语识别方法

唇动视觉信息是说话内容的重要载体。受嘴唇外观、背景信息和说话习惯等影响,即使说话者说相同的内容,唇动视觉信息也会相差很大。为解决唇语视觉信息多样性的问题,提出一种基于long short-term memory(LSTM)的新的唇语识别方法。以往大多数的方法从嘴唇外表信息入手。本方法用嘴唇关键点坐标描述嘴唇形变信息作为唇语视频的特征,它具有类内一致性和类间区分性的特点。然后利用LSTM对特征进行时序编码,它能学习具有区分性和泛化性的空间-时序特征。在公开的唇语数据集GRID、MIRACL-VC和Oulu VS上对本方法做了针对分割的单词或短语的说话者独立的唇语识别评估。在GRID和MIRACL-VC上,本方法的准确率比传统方法至少高30%;在Oulu VS上,本方法的准确率接近于最优结果。以上实验结果表明,本文提出的基于LSTM的唇语识别方法有效地解决了唇语视觉信息多样性的问题。

Tool Health Condition Recognition Method for High Speed Milling of Titanium Alloy Based on Principal Component Analysis (PCA) and Long Short Term Memory (LSTM)

The healthy condition of the milling tool has a very high impact on the machining quality of the titanium components.Therefore,it is important to recognize the healthy condition of the tool and replace the damaged cutter at the right time.In order to recognize the health condition of the milling cutter,a method based on the long short term memory(LSTM)was proposed to recognize tool health state in this paper.The various signals collected in the tool wear experiments were analyzed by time-domain statistics,and then the extracted data were generated by principal component analysis(PCA)method.The preprocessed data extracted by PCA is transmitted to the LSTM model for recognition.Compared with back propagation neural network(BPNN)and support vector machine(SVM),the proposed method can effectively utilize the time-domain regulation in the data to achieve higher recognition speed and accuracy.

Short-Term Relay Quality Prediction Algorithm Based on Long and Short-Term Memory

The fraction defective of semi-finished products is predicted to optimize the process of relay production lines, by which production quality and productivity are increased, and the costs are decreased. The process parameters of relay production lines are studied based on the long-and-short-term memory network. Then, the Keras deep learning framework is utilized to build up a short-term relay quality prediction algorithm for the semi-finished product. A simulation model is used to study prediction algorithm. The simulation results show that the average prediction absolute error of the fraction is less than 5%. This work displays great application potential in the relay production lines.

Predicting and Curing Depression Using Long Short Term Memory and Global Vector

In today’s world, there are many people suffering from mentalhealth problems such as depression and anxiety. If these conditions are notidentified and treated early, they can get worse quickly and have far-reachingnegative effects. Unfortunately, many people suffering from these conditions,especially depression and hypertension, are unaware of their existence until theconditions become chronic. Thus, this paper proposes a novel approach usingBi-directional Long Short-Term Memory (Bi-LSTM) algorithm and GlobalVector (GloVe) algorithm for the prediction and treatment of these conditions.Smartwatches and fitness bands can be equipped with these algorithms whichcan share data with a variety of IoT devices and smart systems to betterunderstand and analyze the user’s condition. We compared the accuracy andloss of the training dataset and the validation dataset of the two modelsnamely, Bi-LSTM without a global vector layer and with a global vector layer.It was observed that the model of Bi-LSTM without a global vector layer hadan accuracy of 83%,while Bi-LSTMwith a global vector layer had an accuracyof 86% with a precision of 86.4%, and an F1 score of 0.861. In addition toproviding basic therapies for the treatment of identified cases, our model alsohelps prevent the deterioration of associated conditions, making our methoda real-world solution.

Short-TermWind Power Prediction Based on Combinatorial Neural Networks

Wind power volatility not only limits the large-scale grid connection but also poses many challenges to safe grid operation.Accurate wind power prediction can mitigate the adverse effects of wind power volatility on wind power grid connections.For the characteristics of wind power antecedent data and precedent data jointly to determine the prediction accuracy of the prediction model,the short-term prediction of wind power based on a combined neural network is proposed.First,the Bi-directional Long Short Term Memory(BiLSTM)network prediction model is constructed,and the bi-directional nature of the BiLSTM network is used to deeply mine the wind power data information and find the correlation information within the data.Secondly,to avoid the limitation of a single prediction model when the wind power changes abruptly,the Wavelet Transform-Improved Adaptive Genetic Algorithm-Back Propagation(WT-IAGA-BP)neural network based on the combination of the WT-IAGA-BP neural network and BiLSTM network is constructed for the short-term prediction of wind power.Finally,comparing with LSTM,BiLSTM,WT-LSTM,WT-BiLSTM,WT-IAGA-BP,and WT-IAGA-BP&LSTM prediction models,it is verified that the wind power short-term prediction model based on the combination of WT-IAGA-BP neural network and BiLSTM network has higher prediction accuracy.

Analyses of fear memory in Arc/Arg3.1-deficient mice: intact short-term memory and impaired long-term and remote memory

Activity-regulated cytoskeleton-associated protein (Arc/Arg3.1) was originally identified in patients with seizures. It is densely distributed in the hip-pocampus and amygdala in particular. Because the expression of Arc/Arg3.1 is regulated by nerve in-puts, it is thought to be an immediate early gene. As shown both in vitro and in vivo, Arc/Arg3.1 is in-volved in synaptic consolidation and regulates some forms of learning and memory in rats and mice [1,2]. Furthermore, a recent study suggests that Arc/Arg3.1 may play a significant role in signal transmission via AMPA-type glutamate receptors [3-5]. Therefore, we conducted a detailed analysis of fear memory in Arc/Arg3.1-deficient mice. As previously reported, the knockout animals exhib-ited impaired fear memory in both contextual and cued test situations. Although Arc/Arg3.1-deficient mice showed almost the same performance as wild-type littermates 4 hr after a conditioning trial, their performance was impaired in the retention test after 24 hr or longer, either with or without reconsolidation. Immunohistochemical analyses showed an abnormal density of GluR1 in the hip-pocampus of Arc/Arg3.1-deficient mice;however, an application of AMPA potentiator did not improve memory performance in the mutant mice. Memory impairment in Arc/Arg3.1-deficient mice is so ro-bust that the mice provide a useful tool for devel-oping treatments for memory impairment.

State of Health Estimation of Lithium-Ion Batteries Using Support Vector Regression and Long Short-Term Memory

Lithium-ion batteries are the most widely accepted type of battery in the electric vehicle industry because of some of their positive inherent characteristics. However, the safety problems associated with inaccurate estimation and prediction of the state of health of these batteries have attracted wide attention due to the adverse negative effect on vehicle safety. In this paper, both machine and deep learning models were used to estimate the state of health of lithium-ion batteries. The paper introduces the definition of battery health status and its importance in the electric vehicle industry. Based on the data preprocessing and visualization analysis, three features related to actual battery capacity degradation are extracted from the data. Two learning models, SVR and LSTM were employed for the state of health estimation and their respective results are compared in this paper. The mean square error and coefficient of determination were the two metrics for the performance evaluation of the models. The experimental results indicate that both models have high estimation results. However, the metrics indicated that the SVR was the overall best model.

Comparison of Two Recurrent Neural Networks for Rainfall-Runoff Modeling in the Zou River Basin at Atchérigbé (Bénin)

Hydrological models are developed to simulate river flows over a watershed for many practical applications in the field of water resource management. The present paper compares the performance of two recurrent neural networks for rainfall-runoff modeling in the Zou River basin at Atchérigbé outlet. To this end, we used daily precipitation data over the period 1988-2010 as input of the models, such as the Long Short-Term Memory (LSTM) and Recurrent Gate Networks (GRU) to simulate river discharge in the study area. The investigated models give good results in calibration (R2 = 0.888, NSE = 0.886, and RMSE = 0.42 for LSTM;R2 = 0.9, NSE = 0.9 and RMSE = 0.397 for GRU) and in validation (R2 = 0.865, NSE = 0.851, and RMSE = 0.329 for LSTM;R2 = 0.9, NSE = 0.865 and RMSE = 0.301 for GRU). This good performance of LSTM and GRU models confirms the importance of models based on machine learning in modeling hydrological phenomena for better decision-making.

Ensembling Neural Networks for User’s Indoor Localization Using Magnetic Field Data from Smartphones

Predominantly the localization accuracy of the magnetic field-based localization approaches is severed by two limiting factors:Smartphone heterogeneity and smaller data lengths.The use of multifarioussmartphones cripples the performance of such approaches owing to the variability of the magnetic field data.In the same vein,smaller lengths of magnetic field data decrease the localization accuracy substantially.The current study proposes the use of multiple neural networks like deep neural network(DNN),long short term memory network(LSTM),and gated recurrent unit network(GRN)to perform indoor localization based on the embedded magnetic sensor of the smartphone.A voting scheme is introduced that takes predictions from neural networks into consideration to estimate the current location of the user.Contrary to conventional magnetic field-based localization approaches that rely on the magnetic field data intensity,this study utilizes the normalized magnetic field data for this purpose.Training of neural networks is carried out using Galaxy S8 data while the testing is performed with three devices,i.e.,LG G7,Galaxy S8,and LG Q6.Experiments are performed during different times of the day to analyze the impact of time variability.Results indicate that the proposed approach minimizes the impact of smartphone variability and elevates the localization accuracy.Performance comparison with three approaches reveals that the proposed approach outperforms them in mean,50%,and 75%error even using a lesser amount of magnetic field data than those of other approaches.

A Novel MegaBAT Optimized Intelligent Intrusion Detection System in Wireless Sensor Networks

Wireless Sensor Network(WSN),whichfinds as one of the major components of modern electronic and wireless systems.A WSN consists of numerous sensor nodes for the discovery of sensor networks to leverage features like data sensing,data processing,and communication.In thefield of medical health care,these network plays a very vital role in transmitting highly sensitive data from different geographic regions and collecting this information by the respective network.But the fear of different attacks on health care data typically increases day by day.In a very short period,these attacks may cause adversarial effects to the WSN nodes.Furthermore,the existing Intrusion Detection System(IDS)suffers from the drawbacks of limited resources,low detection rate,and high computational overhead and also increases the false alarm rates in detecting the different attacks.Given the above-mentioned problems,this paper proposes the novel MegaBAT optimized Long Short Term Memory(MBOLT)-IDS for WSNs for the effective detection of different attacks.In the proposed framework,hyperpara-meters of deep Long Short-Term Memory(LSTM)were optimized by the meta-heuristic megabat algorithm to obtain a low computational overhead and high performance.The experimentations have been carried out using(Wireless Sensor NetworkDetection System)WSN-DS datasets and performance metrics such as accuracy,recall,precision,specificity,and F1-score are calculated and compared with the other existing intelligent IDS.The proposed framework provides outstanding results in detecting the black hole,gray hole,scheduling,flooding attacks and significantly reduces the time complexity,which makes this system suitable for resource-constraint WSNs.

Synthetic well logs generation via Recurrent Neural Networks

To supplement missing logging information without increasing economic cost, a machine learning method to generate synthetic well logs from the existing log data was presented, and the experimental verification and application effect analysis were carried out. Since the traditional Fully Connected Neural Network(FCNN) is incapable of preserving spatial dependency, the Long Short-Term Memory(LSTM) network, which is a kind of Recurrent Neural Network(RNN), was utilized to establish a method for log reconstruction. By this method, synthetic logs can be generated from series of input log data with consideration of variation trend and context information with depth. Besides, a cascaded LSTM was proposed by combining the standard LSTM with a cascade system. Testing through real well log data shows that: the results from the LSTM are of higher accuracy than the traditional FCNN; the cascaded LSTM is more suitable for the problem with multiple series data; the machine learning method proposed provides an accurate and cost effective way for synthetic well log generation.

Bitcoin Candlestick Prediction with Deep Neural Networks Based on Real Time Data

Currently,Bitcoin is the world’s most popular cryptocurrency.The price of Bitcoin is extremely volatile,which can be described as high-benefit and high-risk.To minimize the risk involved,a means of more accurately predicting the Bitcoin price is required.Most of the existing studies of Bitcoin prediction are based on historical(i.e.,benchmark)data,without considering the real-time(i.e.,live)data.To mitigate the issue of price volatility and achieve more precise outcomes,this study suggests using historical and real-time data to predict the Bitcoin candlestick—or open,high,low,and close(OHLC)—prices.Seeking a better prediction model,the present study proposes time series-based deep learning models.In particular,two deep learning algorithms were applied,namely,long short-term memory(LSTM)and gated recurrent unit(GRU).Using real-time data,the Bitcoin candlesticks were predicted for three intervals:the next 4 h,the next 12 h,and the next 24 h.The results showed that the best-performing model was the LSTM-based model with the 4-h interval.In particular,this model achieved a stellar performance with a mean absolute percentage error(MAPE)of 0.63,a root mean square error(RMSE)of 0.0009,a mean square error(MSE)of 9e-07,a mean absolute error(MAE)of 0.0005,and an R-squared coefficient(R2)of 0.994.With these results,the proposed prediction model has demonstrated its efficiency over the models proposed in previous studies.The findings of this study have considerable implications in the business field,as the proposed model can assist investors and traders in precisely identifying Bitcoin sales and buying opportunities.

Short Term Traffic Flow Prediction Using Hybrid Deep Learning

Traffic flow prediction in urban areas is essential in the IntelligentTransportation System (ITS). Short Term Traffic Flow (STTF) predictionimpacts traffic flow series, where an estimation of the number of vehicleswill appear during the next instance of time per hour. Precise STTF iscritical in Intelligent Transportation System. Various extinct systems aim forshort-term traffic forecasts, ensuring a good precision outcome which was asignificant task over the past few years. The main objective of this paper is topropose a new model to predict STTF for every hour of a day. In this paper,we have proposed a novel hybrid algorithm utilizing Principal ComponentAnalysis (PCA), Stacked Auto-Encoder (SAE), Long Short Term Memory(LSTM), and K-Nearest Neighbors (KNN) named PALKNN. Firstly, PCAremoves unwanted information from the dataset and selects essential features.Secondly, SAE is used to reduce the dimension of input data using onehotencoding so the model can be trained with better speed. Thirdly, LSTMtakes the input from SAE, where the data is sorted in ascending orderbased on the important features and generates the derived value. Finally,KNN Regressor takes information from LSTM to predict traffic flow. Theforecasting performance of the PALKNN model is investigated with OpenRoad Traffic Statistics dataset, Great Britain, UK. This paper enhanced thetraffic flow prediction for every hour of a day with a minimal error value.An extensive experimental analysis was performed on the benchmark dataset.The evaluated results indicate the significant improvement of the proposedPALKNN model over the recent approaches such as KNN, SARIMA, LogisticRegression, RNN, and LSTM in terms of root mean square error (RMSE)of 2.07%, mean square error (MSE) of 4.1%, and mean absolute error (MAE)of 2.04%.

Comparative study on the performance of ConvLSTM and ConvGRU in classification problems-taking early warning of short-duration heavy rainfall as an example

卷积长短期记忆单元ConvLSTM和卷积门控循环单元ConvGRU是两种广泛应用的深度学习单元,通过将循环机制与卷积运算相结合,常常用于时空序列的预测.为了明确上述两种模型的收敛速度和分类能力,需要使用相同的模型架构对相同的分类问题进行预测.本研究将北京短时强降水区级预警问题看作深度学习中的二分类问题,使用京津冀雷达网的组合反射率数据和北京区域内的自动气象站降雨数据进行深度学习模型的训练和评估.结果表明,ConvGRU的收敛速度比ConvLSTM快约25%.ConvLSTM和ConvGRU的预警性能随地区,时间,降雨强度的变化趋势相似,但大部分ConvLSTM的得分较高,少数情况下ConvGRU的得分较高.

Deep Learning Network for Energy Storage Scheduling in Power Market Environment Short-Term Load Forecasting Model

In the electricity market,fluctuations in real-time prices are unstable,and changes in short-term load are determined by many factors.By studying the timing of charging and discharging,as well as the economic benefits of energy storage in the process of participating in the power market,this paper takes energy storage scheduling as merely one factor affecting short-term power load,which affects short-term load time series along with time-of-use price,holidays,and temperature.A deep learning network is used to predict the short-term load,a convolutional neural network(CNN)is used to extract the features,and a long short-term memory(LSTM)network is used to learn the temporal characteristics of the load value,which can effectively improve prediction accuracy.Taking the load data of a certain region as an example,the CNN-LSTM prediction model is compared with the single LSTM prediction model.The experimental results show that the CNN-LSTM deep learning network with the participation of energy storage in dispatching can have high prediction accuracy for short-term power load forecasting.

基于改进LSTM的数码雷管模组印刷质量预测

由于数码雷管模组印刷过程中生产工艺复杂、强时序性等特点,其质量的精准预测已成为提高产品质量管理水平的关键。基于此提出一种改进长短期记忆(long short-term memory,LSTM)网络的数码雷管模组印刷质量预测模型。首先根据数码雷管模组印刷过程提炼机器运行参数、环境参数与检测参数作为印刷产品质量的原始特征,并对关键检测参数进行时序特征重构以增强特征表达能力;其次基于改进的LSTM网络建立数码雷管模组印刷特征提取框架,采用卷积神经网络提取空间特征避免LSTM挖掘高维印刷特征时隐含关系的不足,通过全局注意力机制自适应学习不同时刻印刷特征对印刷产品质量的贡献度,为LSTM提取的深层时序特征分配不同权值;最后以深层特征作为输入,通过全连接网络实现数码雷管模组印刷产品的质量预测。实验结果表明,相较于BP神经网络、门控循环单元网络、LSTM等预测方法,改进的LSTM网络有效提高了数码雷管模组印刷产品质量的预测精度。

基于ARIMA-LSTM的矿区地表沉降预测方法

煤矿开采安全问题尤其是采空区地表沉降现象会对人员安全及工程安全造成威胁,研究合适的矿区地表沉降预测方法具有很大意义。矿区地表沉降影响因素复杂,单一的深度学习模型对矿区地表沉降数据拟合效果差且现有的地表沉降预测研究多是单独进行概率预测或考虑时序特性进行点预测,难以在考虑数据的时序特征的同时对其随机性进行定量描述。针对此问题,在对数据本身性质进行观察分析后选择差分整合移动平均自回归(ARIMA)模型进行时序特征的概率预测,结合长短时记忆(LSTM)网络模型来学习复杂的且具有长期依赖性的非线性时序特征。提出基于ARIMA-LSTM的地表沉降预测模型,利用ARIMA模型对数据的时序线性部分进行预测,并将ARIMA模型预测的残差数据辅助LSTM模型训练,在考虑时序特征的同时对数据的随机性进行描述。研究结果表明,相较于单独采用ARIMA或LSTM模型,该方法具有更高的预测精度(MSE为0.262 87,MAE为0.408 15,RMSE为0.512 71)。进一步的对比结果显示,预测结果与雷达卫星影像数据(经SBAS-INSAR处理后)趋势一致,证实了该方法的有效性。