GRU Network Research Articles

An Online Control Method of Reactive Power and Voltage Based on Mechanism–Data Hybrid Drive Model Considering Source–Load Uncertainty

The uncertainty brought about by the high proportion of distributed generations poses great challenges to the operational safety of novel distribution systems. Therefore, this paper proposes an online reactive power and voltage control method that integrates source–load uncertainty and a mechanism–data hybrid drive (MDHD) model. Based on the concept of a mechanism and data hybrid drive, the mechanism-driven deterministic reactive power optimization strategy and the stochastic reactive power optimization strategy are used as training data. By training the data-driven CNN–GRU network model offline, the influence of source–load uncertainty on reactive power optimization can be effectively assessed. On this basis, according to the online source and load predicted data, the proposed hybrid-driven model can be applied to quickly obtain the reactive power optimization strategy to enable fast control of voltage. As observed in the case studies, compared with the traditional deterministic and stochastic reactive power optimization models, the hybrid-driven model not only satisfies the real-time requirement of online voltage control, but also has stronger adaptability to source–load uncertainty.

Prediction of Over-the-Horizon Electromagnetic Wave Propagation in Evaporation Ducts Based on the Gated Recurrent Unit Network Model

Evaporation ducts have a significant impact on electromagnetic (EM) wave propagation. The accurate prediction of EM propagation loss (PL) is important for the practical application of EM systems. However, the random weather processes and interaction of air-sea variables will cause complex spatial and temporal fluctuations in PL, especially for an over-the-horizon link. The physics prediction models show some limitations in practice. This article proposes a data-driven PL prediction method based on the gated recurrent unit network (GRU-PL). An X-band EM propagation system was developed and deployed for one month to collect PL of a 53 km link in the South China Sea for model training and testing. The GRU-PL model uses the air-sea variables as inputs and outputs the predicted PL. The root-mean-square error (RMSE) of the GRU-PL model is 7.04 dB, decreasing by 57.2% compared with the classical method. Furthermore, a method, which uses the inversion evaporation duct heights (EDHs) as the output to train the GRU network and then uses the parabolic equation model to predict the PL based on the GRU predicted EDHs, is also discussed. The RMSE is 7.3 dB. This method shows good potential in PL prediction of dynamically changing EM propagation links.

Research on a Lip Reading Algorithm Based on Efficient-GhostNet

Lip reading technology refers to the analysis of the visual information of the speaker’s mouth movements to recognize the content of the speaker’s speech. As one of the important aspects of human–computer interaction, the technology of lip reading has gradually become popular with the development of deep learning in recent years. At present, most the lip reading networks are very complex, with very large numbers of parameters and computation, and the model generated by training needs to occupy large memory, which brings difficulties for devices with limited storage capacity and computation power, such as mobile terminals. Based on the above problems, this paper optimizes and improves GhostNet, a lightweight network, and improves on it by proposing a more efficient Efficient-GhostNet, which achieves performance improvement while reducing the number of parameters through a local cross-channel interaction strategy, without dimensionality reduction. The improved Efficient-GhostNet is used to perform lip spatial feature extraction, and then the extracted features are inputted to the GRU network to obtain the temporal features of the lip sequences, and finally for prediction. We used Asian volunteers for the recording of the dataset in this paper, while also adopting data enhancement for the dataset, using the angle transformation of the dataset to deflect the recording process of the recorder by 15 degrees each to the left and right, in order to be able to enhance the robustness of the network and better reduce the influence of other factors, as well as to improve the generalization ability of the model so that the model can be more consistent with recognition scenarios in real life. Experiments prove that the improved Efficient-GhostNet + GRU model can achieve the purpose of reducing the number of parameters with comparable accuracy.

Battery health prognosis with gated recurrent unit neural networks and hidden Markov model considering uncertainty quantification

• Battery thermoelectric characteristics in the aging process are depicted by multi-features. • The GRU network is trained to fit the global downward trend. • An HMM is developed to fit the local fluctuations for quantifying the uncertainty. • We conduct experiments on two datasets with fixed and random working conditions. With the widespread use of lithium-ion batteries in various fields, battery failures become the most critical concerns that may lead to enormous economic losses and even serious safety issues. The prognostics and health management of lithium-ion batteries helps to ensure reliable and safe battery operations. Existing studies on the state of health of batteries mainly focus on improving and refining prediction models, while the emerging technologies that address uncertainty issues in the battery degradation process are also receiving more and more attention. In this paper, we propose a new state of health prediction method by using the gated recurrent unit neural networks and the hidden Markov model with considering uncertainty quantification. According to the empirical mode decomposition, the battery capacity is decomposed into the global downward trend and the local fluctuations. We train gated recurrent unit neural networks to fit the long-term global downward trend without gradient vanishing, and a hidden Markov model to fit the local fluctuations for quantifying the uncertainty introduced by the capacity recovery phenomenon in battery degradation. Finally, numerical experiments are conducted on two famous datasets, the experimental results demonstrate that the proposed method outperforms on the accuracy and reliability for battery state of health prediction.

BGAT: Aspect-based sentiment analysis based on bidirectional GRU and graph attention network

In today’s social media and various frequently used lifestyle applications, the phenomenon that people express their sentiment via comments or instant barrage is common. People not only show their joys and sorrows in the process of expression but also present their opinions to one thing in many aspects which include. Nowadays, aspect-based sentiment analysis has become a mature and wildly-used technology. There are many public datasets considered as a benchmark to test model performance, such as Laptop2014, Restaurant2014, Twitter, etc. In our work, we also use these public datasets as the test criteria. Current mainstream models generally use the methods of stacking multi-RNNs layers or combining neural networks and BERT or other pre-trained models. On account of the importance displayed by the dependence between aspect words and sentiment words, we investigate a novel model (BGAT) blending bidirectional gated recurrent unit (BiGRU) and relational graph attention network (RGAT) to learn dependencies information. Extensive experiments have been conducted on five datasets, the results demonstrate the great capability of our model.

LCSNet: End-to-end Lipreading with Channel-aware Feature Selection

Lipreading is a task of decoding the movement of the speaker’s lip region into text. In recent years, lipreading methods based on deep neural network have attracted widespread attention, and the accuracy has far surpassed that of experienced human lipreaders. The visual differences in some phonemes are extremely subtle and pose a great challenge to lipreading. Most of the lipreading existing methods do not process the extracted visual features, which mainly suffer from two problems. First, the extracted features contain lot of useless information such as noise caused by differences in speech speed and lip shape, for example. In addition, the extracted features are not abstract enough to distinguish phonemes with similar pronunciation. These problems have a bad effect on the performance of lipreading. To extract features from the lip regions that are more distinguishable and more relevant to the speech content, this article proposes an end-to-end deep neural network-based lipreading model (LCSNet). The proposed model extracts the short-term spatio-temporal features and the motion trajectory features from the lip region in the video clips. The extracted features are filtered by the channel attention module to eliminate the useless features and then used as input to the proposed Selective Feature Fusion Module (SFFM) to extract the high-level abstract features. Afterwards, these features are used as input to the bidirectional GRU network in time order for temporal modeling to obtain the long-term spatio-temporal features. Finally, a Connectionist Temporal Classification (CTC) decoder is used to generate the output text. The experimental results show that the proposed model achieves a 1.0% CER and 2.3% WER on the GRID corpus database, which, respectively, represents an improvement of 52% and 47% compared to LipNet.

The Design of the 1D CNN–GRU Network Based on the RCS for Classification of Multiclass Missiles

For real-time target classification, a study was conducted to improve the AI-based target classification performance using RCS measurements that are vulnerable to noise, but can be obtained quickly. To compensate for the shortcomings of the RCS, a 1D CNN–GRU network with strengths in feature extraction and time-series processing was considered. The 1D CNN–GRU was experimentally changed and designed to fit the RCS characteristics. The performance of the proposed 1D CNN–GRU was compared and analyzed using the 1D CNN and 1D CNN–LSTM. The designed 1D CNN–GRU had the best classification performance with a high accuracy of 99.50% in complex situations, such as with different missile shapes with the same trajectory and with the same missile shapes that had the same trajectory. In addition, to confirm the general target classification performance for the RCS, a new class was verified. The 1D CNN–GRU had the highest classification performance at 99.40%. Finally, as a result of comparing three networks by adding noise to compensate for the shortcomings of the RCS, the 1D CNN–GRU, which was optimized for both the data set used in this paper and the newly constructed data set, was the most robust to noise.

A New Resource Recommendation Method for Experiential Leaching Based on the Completion Degree of Online Learning Tasks

Experiential Learning (ExL) is an effective way to consolidate theoretical knowledge and deepen understandings, and the recommendation of ExL resources needs to also take the effect of students’ theoretical learning into consideration. However, existing studies generally ignore the stage-by-stage assessment of students’ completion of online learning tasks, and the recommendation performance of existing resource recommendation models for ExL is not satisfactory enough. Therefore, the recommendation method needs to be innovated, and the interpretability of recommendation results is facing challenges. To respond to these issues, this paper studied a new resource recommendation method for ExL based on the completion degree of online learning tasks. At first, the paper gave the principle of recommending ExL resources based on the completion degree of online learning tasks, and built an online learning task completion degree prediction model. Then, this paper adopted a bi-directional GRU network model based on attention mechanism to analyze the recent online learning behavior sequence of students and attain the completion degree of students’ short term learning tasks. After that, a knowledge map representing ExL resources and the correlation of knowledge attributes was drawn; by combining the completion degree of both the short-term and long-term learning tasks, the ExL resources suitable for students were recommended to them. At last, experimental results verified the effectiveness of the constructed model.

A Hybrid Model Based on Complete Ensemble Empirical Mode Decomposition With Adaptive Noise, GRU Network and Whale Optimization Algorithm for Wind Power Prediction

A Hybrid Model Based on Complete Ensemble Empirical Mode Decomposition With Adaptive Noise, GRU Network and Whale Optimization Algorithm for Wind Power Prediction

Multi-Level Attention Based Coreference Resolution With Gated Recurrent Unit and Convolutional Neural Networks

Aiming at the diversity of the entity mentions in the field of animal husbandry in Tibet, the reference resolution model Att-GRU-CNN based on multi-level attention mechanism is proposed. The model uses GRU network for global semantic feature learning and knowledge memory, and uses CNN network to further extract local high-level semantic features. A word-level attention layer is added on the GRU hidden layer, the feature vector of entity mention does dot product operation with the feature vector of each word, and the result is normalized and used as the weight value distribution of the words, thus, the prior knowledge that the mention is the most important in the context is given to the network; A sentence-level attention layer is added on the CNN convolution layer, the convolutional layer output and entity name library matrix do association operation, and the result is used as the weight value distribution of the sentence, so as to strengthen the relation between the sentence where the mention is located and the entity scientific name. At the same time, the data enhancement technology is used to improve the generalization ability of the model. Finally, the ablation experiment on Tibetan animal husbandry dataset verified the effectiveness of each component of the model and the superposition effect after combination; the performance comparison experiment is carried out on public datasets MUC, B3 and CEAF <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$\varphi _{4}$ </tex-math></inline-formula> . The experimental results show that this model has a significant improvement over other models in the accuracy rate, recall rate and F1 score.

Driver Intent-Based Intersection Autonomous Driving Collision Avoidance Reinforcement Learning Algorithm

With the rapid development of artificial intelligent technology, the deep learning method is widely applied to predict human driving intentions due to its relative accuracy of prediction, which is one of critical links for security guarantee in the distributed, mixed driving scenario. In order to sense the intention of human-driven vehicles and reduce the self-driving collision avoidance rate, an improved intention prediction method for human-driving vehicles based on unsupervised, deep inverse reinforcement learning is proposed. Firstly, a contrast discriminator module was proposed to extract richer features. Then, the residual module was created to overcome the drawbacks of gradient disappearance and network degradation with the increase in network layers. Furthermore, the dropout layer was generated to prevent the over-fitting phenomenon in the whole training process of the GRU network, so as to improve the generalization ability of the network model. Finally, abundant experiments were conducted on datasets to evaluate our proposed method. The pass rate of self-driving vehicles with conservative driver probabilities of p = 0.25, p = 0.4, and p = 0.6 improved by a maximum of 8%, 10%, and 3%, compared with the classical method LSTM and VAE + RNN. It indicates that the prediction results of our proposed method fit more with the basic structure of the given traffic scenario in a long-term prediction range, which verifies the effectiveness of our proposed method.

An attention-based bidirectional GRU network for temporal action proposals generation

An attention-based bidirectional GRU network for temporal action proposals generation

A fuzzy convolutional attention-based GRU network for human activity recognition

Human activity recognition has become a pillar of today intelligent Human–Computer Interfaces as it typically provides more comfortable and ubiquitous interaction. This paper proposes a novel fuzzy-based deep learning-based algorithm to predict future sequences of activities from a given sequence of daily living activities of a subject wearing a lower limb exoskeleton. The engineering application concerns the challenging task of recognizing locomotion activities of the wearer in real-time, which is needed to ensure appropriate control of the robot during daily living activities. Indeed, real-time locomotion activity recognition is very challenging for controlling lower-limb exoskeletons. The model proposes a new adaptive kernel, based on the data features derived from the fuzzy rules on the input sequences to enrich the features of the activity sequences. Then, a CNN is applied to extract local subsequences from the whole sequences to identify local patterns in the convolution window. Finally, an attention-based GRU is incorporated into the model to extract meaningful parts of the time-series sequences. The results show high accuracy in the estimation of the transition between gait modes which is critical to ensure smooth control of the exoskeleton. The performance of the model is evaluated using the dynamic activity data gathered from different subjects. The proposed model outperforms the traditional models used in the literature.

Detecting Nontechnical Losses in Smart Meters Using a MLP-GRU Deep Model and Augmenting Data via Theft Attacks

The current study uses a data-driven method for Nontechnical Loss (NTL) detection using smart meter data. Data augmentation is performed using six distinct theft attacks on benign users’ samples to balance the data from honest and theft samples. The theft attacks help to generate synthetic patterns that mimic real-world electricity theft patterns. Moreover, we propose a hybrid model including the Multi-Layer Perceptron and Gated Recurrent Unit (MLP-GRU) networks for detecting electricity theft. In the model, the MLP network examines the auxiliary data to analyze nonmalicious factors in daily consumption data, whereas the GRU network uses smart meter data acquired from the Pakistan Residential Electricity Consumption (PRECON) dataset as the input. Additionally, a random search algorithm is used for tuning the hyperparameters of the proposed deep learning model. In the simulations, the proposed model is compared with the MLP-Long Term Short Memory (LSTM) scheme and other traditional schemes. The results show that the proposed model has scores of 0.93 and 0.96 for the area under the precision–recall curve and the area under the receiver operating characteristic curve, respectively. The precision–recall curve and the area under the receiver operating characteristic curve scores for the MLP-LSTM are 0.93 and 0.89, respectively.

Parameter Optimization of Educational Network Ecosystem Based on BERT Deep Learning Model

The key sentimental words in the text cannot be paid attention to effectively, and language knowledge such as the text information and the sentimental resources are relied on. Therefore, it is necessary to make full use of this unique sentimental information to achieve the best performance of the model. In order to solve the problems, a method based on the fusion of the convolutional neural network and the bidirectional GRU network text sentiment analysis capsule model to analyze the ideological and political education of public opinion is put forward. In this model, each sentiment category is combined with the attention mechanism to generate feature vectors to construct sentiment capsules. Finally, the text sentiment categories are judged according to the attributes of the capsules. The model is tested on MR, IMDB, SST-5, and the data set of the ideological and political education review. Experimental results show that compared with MC-CNN-LSTM, the readiness rate of the proposed model is improved by 5.1%, 2.8%, 2.8%, and 1.6% on four public Chinese and English data sets, respectively. Compared with LR-Bi-LSTM, NSCL, and multi-Bi-LSTM models, the accuracy of the proposed MC-BiGRU-Capsule model on MR and SST-5 data are 3.2%, 2.4%, and 3.4% higher than that of the LR-Bi-LSTM, NSCL, and multi-Bi-LSTM models, respectively. It also shows a better classification effect on multiclassification data sets. It is concluded that compared with other baseline models, this method has a better classification effect.

Decoupled Object-Independent Image Features for Fine Phasing of Segmented Mirrors Using Deep Learning

A segmented primary mirror is very important for extra-large astronomical telescopes, in order to detect the phase error between segmented mirrors. Traditional iterative algorithms are hard to detect co−phasing aberrations in real time due to the long-time iterative process. Deep learning has shown large potential in wavefront sensing, and it gradually focuses on detecting piston error. However, the current methods based on deep learning are mainly applied to coarse phase sensing, and only consider the detection of piston error with no tip/tilt errors, which is inconsistent with reality. In this paper, by innovatively designing the form of pupil mask, and further updating the OTF in the frequency domain, we obtain a new decoupled independent feature image that can simultaneously detect the piston error and tilt/tilt error of all sub-mirrors, which is effectively decoupled, and eliminates the dependence of the data set on the imaging object. Then, the Bi−GRU network is used to recover phase error information with high accuracy from the feature image proposed in this paper. The network’s detection accuracy ability is verified under single wavelength and broadband spectrum in simulation. This paper demonstrates that co−phasing errors can be accurately decoupled and extracted by the new feature image we proposed and will contribute to the fine phasing accuracy and practicability of the extended scenes for the segmented telescopes.

Short-Term Prediction Method of Solar Photovoltaic Power Generation Based on Machine Learning in Smart Grid

In order to improve the accuracy of ultra short-term power prediction of the photovoltaic power generation system, a short-term photovoltaic power prediction method based on an adaptive k-means and Gru machine learning model is proposed. This method first introduces the construction process of the model and then builds a short-term photovoltaic power generation prediction model based on an adaptive k-means and Gru machine learning models. Then, the network structure and key parameters are determined through experiments, and the initial training set of the prediction model is selected according to the short-term photovoltaic power generation characteristics. And the adaptive k-means is used to cluster the initial training set and the photovoltaic power on the forecast day. The Gru model is trained on the initial training set data of each category, and the generated power is predicted in combination with the trained Gru model. Finally, considering three typical weather types, the proposed method is used for simulation analysis and compared with the other three traditional photovoltaic power generation single prediction models. The comparison results show that the proposed short-term photovoltaic power generation prediction method based on an adaptive k-means and Gru network has better effect, better robustness, and less error.

Joint Event Relation Identification Based on Multiscale Convolutional Neural Network and Sharing Strategy

At present, most of the event relation identification work mainly focuses on the sequential temporal and explicit causal relation between events. These methods usually ignore the role of synchronous temporal and implicit causal relation in sentences, which makes the semantic understanding of the model deviate from the text. In this paper, we propose a joint event relation identification model. The model uses bidirectional GRU and multiscale convolutional neural network to obtain the context semantic features and multiscale local semantic features of text, respectively. Then, these two kinds of features are fused to fully obtain the semantics of the text itself. In addition, we build encoders and decoders of event temporal and causal relation, respectively, to obtain the event temporal and causal semantic features from text. In this process, considering the correlation between event timing and causality, we use three different parameter sharing strategies to realize the interaction between event temporal and causal semantic features. The experimental results on the legal field dataset we constructed show that our model has made significant improvements compared with the baseline model. Through experimental analysis, our method can effectively improve the identification performance of synchronous temporal and implicit causality relation.

Mitigation of Ice-Induced Vibration of Offshore Platform Based on Gated Recurrent Neural Network

Ice-induced vibration is one of the major risks that face the offshore platform located in cold regions. In this paper, the gated recurrent neural network (GRNN) is utilized to predict and suppress the response of offshore platforms subjected to ice load. First, a simplified model of the offshore platform is derived and validated based on the finite element model (FEM). The time history of the floating ice load is generated using the harmonic superposition method. Gated Recurrent Unit Network (GRU) and the Long-Short-Term Memory Network (LSTM) are composed in MATLAB to predict the behavior of the off-shore platform. Afterward, the linear quadratic regulator (LQR) control algorithm is used to calculate the controlling force for the training of the GRU/LSTM-based prediction controller. Numerical results show that the ice-induced vibration response prediction method based on GRU network design can predict the structural response with satisfying accuracy, and the ice-induced vibration response control method based on the LSTM network and GRU network design can learn the LQR method well and achieve good control effect. Time lag and other problems that the vibration control programs often encountered were solved well.

Remaining useful life prediction of bearings by a new reinforced memory GRU network

The remaining useful life (RUL) prediction of bearings has great significance in the predictive maintenance of mechanical equipment. Owing to the difficulty of collecting abundant lifecycle datasets with correct labels, it is quite necessary to explore a prediction method with high precision and robustness in the case of small samples. It follows that a novel RUL prediction approach is put forward to overcome this problem. First, for reducing the man-made interference and the demand for expert knowledge, an unsupervised health indicator (HI) is constructed by Gaussian mixture model (GMM) and Kullback-Leibler divergence (KLD), which is named as KLD-based HI. Then because of the rapid forgetting of historical trend information in the current RNN-based prediction models, a novel reinforced memory gated recurrent unit (RMGRU) network is proposed by reusing the state information at the previous moment. According to the constructed KLD-based HI vector, the unknown HIs are successively predicted by RMGRU until the predicted HI value exceeds the failure threshold, and then RUL is calculated. The contrast experiment on IEEE 2012PHM bearing datasets shows the superiority of the bearing RUL prediction approach based on RMGRU over the classical time series forecasting methods. It can be concluded that this method has great application potential in bearing RUL prediction.