Early identification of patients at risk for iron-deficiency anemia using deep learning techniques

OBJECTIVES

Iron-deficiency anemia (IDA) is a common health problem worldwide, and up to 10% of adult patients with incidental IDA may have gastrointestinal cancer. A diagnosis of IDA can be established through a combination of laboratory tests, but it is often underrecognized until a patient becomes symptomatic. Based on advances in machine learning, we hypothesized that we could reduce the time to diagnosis by developing an IDA prediction model. Our goal was to develop 3 neural networks by using retrospective longitudinal outpatient laboratory data to predict the risk of IDA 3 to 6 months before traditional diagnosis.

METHODS

We analyzed retrospective outpatient electronic health record data between 2009 and 2020 from an academic medical center in northern Texas. We included laboratory features from 30,603 patients to develop 3 types of neural networks: artificial neural networks, long short-term memory cells, and gated recurrent units. The classifiers were trained using the Adam Optimizer across 200 random training-validation splits. We calculated accuracy, area under the receiving operating characteristic curve, sensitivity, and specificity in the testing split.

RESULTS

Although all models demonstrated comparable performance, the gated recurrent unit model outperformed the other 2, achieving an accuracy of 0.83, an area under the receiving operating characteristic curve of 0.89, a sensitivity of 0.75, and a specificity of 0.85 across 200 epochs.

CONCLUSIONS

Our results showcase the feasibility of employing deep learning techniques for early prediction of IDA in the outpatient setting based on sequences of laboratory data, offering a substantial lead time for clinical intervention.

Combining data augmentation and deep learning for improved epilepsy detection

Introduction

In recent years, the use of EEG signals for seizure detection has gained widespread academic attention. Aiming at the problem of overfitting deep learning models due to the small number of EEG signal data during epilepsy detection, this paper proposes an epilepsy detection method that combines data augmentation and deep learning.

Methods

First, the Adversarial and Mixup Data Augmentation (AMDA) method is used to realize the data augmentation, which effectively enriches the number of training samples. To further improve the classification accuracy and robustness of epilepsy detection, this paper proposes a one-dimensional convolutional neural network and gated recurrent unit (AM-1D CNN-GRU) network model based on attention mechanism for epilepsy detection.

Results and discussion

The experimental results show that the performance of epilepsy detection achieved by using augmented data is significantly improved, and the accuracy, sensitivity, and area under the subject's working characteristic curve are up to 96.06, 95.48%, and 0.9637, respectively. Compared with the non-augmented data, all indicators are increased by more than 6.2%. Meanwhile, the detection performance was significantly improved compared with other epilepsy detection methods. The results of this research can provide a reference for the clinical application of epilepsy detection.

DeepPRMS: advanced deep learning model to predict protein arginine methylation sites

Protein methylation is a form of post-translational modifications of protein, which is crucial for various cellular processes, including transcription activity and DNA repair. Correctly predicting protein methylation sites is fundamental for research and drug discovery. Some experimental techniques, such as methyl-specific antibodies, chromatin immune precipitation and mass spectrometry, exist for predicting protein methylation sites, but these techniques are time-consuming and costly. The ability to predict methylation sites using in silico techniques may help researchers identify potential candidate sites for future examination and make it easier to carry out site-specific investigations and downstream characterizations. In this research, we proposed a novel deep learning-based predictor, named DeepPRMS, to identify protein methylation sites in primary sequences. The DeepPRMS utilizes the gated recurrent unit (GRU) and convolutional neural network (CNN) algorithms to extract the sequential and spatial information from the primary sequences. GRU is used to extract sequential information, while CNN is used for spatial information. We combined the latent representation of GRU and CNN models to have a better interaction among them. Based on the independent test data set, DeepPRMS obtained an accuracy of 85.32%, a specificity of 84.94%, Matthew's correlation coefficient of 0.71 and a sensitivity of 85.80%. The results indicate that DeepPRMS can predict protein methylation sites with high accuracy and outperform the state-of-the-art models. The DeepPRMS is expected to effectively guide future research experiments for identifying potential methylated protein sites. The web server is available at http://deepprms.nitsri.ac.in/.

Melanoma skin cancer detection using mask-RCNN with modified GRU model

Introduction: Melanoma Skin Cancer (MSC) is a type of cancer in the human body; therefore, early disease diagnosis is essential for reducing the mortality rate. However, dermoscopic image analysis poses challenges due to factors such as color illumination, light reflections, and the varying sizes and shapes of lesions. To overcome these challenges, an automated framework is proposed in this manuscript. Methods: Initially, dermoscopic images are acquired from two online benchmark datasets: International Skin Imaging Collaboration (ISIC) 2020 and Human against Machine (HAM) 10000. Subsequently, a normalization technique is employed on the dermoscopic images to decrease noise impact, outliers, and variations in the pixels. Furthermore, cancerous regions in the pre-processed images are segmented utilizing the mask-faster Region based Convolutional Neural Network (RCNN) model. The mask-RCNN model offers precise pixellevel segmentation by accurately delineating object boundaries. From the partitioned cancerous regions, discriminative feature vectors are extracted by applying three pre-trained CNN models, namely ResNeXt101, Xception, and InceptionV3. These feature vectors are passed into the modified Gated Recurrent Unit (GRU) model for MSC classification. In the modified GRU model, a swish-Rectified Linear Unit (ReLU) activation function is incorporated that efficiently stabilizes the learning process with better convergence rate during training. Results and discussion: The empirical investigation demonstrate that the modified GRU model attained an accuracy of 99.95% and 99.98% on the ISIC 2020 and HAM 10000 datasets, where the obtained results surpass the conventional detection models.

Neural Networks With Gated Recurrent Units Reduce Glucose Forecasting Error Due to Changes in Sensor Location

BACKGROUND

Continuous glucose monitors (CGMs) have become important tools for providing estimates of glucose to patients with diabetes. Recently, neural networks (NNs) have become a common method for forecasting glucose values using data from CGMs. One method of forecasting glucose values is a time-delay feedforward (FF) NN, but a change in the CGM location on a participant can increase forecast error in a FF NN.

METHODS

In response, we examined a NN with gated recurrent units (GRUs) as a method of reducing forecast error due to changes in sensor location.

RESULTS

We observed that for 13 participants with type 2 diabetes wearing blinded CGMs on both arms for 12 weeks (FreeStyle Libre Pro-Abbott), GRU NNs did not produce significantly different errors in glucose prediction due to sensor location changes (P < .05).

CONCLUSION

We observe that GRU NNs can mitigate error in glucose prediction due to differences in CGM location.

Machine learning model for aberrant driving behaviour prediction using heart rate variability: a pilot study involving highway bus drivers

Objectives. Current approaches via physiological features detecting aberrant driving behaviour (ADB), including speeding, abrupt steering, hard braking and aggressive acceleration, are developing. This study proposes using machine learning approaches incorporating heart rate variability (HRV) parameters to predict ADB occurrence. Methods. Naturalistic driving data of 10 highway bus drivers in Taiwan from their daily routes were collected for 4 consecutive days. Their driving behaviours and physiological data during a driving task were determined using a navigation mobile application and heart rate watch. Participants' self-reported data on sleep, driving-related experience, open-source data on weather and the traffic congestion level were obtained. Five machine learning models - logistic regression, random forest, naive Bayes, support vector machine and gated recurrent unit (GRU) - were employed to predict ADBs. Results. Most drivers with ADB had low sleep efficiency (≤80%), with significantly higher scores in driver behaviour questionnaire subcategories of lapses and errors and in the Karolinska sleepiness scale than those without ADBs. Moreover, HRV parameters were significantly different between baseline and pre-ADB event measurements. GRU had the highest accuracy (81.16-84.22%). Conclusions. Sleep deficit may be related to the increased fatigue level and ADB occurrence predicted from HRV-based models among bus drivers.

H2Opred: a robust and efficient hybrid deep learning model for predicting 2'-O-methylation sites in human RNA

2'-O-methylation (2OM) is the most common post-transcriptional modification of RNA. It plays a crucial role in RNA splicing, RNA stability and innate immunity. Despite advances in high-throughput detection, the chemical stability of 2OM makes it difficult to detect and map in messenger RNA. Therefore, bioinformatics tools have been developed using machine learning (ML) algorithms to identify 2OM sites. These tools have made significant progress, but their performances remain unsatisfactory and need further improvement. In this study, we introduced H2Opred, a novel hybrid deep learning (HDL) model for accurately identifying 2OM sites in human RNA. Notably, this is the first application of HDL in developing four nucleotide-specific models [adenine (A2OM), cytosine (C2OM), guanine (G2OM) and uracil (U2OM)] as well as a generic model (N2OM). H2Opred incorporated both stacked 1D convolutional neural network (1D-CNN) blocks and stacked attention-based bidirectional gated recurrent unit (Bi-GRU-Att) blocks. 1D-CNN blocks learned effective feature representations from 14 conventional descriptors, while Bi-GRU-Att blocks learned feature representations from five natural language processing-based embeddings extracted from RNA sequences. H2Opred integrated these feature representations to make the final prediction. Rigorous cross-validation analysis demonstrated that H2Opred consistently outperforms conventional ML-based single-feature models on five different datasets. Moreover, the generic model of H2Opred demonstrated a remarkable performance on both training and testing datasets, significantly outperforming the existing predictor and other four nucleotide-specific H2Opred models. To enhance accessibility and usability, we have deployed a user-friendly web server for H2Opred, accessible at https://balalab-skku.org/H2Opred/. This platform will serve as an invaluable tool for accurately predicting 2OM sites within human RNA, thereby facilitating broader applications in relevant research endeavors.

Human Activity Recognition Using Cascaded Dual Attention CNN and Bi-Directional GRU Framework

Vision-based human activity recognition (HAR) has emerged as one of the essential research areas in video analytics. Over the last decade, numerous advanced deep learning algorithms have been introduced to recognize complex human actions from video streams. These deep learning algorithms have shown impressive performance for the video analytics task. However, these newly introduced methods either exclusively focus on model performance or the effectiveness of these models in terms of computational efficiency, resulting in a biased trade-off between robustness and computational efficiency in their proposed methods to deal with challenging HAR problem. To enhance both the accuracy and computational efficiency, this paper presents a computationally efficient yet generic spatial-temporal cascaded framework that exploits the deep discriminative spatial and temporal features for HAR. For efficient representation of human actions, we propose an efficient dual attentional convolutional neural network (DA-CNN) architecture that leverages a unified channel-spatial attention mechanism to extract human-centric salient features in video frames. The dual channel-spatial attention layers together with the convolutional layers learn to be more selective in the spatial receptive fields having objects within the feature maps. The extracted discriminative salient features are then forwarded to a stacked bi-directional gated recurrent unit (Bi-GRU) for long-term temporal modeling and recognition of human actions using both forward and backward pass gradient learning. Extensive experiments are conducted on three publicly available human action datasets, where the obtained results verify the effectiveness of our proposed framework (DA-CNN+Bi-GRU) over the state-of-the-art methods in terms of model accuracy and inference runtime across each dataset. Experimental results show that the DA-CNN+Bi-GRU framework attains an improvement in execution time up to 167× in terms of frames per second as compared to most of the contemporary action-recognition methods.

Fog-Based Smart Cardiovascular Disease Prediction System Powered by Modified Gated Recurrent Unit

The ongoing fast-paced technology trend has brought forth ceaseless transformation. In this regard, cloud computing has long proven to be the paramount deliverer of services such as computing power, software, networking, storage, and databases on a pay-per-use basis. The cloud is a big proponent of the internet of things (IoT), furnishing the computation and storage requisite to address internet-of-things applications. With the proliferating IoT devices triggering a continual data upsurge, the cloud-IoT interaction encounters latency, bandwidth, and connectivity restraints. The inclusion of the decentralized and distributed fog computing layer amidst the cloud and IoT layer extends the cloud's processing, storage, and networking services close to end users. This hierarchical edge-fog-cloud model distributes computation and intelligence, yielding optimal solutions while tackling constraints like massive data volume, latency, delay, and security vulnerability. The healthcare domain, warranting time-critical functionalities, can reap benefits from the cloud-fog-IoT interplay. This research paper propounded a fog-assisted smart healthcare system to diagnose heart or cardiovascular disease. It combined a fuzzy inference system (FIS) with the recurrent neural network model's variant of the gated recurrent unit (GRU) for pre-processing and predictive analytics tasks. The proposed system showcases substantially improved performance results, with classification accuracy at 99.125%. With major processing of healthcare data analytics happening at the fog layer, it is observed that the proposed work reveals optimized results concerning delays in terms of latency, response time, and jitter, compared to the cloud. Deep learning models are adept at handling sophisticated tasks, particularly predictive analytics. Time-critical healthcare applications reap benefits from deep learning's exclusive potential to furnish near-perfect results, coupled with the merits of the decentralized fog model, as revealed by the experimental results.

Unsupervised Anomaly Detection for Cars CAN Sensors Time Series Using Small Recurrent and Convolutional Neural Networks

Predictive maintenance in the car industry is an active field of research for machine learning and anomaly detection. The capability of cars to produce time series data from sensors is growing as the car industry is heading towards more connected and electric vehicles. Unsupervised anomaly detectors are therefore very adapted to process those complex multidimensional time series and highlight abnormal behaviors. We propose to use recurrent and convolutional neural networks based on unsupervised anomaly detectors with simple architectures on real, multidimensional time series generated by the car sensors and extracted from the Controller Area Network bus (CAN). Our method is then evaluated through known specific anomalies. As the computational costs of Machine Learning algorithms are a rising issue regarding embedded scenarios such as car anomaly detection, we also focus on creating anomaly detectors that are as small as possible. Using a state-of-the-art methodology incorporating a time series predictor and a prediction-error-based anomaly detector, we show that we can obtain roughly the same anomaly detection performance with smaller predictors, reducing parameters and calculations by up to 23% and 60%, respectively. Finally, we introduce a method to correlate variables with specific anomalies by using anomaly detector results and labels.

Machine Learning for Optical Motion Capture-Driven Musculoskeletal Modelling from Inertial Motion Capture Data

Marker-based Optical Motion Capture (OMC) systems and associated musculoskeletal (MSK) modelling predictions offer non-invasively obtainable insights into muscle and joint loading at an in vivo level, aiding clinical decision-making. However, an OMC system is lab-based, expensive, and requires a line of sight. Inertial Motion Capture (IMC) techniques are widely-used alternatives, which are portable, user-friendly, and relatively low-cost, although with lesser accuracy. Irrespective of the choice of motion capture technique, one typically uses an MSK model to obtain the kinematic and kinetic outputs, which is a computationally expensive tool increasingly well approximated by machine learning (ML) methods. Here, an ML approach is presented that maps experimentally recorded IMC input data to the human upper-extremity MSK model outputs computed from ('gold standard') OMC input data. Essentially, this proof-of-concept study aims to predict higher-quality MSK outputs from the much easier-to-obtain IMC data. We use OMC and IMC data simultaneously collected for the same subjects to train different ML architectures that predict OMC-driven MSK outputs from IMC measurements. In particular, we employed various neural network (NN) architectures, such as Feed-Forward Neural Networks (FFNNs) and Recurrent Neural Networks (RNNs) (vanilla, Long Short-Term Memory, and Gated Recurrent Unit) and a comprehensive search for the best-fit model in the hyperparameters space in both subject-exposed (SE) as well as subject-naive (SN) settings. We observed a comparable performance for both FFNN and RNN models, which have a high degree of agreement (ravg,SE,FFNN=0.90±0.19, ravg,SE,RNN=0.89±0.17, ravg,SN,FFNN=0.84±0.23, and ravg,SN,RNN=0.78±0.23) with the desired OMC-driven MSK estimates for held-out test data. The findings demonstrate that mapping IMC inputs to OMC-driven MSK outputs using ML models could be instrumental in transitioning MSK modelling from 'lab to field'.

Multi-Head Spatiotemporal Attention Graph Convolutional Network for Traffic Prediction

Intelligent transportation systems (ITSs) have become an indispensable component of modern global technological development, as they play a massive role in the accurate statistical estimation of vehicles or individuals commuting to a particular transportation facility at a given time. This provides the perfect backdrop for designing and engineering an adequate infrastructural capacity for transportation analyses. However, traffic prediction remains a daunting task due to the non-Euclidean and complex distribution of road networks and the topological constraints of urbanized road networks. To solve this challenge, this paper presents a traffic forecasting model which combines a graph convolutional network, a gated recurrent unit, and a multi-head attention mechanism to simultaneously capture and incorporate the spatio-temporal dependence and dynamic variation in the topological sequence of traffic data effectively. By achieving 91.8% accuracy on the Los Angeles highway traffic (Los-loop) test data for 15-min traffic prediction and an R2 score of 85% on the Shenzhen City (SZ-taxi) test dataset for 15- and 30-min predictions, the proposed model demonstrated that it can learn the global spatial variation and the dynamic temporal sequence of traffic data over time. This has resulted in state-of-the-art traffic forecasting for the SZ-taxi and Los-loop datasets.

Predicting the Output of Solar Photovoltaic Panels in the Absence of Weather Data Using Only the Power Output of the Neighbouring Sites

There is an increasing need for capable models in the forecast of the output of solar photovoltaic panels. These models are vital for optimizing the performance and maintenance of PV systems. There is also a shortage of studies on forecasts of the output power of solar photovoltaics sites in the absence of meteorological data. Unlike common methods, this study explores numerous machine learning algorithms for forecasting the output of solar photovoltaic panels in the absence of weather data such as temperature, humidity and wind speed, which are often used when forecasting the output of solar PV panels. The considered models include Long Short-Term Memory (LSTM), Gated Recurrent Unit (GRU), Recurrent Neural Network (RNN) and Transformer. These models were used with the data collected from 50 different solar photo voltaic sites in South Korea, which consist of readings of the output of each of the sites collected at regular intervals. This study focuses on obtaining multistep forecasts for the multi-in multi-out, multi-in uni-out and uni-in uni-out settings. Detailed experimentation was carried out in each of these settings. Finally, for each of these settings and different lookback and forecast lengths, the best models were also identified.

Putting the Personalized Metabolic Avatar into Production: A Comparison between Deep-Learning and Statistical Models for Weight Prediction

Nutrition is a cross-cutting sector in medicine, with a huge impact on health, from cardiovascular disease to cancer. Employment of digital medicine in nutrition relies on digital twins: digital replicas of human physiology representing an emergent solution for prevention and treatment of many diseases. In this context, we have already developed a data-driven model of metabolism, called a "Personalized Metabolic Avatar" (PMA), using gated recurrent unit (GRU) neural networks for weight forecasting. However, putting a digital twin into production to make it available for users is a difficult task that as important as model building. Among the principal issues, changes to data sources, models and hyperparameters introduce room for error and overfitting and can lead to abrupt variations in computational time. In this study, we selected the best strategy for deployment in terms of predictive performance and computational time. Several models, such as the Transformer model, recursive neural networks (GRUs and long short-term memory networks) and the statistical SARIMAX model were tested on ten users. PMAs based on GRUs and LSTM showed optimal and stable predictive performances, with the lowest root mean squared errors (0.38 ± 0.16-0.39 ± 0.18) and acceptable computational times of the retraining phase (12.7 ± 1.42 s-13.5 ± 3.60 s) for a production environment. While the Transformer model did not bring a substantial improvement over RNNs in term of predictive performance, it increased the computational time for both forecasting and retraining by 40%. The SARIMAX model showed the worst performance in term of predictive performance, though it had the best computational time. For all the models considered, the extent of the data source was a negligible factor, and a threshold was established for the number of time points needed for a successful prediction.

GRU-SVM Based Threat Detection in Cognitive Radio Network

Cognitive radio networks are vulnerable to numerous threats during spectrum sensing. Different approaches can be used to lessen these attacks as the malicious users degrade the performance of the network. The cutting-edge technologies of machine learning and deep learning step into cognitive radio networks (CRN) to detect network problems. Several studies have been conducted utilising various deep learning and machine learning methods. However, only a small number of analyses have used gated recurrent units (GRU), and that too in software defined networks, but these are seldom used in CRN. In this paper, we used GRU in CRN to train and test the dataset of spectrum sensing results. One of the deep learning models with less complexity and more effectiveness for small datasets is GRU, the lightest variant of the LSTM. The support vector machine (SVM) classifier is employed in this study's output layer to distinguish between authorised users and malicious users in cognitive radio network. The novelty of this paper is the application of combined models of GRU and SVM in cognitive radio networks. A high testing accuracy of 82.45%, training accuracy of 80.99% and detection probability of 1 is achieved at 65 epochs in this proposed work.

A novel hybrid CNN and BiGRU-Attention based deep learning model for protein function prediction

Proteins are the building blocks of all living things. Protein function must be ascertained if the molecular mechanism of life is to be understood. While CNN is good at capturing short-term relationships, GRU and LSTM can capture long-term dependencies. A hybrid approach that combines the complementary benefits of these deep-learning models motivates our work. Protein Language models, which use attention networks to gather meaningful data and build representations for proteins, have seen tremendous success in recent years processing the protein sequences. In this paper, we propose a hybrid CNN + BiGRU - Attention based model with protein language model embedding that effectively combines the output of CNN with the output of BiGRU-Attention for predicting protein functions. We evaluated the performance of our proposed hybrid model on human and yeast datasets. The proposed hybrid model improves the Fmax value over the state-of-the-art model SDN2GO for the cellular component prediction task by 1.9 %, for the molecular function prediction task by 3.8 % and for the biological process prediction task by 0.6 % for human dataset and for yeast dataset the cellular component prediction task by 2.4 %, for the molecular function prediction task by 5.2 % and for the biological process prediction task by 1.2 %.

CEGH: A Hybrid Model Using CEEMD, Entropy, GRU, and History Attention for Intraday Stock Market Forecasting

Intraday stock time series are noisier and more complex than other financial time series with longer time horizons, which makes it challenging to predict. We propose a hybrid CEGH model for intraday stock market forecasting. The CEGH model contains four stages. First, we use complete ensemble empirical mode decomposition (CEEMD) to decompose the original intraday stock market data into different intrinsic mode functions (IMFs). Then, we calculate the approximate entropy (ApEn) values and sample entropy (SampEn) values of each IMF to eliminate noise. After that, we group the retained IMFs into four groups and predict the comprehensive signals of those groups using a feedforward neural network (FNN) or gate recurrent unit with history attention (GRU-HA). Finally, we obtain the final prediction results by integrating the prediction results of each group. The experiments were conducted on the U.S. and China stock markets to evaluate the proposed model. The results demonstrate that the CEGH model improved forecasting performance considerably. The creation of a collaboration between CEEMD, entropy-based denoising, and GRU-HA is our major contribution. This hybrid model could improve the signal-to-noise ratio of stock data and extract global dependence more comprehensively in intraday stock market forecasting.

Track Geometry Prediction Using Three-Dimensional Recurrent Neural Network-Based Models Cross-Functionally Co-Simulated with BIM

Railway track maintenance plays an important role in enabling safe, reliable, and seamless train operations and passenger comfort. Due to the increasing rail transportation, rolling stocks tend to run faster and the load tends to increase continuously. As a result, the track deteriorates quicker, and maintenance needs to be performed more frequently. However, more frequent maintenance activities do not guarantee a better overall performance of the railway system. It is crucial for rail infrastructure managers to optimize predictive and preventative maintenance. This study is the world's first to develop deep machine learning models using three-dimensional recurrent neural network-based co-simulation models to predict track geometry parameters in the next year. Different recurrent neural network-based techniques are used to develop predictive models. In addition, a building information modeling (BIM) model is developed to integrate and cross-functionally co-simulate the track geometry measurement with the prediction for predictive and preventative maintenance purposes. From the study, the developed BIM models can be used to exchange information for predictive maintenance. Machine learning models provide the average R² of 0.95 and the average mean absolute error of 0.56 mm. The insightful breakthrough demonstrates the potential of machine learning and BIM for predictive maintenance, which can promote the safety and cost effectiveness of railway maintenance.

Wireless Sensing Technology Combined with Facial Expression to Realize Multimodal Emotion Recognition

Emotions significantly impact human physical and mental health, and, therefore, emotion recognition has been a popular research area in neuroscience, psychology, and medicine. In this paper, we preprocess the raw signals acquired by millimeter-wave radar to obtain high-quality heartbeat and respiration signals. Then, we propose a deep learning model incorporating a convolutional neural network and gated recurrent unit neural network in combination with human face expression images. The model achieves a recognition accuracy of 84.5% in person-dependent experiments and 74.25% in person-independent experiments. The experiments show that it outperforms a single deep learning model compared to traditional machine learning algorithms.

A Dual-Stage Attention Model for Tool Wear Prediction in Dry Milling Operation

The intelligent monitoring of tool wear status and wear prediction are important factors affecting the intelligent development of the modern machinery industry. Many scholars have used deep learning methods to achieve certain results in tool wear prediction. However, due to the instability and variability of the signal data, some neural network models may have gradient decay between layers. Most methods mainly focus on feature selection of the input data but ignore the influence degree of different features to tool wear. In order to solve these problems, this paper proposes a dual-stage attention model for tool wear prediction. A CNN-BiGRU-attention network model is designed, which introduces the self-attention to extract deep features and embody more important features. The IndyLSTM is used to construct a stable network to solve the gradient decay problem between layers. Moreover, the attention mechanism is added to the network to obtain the important information of output sequence, which can improve the accuracy of the prediction. Experimental study is carried out for tool wear prediction in a dry milling operation to demonstrate the viability of this method. Through the experimental comparison and analysis with regression prediction evaluation indexes, it proves the proposed method can effectively characterize the degree of tool wear, reduce the prediction errors, and achieve good prediction results.

Multi-Layer Graph Attention Network for Sleep Stage Classification Based on EEG

Graph neural networks have been successfully applied to sleep stage classification, but there are still challenges: (1) How to effectively utilize epoch information of EEG-adjacent channels owing to their different interaction effects. (2) How to extract the most representative features according to confused transitional information in confused stages. (3) How to improve classification accuracy of sleep stages compared with existing models. To address these shortcomings, we propose a multi-layer graph attention network (MGANet). Node-level attention prompts the graph attention convolution and GRU to focus on and differentiate the interaction between channels in the time-frequency domain and the spatial domain, respectively. The multi-head spatial-temporal mechanism balances the channel weights and dynamically adjusts channel features, and a multi-layer graph attention network accurately expresses the spatial sleep information. Moreover, stage-level attention is applied to easily confused sleep stages, which effectively improves the limitations of a graph convolutional network in large-scale graph sleep stages. The experimental results demonstrated classification accuracy; MF1 and Kappa reached 0.825, 0.814, and 0.775 and 0.873, 0.801, and 0.827 for the ISRUC and SHHS datasets, respectively, which showed that MGANet outperformed the state-of-the-art baselines.

Gated Recurrent Unit Network for Psychological Stress Classification Using Electrocardiograms from Wearable Devices

In recent years, research on human psychological stress using wearable devices has gradually attracted attention. However, the physical and psychological differences among individuals and the high cost of data collection are the main challenges for further research on this problem. In this work, our aim is to build a model to detect subjects' psychological stress in different states through electrocardiogram (ECG) signals. Therefore, we design a VR high-altitude experiment to induce psychological stress for the subject to obtain the ECG signal dataset. In the experiment, participants wear smart ECG T-shirts with embedded sensors to complete different tasks so as to record their ECG signals synchronously. Considering the temporal continuity of individual psychological stress, a deep, gated recurrent unit (GRU) neural network is developed to capture the mapping relationship between subjects' ECG signals and stress in different states through heart rate variability features at different moments, so as to build a neural network model from the ECG signal to psychological stress detection. The experimental results show that compared with all comparison methods, our method has the best classification performance on the four stress states of resting, VR scene adaptation, VR task and recovery, and it can be a remote stress monitoring solution for some special industries.

Applying Deep Learning on a Few EEG Electrodes during Resting State Reveals Depressive States. A Data Driven Study

The growing number of depressive people and the overload in primary care services make it necessary to identify depressive states with easily accessible biomarkers such as mobile electroencephalography (EEG). Some studies have addressed this issue by collecting and analyzing EEG resting state in a search of appropriate features and classification methods. Traditionally, EEG resting state classification methods for depression were mainly based on linear or a combination of linear and non-linear features. We hypothesize that participants with ongoing depressive states differ from controls in complex patterns of brain dynamics that can be captured in EEG resting state data, using only nonlinear measures on a few electrodes, making it possible to develop cheap and wearable devices that could be even monitored through smartphones. To validate such a perspective, a resting-state EEG study was conducted with 50 participants, half with depressive state (DEP) and half controls (CTL). A data-driven approach was applied to select the most appropriate time window and electrodes for the EEG analyses, as suggested by Giacometti⁠, as well as the most efficient nonlinear features and classifiers, to distinguish between CTL and DEP participants. Nonlinear features showing temporo-spatial and spectral complexity were selected. The results confirmed that computing nonlinear features from a few selected electrodes in a 15 s time window are sufficient to classify DEP and CTL participants accurately. Finally, after training and testing internally the classifier, the trained machine was applied to EEG resting state data (CTL and DEP) from a publicly available database, validating the capacity of generalization of the classifier with data from different equipment, population, and environment obtaining an accuracy near 100%.

Intrusion Detection in IoT Using Deep Learning

Cybersecurity has been widely used in various applications, such as intelligent industrial systems, homes, personal devices, and cars, and has led to innovative developments that continue to face challenges in solving problems related to security methods for IoT devices. Effective security methods, such as deep learning for intrusion detection, have been introduced. Recent research has focused on improving deep learning algorithms for improved security in IoT. This research explores intrusion detection methods implemented using deep learning, compares the performance of different deep learning methods, and identifies the best method for implementing intrusion detection in IoT. This research is conducted using deep learning models based on convolutional neural networks (CNNs), long short-term memory (LSTM), and gated recurrent units (GRUs). A standard dataset for intrusion detection in IoT is considered to evaluate the proposed model. Finally, the empirical results are analyzed and compared with the existing approaches for intrusion detection in IoT. The proposed method seemed to have the highest accuracy compared to the existing methods.

Visualization deep learning model for automatic arrhythmias classification

With the improvement of living standards, heart disease has become one of the common diseases that threaten human health. As a widely used reliable, non-invasive technology, the Electrocardiography (ECG) data has been increasingly used for diagnosing cardiovascular diseases. With the rapid growth of ECG examinations and the shortage of cardiologists, accurate and automatic recognition of ECG patterns has become a research hotspot. In order to improve the accuracy in detecting abnormal ECG patterns, this paper proposes a hybrid 1D Resnet-GRU consisting of the Resnet and gated recurrent unit (GRU) modules to implement classification of arrhythmias from 12-lead ECG recordings. In addition, the focal Loss function is used to solve the problem of unbalanced datasets. Based on the proposed 1D Resnet-GRU model, the Grad-CAM++ mechanism, one of class-discriminative visualization methods, has been employed to the trained network model and generate thermal images superimposed on raw signals, so as to improve interpretability and transparency for arrhythmia classification. Experimental results show the proposed Resnet-GRU method can improve the F_1 score and accuracy effectively, and Grad-CAM++ can provide interpretability and clinical information for arrhythmia classification.

Using Complementary Ensemble Empirical Mode Decomposition and Gated Recurrent Unit to Predict Landslide Displacements in Dam Reservoir

It is crucial to predict landslide displacement accurately for establishing a reliable early warning system. Such a requirement is more urgent for landslides in the reservoir area. The main reason is that an inaccurate prediction can lead to riverine disasters and secondary surge disasters. Machine learning (ML) methods have been developed and commonly applied in landslide displacement prediction because of their powerful nonlinear processing ability. Recently, deep ML methods have become popular, as they can deal with more complicated problems than conventional ML methods. However, it is usually not easy to obtain a well-trained deep ML model, as many hyperparameters need to be trained. In this paper, a deep ML method—the gated recurrent unit (GRU)—with the advantages of a powerful prediction ability and fewer hyperparameters, was applied to forecast landslide displacement in the dam reservoir. The accumulated displacement was firstly decomposed into a trend term, a periodic term, and a stochastic term by complementary ensemble empirical mode decomposition (CEEMD). A univariate GRU model and a multivariable GRU model were employed to forecast trend and stochastic displacements, respectively. A multivariable GRU model was applied to predict periodic displacement, and another two popular ML methods—long short-term memory neural networks (LSTM) and random forest (RF)—were used for comparison. Precipitation, reservoir level, and previous displacement were considered to be candidate-triggering factors for inputs of the models. The Baijiabao landslide, located in the Three Gorges Reservoir Area (TGRA), was taken as a case study to test the prediction ability of the model. The results demonstrated that the GRU algorithm provided the most encouraging results. Such a satisfactory prediction accuracy of the GRU algorithm depends on its ability to fully use the historical information while having fewer hyperparameters to train. It is concluded that the proposed model can be a valuable tool for predicting the displacements of landslides in the TGRA and other dam reservoirs.

HTR for Greek Historical Handwritten Documents

Offline handwritten text recognition (HTR) for historical documents aims for effective transcription by addressing challenges that originate from the low quality of manuscripts under study as well as from several particularities which are related to the historical period of writing. In this paper, the challenge in HTR is related to a focused goal of the transcription of Greek historical manuscripts that contain several particularities. To this end, in this paper, a convolutional recurrent neural network architecture is proposed that comprises octave convolution and recurrent units which use effective gated mechanisms. The proposed architecture has been evaluated on three newly created collections from Greek historical handwritten documents that will be made publicly available for research purposes as well as on standard datasets like IAM and RIMES. For evaluation we perform a concise study which shows that compared to state of the art architectures, the proposed one deals effectively with the challenging Greek historical manuscripts.

De Novo Molecular Design of Caspase-6 Inhibitors by a GRU-Based Recurrent Neural Network Combined with a Transfer Learning Approach

Due to their potential in the treatment of neurodegenerative diseases, caspase-6 inhibitors have attracted widespread attention. However, the existing caspase-6 inhibitors showed more or less inevitable deficiencies that restrict their clinical development and applications. Therefore, there is an urgent need to develop novel caspase-6 candidate inhibitors. Herein, a gated recurrent unit (GRU)-based recurrent neural network (RNN) combined with transfer learning was used to build a molecular generative model of caspase-6 inhibitors. The results showed that the GRU-based RNN model can accurately learn the SMILES grammars of about 2.4 million chemical molecules including ionic and isomeric compounds and can generate potential caspase-6 inhibitors after transfer learning of the known 433 caspase-6 inhibitors. Based on the novel molecules derived from the molecular generative model, an optimal logistic regression model and Surflex-dock were employed for predicting and ranking the inhibitory activities. According to the prediction results, three potential caspase-6 inhibitors with different scaffolds were selected as the promising candidates for further research. In general, this paper provides an efficient combinational strategy for de novo molecular design of caspase-6 inhibitors.

Improved Prediction Model of Protein Lysine Crotonylation Sites Using Bidirectional Recurrent Neural Networks

Histone lysine crotonylation (Kcr) is a post-translational modification of histone proteins that is involved in the regulation of gene transcription, acute and chronic kidney injury, spermatogenesis, depression, cancer, and so forth. The identification of Kcr sites in proteins is important for characterizing and regulating primary biological mechanisms. The use of computational approaches such as machine learning and deep learning algorithms have emerged in recent years as the traditional wet-lab experiments are time-consuming and costly. We propose as part of this study a deep learning model based on a recurrent neural network (RNN) termed as Sohoko-Kcr for the prediction of Kcr sites. Through the embedded encoding of the peptide sequences, we investigate the efficiency of RNN-based models such as long short-term memory (LSTM), bidirectional LSTM (BiLSTM), and bidirectional gated recurrent unit (BiGRU) networks using cross-validation and independent tests. We also established the comparison between Sohoko-Kcr and other published tools to verify the efficiency of our model based on 3-fold, 5-fold, and 10-fold cross-validations using independent set tests. The results then show that the BiGRU model has consistently displayed outstanding performance and computational efficiency. Based on the proposed model, a webserver called Sohoko-Kcr was deployed for free use and is accessible at https://sohoko-research-9uu23.ondigitalocean.app.

Wearable IMU-Based Human Activity Recognition Algorithm for Clinical Balance Assessment Using 1D-CNN and GRU Ensemble Model

In this study, a wearable inertial measurement unit system was introduced to assess patients via the Berg balance scale (BBS), a clinical test for balance assessment. For this purpose, an automatic scoring algorithm was developed. The principal aim of this study is to improve the performance of the machine-learning-based method by introducing a deep-learning algorithm. A one-dimensional (1D) convolutional neural network (CNN) and a gated recurrent unit (GRU) that shows good performance in multivariate time-series data were used as model components to find the optimal ensemble model. Various structures were tested, and a stacking ensemble model with a simple meta-learner after two 1D-CNN heads and one GRU head showed the best performance. Additionally, model performance was enhanced by improving the dataset via preprocessing. The data were down sampled, an appropriate sampling rate was found, and the training and evaluation times of the model were improved. Using an augmentation process, the data imbalance problem was solved, and model accuracy was improved. The maximum accuracy of 14 BBS tasks using the model was 98.4%, which is superior to the results of previous studies.

Temperature Prediction Using Multivariate Time Series Deep Learning in the Lining of an Electric Arc Furnace for Ferronickel Production

The analysis of data from sensors in structures subjected to extreme conditions such as the ones used in smelting processes is a great decision tool that allows knowing the behavior of the structure under different operational conditions. In this industry, the furnaces and the different elements are fully instrumented, including sensors to measure variables such as temperature, pressure, level, flow, power, electrode positions, among others. From the point of view of engineering and data analytics, this quantity of data presents an opportunity to understand the operation of the system under normal conditions or to explore new ways of operation by using information from models provided by using deep learning approaches. Although some approaches have been developed with application to this industry, it is still an open research area. As a contribution, this paper presents an applied deep learning temperature prediction model for a 75 MW electric arc furnace, which is used for ferronickel production. In general, the methodology proposed considers two steps: first, a data cleaning process to increase the quality of the data, eliminating both redundant information as well as atypical and unusual data, and second, a multivariate time series deep learning model to predict the temperatures in the furnace lining. The developed deep learning model is a sequential one based on GRU (gated recurrent unit) layer plus a dense layer. The GRU + Dense model achieved an average root mean square error (RMSE) of 1.19 °C in the test set of 16 different thermocouples radially distributed on the furnace.

Extended-Range Prediction Model Using NSGA-III Optimized RNN-GRU-LSTM for Driver Stress and Drowsiness

Road traffic accidents have been listed in the top 10 global causes of death for many decades. Traditional measures such as education and legislation have contributed to limited improvements in terms of reducing accidents due to people driving in undesirable statuses, such as when suffering from stress or drowsiness. Attention is drawn to predicting drivers’ future status so that precautions can be taken in advance as effective preventative measures. Common prediction algorithms include recurrent neural networks (RNNs), gated recurrent units (GRUs), and long short-term memory (LSTM) networks. To benefit from the advantages of each algorithm, nondominated sorting genetic algorithm-III (NSGA-III) can be applied to merge the three algorithms. This is named NSGA-III-optimized RNN-GRU-LSTM. An analysis can be made to compare the proposed prediction algorithm with the individual RNN, GRU, and LSTM algorithms. Our proposed model improves the overall accuracy by 11.2–13.6% and 10.2–12.2% in driver stress prediction and driver drowsiness prediction, respectively. Likewise, it improves the overall accuracy by 6.9–12.7% and 6.9–8.9%, respectively, compared with boosting learning with multiple RNNs, multiple GRUs, and multiple LSTMs algorithms. Compared with existing works, this proposal offers to enhance performance by taking some key factors into account—namely, using a real-world driving dataset, a greater sample size, hybrid algorithms, and cross-validation. Future research directions have been suggested for further exploration and performance enhancement.

Multi-class motor imagery EEG classification method with high accuracy and low individual differences based on hybrid neural network

Objective.Most current methods of classifying different patterns for motor imagery EEG signals require complex pre-processing and feature extraction steps, which consume time and lack adaptability, ignoring individual differences in EEG signals. It is essential to improve algorithm performance with the increased classes and diversity of subjects.Approach.This study introduces deep learning method for end-to-end learning to complete the classification of four-class MI tasks, aiming to improve the recognition rate and balance the classification accuracy among different subjects. A new one-dimensional input data representation method is proposed. This representation method can increase the number of samples and ignore the influence of channel correlation. In addition, a cascade network of convolutional neural network and gated recurrent unit is designed to learn time-frequency information from EEG data without extracting features manually, this model can capture the hidden representations related to different MI mode of each people.Main results. Experiments on BCI Competition 2a dataset and actual collected dataset achieve high accuracy near 99.40% and 92.56%, and the standard deviation is 0.34 and 1.35 respectively. Results demonstrate that the proposed method outperforms the advanced methods and baseline models.Significance.Experimental results show that the proposed method improves the accuracy of multi-classification and overcomes the impact of individual differences on classification by training neural network subject-dependent, which promotes the development of actual brain-computer interface systems.

An Improved Feature Selection Method Based on Random Forest Algorithm for Wind Turbine Condition Monitoring

Feature selection and dimensionality reduction are important for the performance of wind turbine condition monitoring models using supervisory control and data acquisition (SCADA) data. In this paper, an improved random forest algorithm, namely Feature Simplification Random Forest (FS_RF), is proposed, which is capable of identifying features closely correlated with wind turbine working conditions. The Euclidian distances are employed to distinguish the weight of the same feature among different samples, and its importance is measured by means of the random forest algorithm. The selected features are finally verified by a two-layer gated recurrent unit (GRU) neural network facilitating condition monitoring. The experimental results demonstrate the capacity and effectiveness of the proposed method for wind turbine condition monitoring.

CricShotClassify: An Approach to Classifying Batting Shots from Cricket Videos Using a Convolutional Neural Network and Gated Recurrent Unit

Recognizing the sport of cricket on the basis of different batting shots can be a significant part of context-based advertisement to users watching cricket, generating sensor-based commentary systems and coaching assistants. Due to the similarity between different batting shots, manual feature extraction from video frames is tedious. This paper proposes a hybrid deep-neural-network architecture for classifying 10 different cricket batting shots from offline videos. We composed a novel dataset, CricShot10, comprising uneven lengths of batting shots and unpredictable illumination conditions. Impelled by the enormous success of deep-learning models, we utilized a convolutional neural network (CNN) for automatic feature extraction, and a gated recurrent unit (GRU) to deal with long temporal dependency. Initially, conventional CNN and dilated CNN-based architectures were developed. Following that, different transfer-learning models were investigated—namely, VGG16, InceptionV3, Xception, and DenseNet169—which freeze all the layers. Experiment results demonstrated that the VGG16–GRU model outperformed the other models by attaining 86% accuracy. We further explored VGG16 and two models were developed, one by freezing all but the final 4 VGG16 layers, and another by freezing all but the final 8 VGG16 layers. On our CricShot10 dataset, these two models were 93% accurate. These results verify the effectiveness of our proposed architecture compared with other methods in terms of accuracy.

An Attention-Based Multilayer GRU Model for Multistep-Ahead Short-Term Load Forecasting†

Recently, multistep-ahead prediction has attracted much attention in electric load forecasting because it can deal with sudden changes in power consumption caused by various events such as fire and heat wave for a day from the present time. On the other hand, recurrent neural networks (RNNs), including long short-term memory and gated recurrent unit (GRU) networks, can reflect the previous point well to predict the current point. Due to this property, they have been widely used for multistep-ahead prediction. The GRU model is simple and easy to implement; however, its prediction performance is limited because it considers all input variables equally. In this paper, we propose a short-term load forecasting model using an attention based GRU to focus more on the crucial variables and demonstrate that this can achieve significant performance improvements, especially when the input sequence of RNN is long. Through extensive experiments, we show that the proposed model outperforms other recent multistep-ahead prediction models in the building-level power consumption forecasting.

Optimal Artificial Neural Network Type Selection Method for Usage in Smart House Systems

In the process of the “smart” house systems work, there is a need to process fuzzy input data. The models based on the artificial neural networks are used to process fuzzy input data from the sensors. However, each artificial neural network has a certain advantage and, with a different accuracy, allows one to process different types of data and generate control signals. To solve this problem, a method of choosing the optimal type of artificial neural network has been proposed. It is based on solving an optimization problem, where the optimization criterion is an error of a certain type of artificial neural network determined to control the corresponding subsystem of a “smart” house. In the process of learning different types of artificial neural networks, the same historical input data are used. The research presents the dependencies between the types of neural networks, the number of inner layers of the artificial neural network, the number of neurons on each inner layer, the error of the settings parameters calculation of the relative expected results.

A Real-Time Trajectory Prediction Method of Small-Scale Quadrotors Based on GPS Data and Neural Network

This paper proposes a real-time trajectory prediction method for quadrotors based on a bidirectional gated recurrent unit model. Historical trajectory data of ten types of quadrotors were obtained. The bidirectional gated recurrent units were constructed and utilized to learn the historic data. The prediction results were compared with the traditional gated recurrent unit method to test its prediction performance. The efficiency of the proposed algorithm was investigated by comparing the training loss and training time. The results over the testing datasets showed that the proposed model produced better prediction results than the baseline models for all scenarios of the testing datasets. It was also found that the proposed model can converge to a stable state faster than the traditional gated recurrent unit model. Moreover, various types of training samples were applied and compared. With the same randomly selected test datasets, the performance of the prediction model can be improved by selecting the historical trajectory samples of the quadrotors close to the weight or volume of the target quadrotor for training. In addition, the performance of stable trajectory samples is significantly better than that with unstable trajectory segments with a frequent change of speed and direction with large angles.

A Comparative Analysis of Hybrid Deep Learning Models for Human Activity Recognition

Recent advances in artificial intelligence and machine learning (ML) led to effective methods and tools for analyzing the human behavior. Human Activity Recognition (HAR) is one of the fields that has seen an explosive research interest among the ML community due to its wide range of applications. HAR is one of the most helpful technology tools to support the elderly's daily life and to help people suffering from cognitive disorders, Parkinson's disease, dementia, etc. It is also very useful in areas such as transportation, robotics and sports. Deep learning (DL) is a branch of ML based on complex Artificial Neural Networks (ANNs) that has demonstrated a high level of accuracy and performance in HAR. Convolutional Neural Networks (CNNs) and Recurrent Neural Networks (RNNs) are two types of DL models widely used in the recent years to address the HAR problem. The purpose of this paper is to investigate the effectiveness of their integration in recognizing daily activities, e.g., walking. We analyze four hybrid models that integrate CNNs with four powerful RNNs, i.e., LSTMs, BiLSTMs, GRUs and BiGRUs. The outcomes of our experiments on the PAMAP2 dataset indicate that our proposed hybrid models achieve an outstanding level of performance with respect to several indicative measures, e.g., F-score, accuracy, sensitivity, and specificity.

Mixture Optimization of Recycled Aggregate Concrete Using Hybrid Machine Learning Model

Recycled aggregate concrete (RAC) contributes to mitigating the depletion of natural aggregates, alleviating the carbon footprint of concrete construction, and averting the landfilling of colossal amounts of construction and demolition waste. However, complexities in the mixture optimization of RAC due to the variability of recycled aggregates and lack of accuracy in estimating its compressive strength require novel and sophisticated techniques. This paper aims at developing state-of-the-art machine learning models to predict the RAC compressive strength and optimize its mixture design. Results show that the developed models including Gaussian processes, deep learning, and gradient boosting regression achieved robust predictive performance, with the gradient boosting regression trees yielding highest prediction accuracy. Furthermore, a particle swarm optimization coupled with gradient boosting regression trees model was developed to optimize the mixture design of RAC for various compressive strength classes. The hybrid model achieved cost-saving RAC mixture designs with lower environmental footprint for different target compressive strength classes. The model could be further harvested to achieve sustainable concrete with optimal recycled aggregate content, least cost, and least environmental footprint.

Detecting Coal Pulverizing System Anomaly Using a Gated Recurrent Unit and Clustering

The coal pulverizing system is an important auxiliary system in thermal power generation systems. The working condition of a coal pulverizing system may directly affect the safety and economy of power generation. Prognostics and health management is an effective approach to ensure the reliability of coal pulverizing systems. As the coal pulverizing system is a typical dynamic and nonlinear high-dimensional system, it is difficult to construct accurate mathematical models used for anomaly detection. In this paper, a novel data-driven integrated framework for anomaly detection of the coal pulverizing system is proposed. A neural network model based on gated recurrent unit (GRU) networks, a type of recurrent neural network (RNN), is constructed to describe the temporal characteristics of high-dimensional data and predict the system condition value. Then, aiming at the prediction error, a novel unsupervised clustering algorithm for anomaly detection is proposed. The proposed framework is validated by a real case study from an industrial coal pulverizing system. The results show that the proposed framework can detect the anomaly successfully.

A Novel Fault Diagnosis Approach for Chillers Based on 1-D Convolutional Neural Network and Gated Recurrent Unit

The safety of an Internet Data Center (IDC) is directly determined by the reliability and stability of its chiller system. Thus, combined with deep learning technology, an innovative hybrid fault diagnosis approach (1D-CNN_GRU) based on the time-series sequences is proposed in this study for the chiller system using 1-Dimensional Convolutional Neural Network (1D-CNN) and Gated Recurrent Unit (GRU). Firstly, 1D-CNN is applied to automatically extract the local abstract features of the sensor sequence data. Secondly, GRU with long and short term memory characteristics is applied to capture the global features, as well as the dynamic information of the sequence. Moreover, batch normalization and dropout are introduced to accelerate network training and address the overfitting issue. The effectiveness and reliability of the proposed hybrid algorithm are assessed on the RP-1043 dataset; based on the experimental results, 1D-CNN_GRU displays the best performance compared with the other state-of-the-art algorithms. Further, the experimental results reveal that 1D-CNN_GRU has a superior identification rate for minor faults.

Development of a Real-Time Risk Prediction Model for In-Hospital Cardiac Arrest in Critically Ill Patients Using Deep Learning: Retrospective Study

BACKGROUND

Cardiac arrest is the most serious death-related event in intensive care units (ICUs), but it is not easily predicted because of the complex and time-dependent data characteristics of intensive care patients. Given the complexity and time dependence of ICU data, deep learning-based methods are expected to provide a good foundation for developing risk prediction models based on large clinical records.

OBJECTIVE

This study aimed to implement a deep learning model that estimates the distribution of cardiac arrest risk probability over time based on clinical data and assesses its potential.

METHODS

A retrospective study of 759 ICU patients was conducted between January 2013 and July 2015. A character-level gated recurrent unit with a Weibull distribution algorithm was used to develop a real-time prediction model. Fivefold cross-validation testing (training set: 80% and validation set: 20%) determined the consistency of model accuracy. The time-dependent area under the curve (TAUC) was analyzed based on the aggregation of 5 validation sets.

RESULTS

The TAUCs of the implemented model were 0.963, 0.942, 0.917, 0.875, 0.850, 0.842, and 0.761 before cardiac arrest at 1, 8, 16, 24, 32, 40, and 48 hours, respectively. The sensitivity was between 0.846 and 0.909, and specificity was between 0.923 and 0.946. The distribution of risk between the cardiac arrest group and the non-cardiac arrest group was generally different, and the difference rapidly increased as the time left until cardiac arrest reduced.

CONCLUSIONS

A deep learning model for forecasting cardiac arrest was implemented and tested by considering the cumulative and fluctuating effects of time-dependent clinical data gathered from a large medical center. This real-time prediction model is expected to improve patient's care by allowing early intervention in patients at high risk of unexpected cardiac arrests.

Continuous Finger Gesture Recognition Based on Flex Sensors

The goal of this work is to present a novel continuous finger gesture recognition system based on flex sensors. The system is able to carry out accurate recognition of a sequence of gestures. Wireless smart gloves equipped with flex sensors were implemented for the collection of the training and testing sets. Given the sensory data acquired from the smart gloves, the gated recurrent unit (GRU) algorithm was then adopted for gesture spotting. During the training process for the GRU, the movements associated with different fingers and the transitions between two successive gestures were taken into consideration. On the basis of the gesture spotting results, the maximum a posteriori (MAP) estimation was carried out for the final gesture classification. Because of the effectiveness of the proposed spotting scheme, accurate gesture recognition was achieved even for complicated transitions between successive gestures. From the experimental results, it can be observed that the proposed system is an effective alternative for robust recognition of a sequence of finger gestures.

Inferring Drug-Related Diseases Based on Convolutional Neural Network and Gated Recurrent Unit

Predicting novel uses for drugs using their chemical, pharmacological, and indication information contributes to minimizing costs and development periods. Most previous prediction methods focused on integrating the similarity and association information of drugs and diseases. However, they tended to construct shallow prediction models to predict drug-associated diseases, which make deeply integrating the information difficult. Further, path information between drugs and diseases is important auxiliary information for association prediction, while it is not deeply integrated. We present a deep learning-based method, CGARDP, for predicting drug-related candidate disease indications. CGARDP establishes a feature matrix by exploiting a variety of biological premises related to drugs and diseases. A novel model based on convolutional neural network (CNN) and gated recurrent unit (GRU) is constructed to learn the local and path representations for a drug-disease pair. The CNN-based framework on the left of the model learns the local representation of the drug-disease pair from their feature matrix. As the different paths have discriminative contributions to the drug-disease association prediction, we construct an attention mechanism at the path level to learn the informative paths. In the right part, a GRU-based framework learns the path representation based on path information between the drug and the disease. Cross-validation results indicate that CGARDP performs better than several state-of-the-art methods. Further, CGARDP retrieves more real drug-disease associations in the top part of the prediction result that are of concern to biologists. Case studies on five drugs demonstrate that CGARDP can discover potential drug-related disease indications.

MildInt: Deep Learning-Based Multimodal Longitudinal Data Integration Framework

As large amounts of heterogeneous biomedical data become available, numerous methods for integrating such datasets have been developed to extract complementary knowledge from multiple domains of sources. Recently, a deep learning approach has shown promising results in a variety of research areas. However, applying the deep learning approach requires expertise for constructing a deep architecture that can take multimodal longitudinal data. Thus, in this paper, a deep learning-based python package for data integration is developed. The python package deep learning-based multimodal longitudinal data integration framework (MildInt) provides the preconstructed deep learning architecture for a classification task. MildInt contains two learning phases: learning feature representation from each modality of data and training a classifier for the final decision. Adopting deep architecture in the first phase leads to learning more task-relevant feature representation than a linear model. In the second phase, linear regression classifier is used for detecting and investigating biomarkers from multimodal data. Thus, by combining the linear model and the deep learning model, higher accuracy and better interpretability can be achieved. We validated the performance of our package using simulation data and real data. For the real data, as a pilot study, we used clinical and multimodal neuroimaging datasets in Alzheimer's disease to predict the disease progression. MildInt is capable of integrating multiple forms of numerical data including time series and non-time series data for extracting complementary features from the multimodal dataset.

Household Power Demand Prediction Using Evolutionary Ensemble Neural Network Pool with Multiple Network Structures

The progress of technology on energy and IoT fields has led to an increasingly complicated electric environment in low-voltage local microgrid, along with the extensions of electric vehicle, micro-generation, and local storage. It is required to establish a home energy management system (HEMS) to efficiently integrate and manage household energy micro-generation, consumption and storage, in order to realize decentralized local energy systems at the community level. Domestic power demand prediction is of great importance for establishing HEMS on realizing load balancing as well as other smart energy solutions with the support of IoT techniques. Artificial neural networks with various network types (e.g., DNN, LSTM/GRU based RNN) and other configurations are widely utilized on energy predictions. However, the selection of network configuration for each research is generally a case by case study achieved through empirical or enumerative approaches. Moreover, the commonly utilized network initialization methods assign parameter values based on random numbers, which cause diversity on model performance, including learning efficiency, forecast accuracy, etc. In this paper, an evolutionary ensemble neural network pool (EENNP) method is proposed to achieve a population of well-performing networks with proper combinations of configuration and initialization automatically. In the experimental study, power demand predictions of multiple households are explored in three application scenarios: optimizing potential network configuration set, forecasting single household power demand, and refilling missing data. The impacts of evolutionary parameters on model performance are investigated. The experimental results illustrate that the proposed method achieves better solutions on the considered scenarios. The optimized potential network configuration set using EENNP achieves a similar result to manual optimization. The results of household demand prediction and missing data refilling perform better than the naïve and simple predictors.

Individual Identification Using the Functional Brain Fingerprint Detected by the Recurrent Neural Network

Individual identification based on brain function has gained traction in literature. Investigating individual differences in brain function can provide additional insights into the brain. In this work, we introduce a recurrent neural network-based model for identifying individuals based on only a short segment of resting-state functional magnetic resonance imaging data. In addition, we demonstrate how the global signal and differences in atlases affect individual identifiability. Furthermore, we investigate neural network features that exhibit the uniqueness of each individual. The results indicate that our model is able to identify individuals based on neural features and provides additional information regarding brain dynamics.