Search

Accurate classification of electromyographic (EMG) signals is vital in biomedical applications. This study evaluates different architectures of recurrent neural networks for the classification of EMG signals associated with five movements of the right upper extremity. A Butterworth filter was implemented for signal preprocessing, followed by segmentation into 250 ms windows, with an overlap of 190 ms. The resulting dataset was divided into training, validation, and testing subsets. The Grey Wolf Optimization algorithm was applied to the gated recurrent unit (GRU), long short-term memory (LSTM) architectures, and bidirectional recurrent neural networks. In parallel, a performance comparison with support vector machines (SVMs) was performed. The results obtained in the first experimental phase revealed that all the RNN networks evaluated reached a 100% accuracy, standing above the 93% achieved by the SVM. Regarding classification speed, LSTM ranked as the fastest architecture, recording a time of 0.12 ms, followed by GRU with 0.134 ms. Bidirectional recurrent neural networks showed a response time of 0.2 ms, while SVM had the longest time at 2.7 ms. In the second experimental phase, a slight decrease in the accuracy of the RNN models was observed, standing at 98.46% for LSTM, 96.38% for GRU, and 97.63% for the bidirectional network. The findings of this study highlight the effectiveness and speed of recurrent neural networks in the EMG signal classification task. [ABSTRACT FROM AUTHOR]

Indonesia, with the most significant Muslim population worldwide, mandates the consumption of halal food. However, many websites, including Google Maps, do not provide information about halal restaurants. Data integration is essential for obtaining comprehensive and accurate information on halal restaurants from diverse sources, such as the Indonesia Halal Product Assurance Agency (BPJPH) and Google Maps. Preprocessing of these two datasets and their labeling using the Jaccard index were conducted. The Bidirectional Recurrent Neural Networks (BRNN) model was constructed using deepmatcher and evaluated using the F1-score metric. The integration of these two datasets resulted in 155 rows of matching pairs of data. [ABSTRACT FROM AUTHOR]

Collaborative robots, or cobots, are designed to work alongside humans and to alleviate their physical burdens, such as lifting heavy objects or performing tedious tasks. Ensuring the safety of human-robot interaction (HRI) is paramount for effective collaboration. To achieve this, it is essential to have a reliable dynamic model of the cobot that enables the implementation of torque control strategies. These strategies aim to achieve accurate motion while minimizing the amount of torque exerted by the robot. However, modeling the complex non-linear dynamics of cobots with elastic actuators poses a challenge for traditional analytical modeling techniques. Instead, cobot dynamic modeling needs to be learned through data-driven approaches, rather than analytical equation-drivenmodeling. In this study, we propose and evaluate three machine learning (ML) approaches based on bidirectional recurrent neural networks (BRNNs) for learning the inverse dynamic model of a cobot equipped with elastic actuators. We also provide our ML approaches with a representative training dataset of the cobot's joint positions, velocities, and corresponding torque values. The first ML approach uses a non-parametric configuration, while the other two implement semi-parametric configurations. All three ML approaches outperform the rigid-bodied dynamic model provided by the cobot's manufacturer in terms of torque precision while maintaining their generalization capabilities and real-time operation due to the optimized sample dataset size and network dimensions. Despite the similarity in torque estimation of these three configurations, the non-parametric configuration was specifically designed for worst-case scenarios where the robot dynamics are completely unknown. Finally, we validate the applicability of ourML approaches by integrating the worst-case non-parametric configuration as a controller within a feedforward loop. We verify the accuracy of the learned inverse dynamicmodel by comparing it to the actual cobot performance. Our non-parametric architecture outperforms the robot's default factory position controller in terms of accuracy. [ABSTRACT FROM AUTHOR]

Collaborative robots, or cobots, are designed to work alongside humans and to alleviate their physical burdens, such as lifting heavy objects or performing tedious tasks. Ensuring the safety of human–robot interaction (HRI) is paramount for effective collaboration. To achieve this, it is essential to have a reliable dynamic model of the cobot that enables the implementation of torque control strategies. These strategies aim to achieve accurate motion while minimizing the amount of torque exerted by the robot. However, modeling the complex non-linear dynamics of cobots with elastic actuators poses a challenge for traditional analytical modeling techniques. Instead, cobot dynamic modeling needs to be learned through data-driven approaches, rather than analytical equation-driven modeling. In this study, we propose and evaluate three machine learning (ML) approaches based on bidirectional recurrent neural networks (BRNNs) for learning the inverse dynamic model of a cobot equipped with elastic actuators. We also provide our ML approaches with a representative training dataset of the cobot's joint positions, velocities, and corresponding torque values. The first ML approach uses a non-parametric configuration, while the other two implement semi-parametric configurations. All three ML approaches outperform the rigid-bodied dynamic model provided by the cobot's manufacturer in terms of torque precision while maintaining their generalization capabilities and real-time operation due to the optimized sample dataset size and network dimensions. Despite the similarity in torque estimation of these three configurations, the non-parametric configuration was specifically designed for worst-case scenarios where the robot dynamics are completely unknown. Finally, we validate the applicability of our ML approaches by integrating the worst-case non-parametric configuration as a controller within a feedforward loop. We verify the accuracy of the learned inverse dynamic model by comparing it to the actual cobot performance. Our non-parametric architecture outperforms the robot's default factory position controller in terms of accuracy.

Poetry has a prominent history in Arabic literature. The classical Arabic poetry has 16 m that vary in rhythm and target purpose. Chanting a poem eloquently requires knowing the poem’s meter and obtaining a diacritized version of its verses (letters inscribed with their short vowels); diacritics are often not inscribed in Arabic texts. This work proposes solutions to classify input Arabic text into the 16 poetry meters and prose. It also investigates the automatic diacritization of Arabic poetry. We adopt machine learning approach using a large dataset of 1657 k verses of poems and prose to develop neural networks to classify and diacritize Arabic poetry. We propose deep and narrow recurrent neural networks with bidirectional long short-term memory cells for solving these problems. The proposed model classifies the input text with an average accuracy of 97.27%, which is significantly higher than previous work. We also propose a solution that achieves an accuracy that approaches 100% when multiple verses of the same poem are available through predicting the class from the aggregate probabilities of the multiple verses. Diacritizing poetry is much harder than diacritizing prose due to the poet’s meticulous selection of phrases and relaxation of some diacritization rules.

Estimating human pose from images is the key to enabling machines to understand human actions. Existing works on human pose estimation mainly focus on designing more resultful deep neural networks to regress the locations of human joints. Although the human pose is obedient to anatomy and shows rich anatomical features, reasoning human body structure by machine in a complex environment is still an open problem. This paper proposes AnatPose which can effectively capture the structural dependency among human body parts by both deep neural network architecture and learning objectives: (1) For the deep neural network architecture, a bidirectional learning paradigm is proposed to learn body-part proportions and dependencies by organizing human body parts as sequential data. This innovation enables the messages to pass in a bidirectional way and makes the human body exchange information about each part deeper during training. (2) For the learning objective, the proposed AnatPose learns a probabilistic representation of multi-scale anatomical features, including keypoint heatmaps, bone heatmaps, and symmetry heatmaps. This innovation enables the multi-scale anatomical features to successfully capture the structural dependency at both low-level joints and high-level associations from the anatomical priors of the human body. Extensive experimental results demonstrate that the proposed AnatPose shows state-of-the-art performance on three challenging datasets. It achieves a PCK@0.2 detection rate of 95.2% on the LSP dataset, a PCKh@0.5 detection rate of 92.9% on the MPII dataset, and an mAP of 76.6% on the Microsoft COCO dataset. Benefiting from its state-of-the-art accuracy, the proposed approach is expected to be widely used in various human pose estimation-driven applications. • AnatPose organizes human body parts as anatomically sequential data. • AnatPose learns the anatomical properties by a bidirectional recurrent neural network. • An anatomically significant objective is designed for optimizing the neural network. • Extensive results show superior performance on human pose estimation. [ABSTRACT FROM AUTHOR]

Accurate classification of electromyographic (EMG) signals is vital in biomedical applications. This study evaluates different architectures of recurrent neural networks for the classification of EMG signals associated with five movements of the right upper extremity. A Butterworth filter was implemented for signal preprocessing, followed by segmentation into 250 ms windows, with an overlap of 190 ms. The resulting dataset was divided into training, validation, and testing subsets. The Grey Wolf Optimization algorithm was applied to the gated recurrent unit (GRU), long short-term memory (LSTM) architectures, and bidirectional recurrent neural networks. In parallel, a performance comparison with support vector machines (SVMs) was performed. The results obtained in the first experimental phase revealed that all the RNN networks evaluated reached a 100% accuracy, standing above the 93% achieved by the SVM. Regarding classification speed, LSTM ranked as the fastest architecture, recording a time of 0.12 ms, followed by GRU with 0.134 ms. Bidirectional recurrent neural networks showed a response time of 0.2 ms, while SVM had the longest time at 2.7 ms. In the second experimental phase, a slight decrease in the accuracy of the RNN models was observed, standing at 98.46% for LSTM, 96.38% for GRU, and 97.63% for the bidirectional network. The findings of this study highlight the effectiveness and speed of recurrent neural networks in the EMG signal classification task.

Poetry has a prominent history in Arabic literature. The classical Arabic poetry has 16 m that vary in rhythm and target purpose. Chanting a poem eloquently requires knowing the poem's meter and obtaining a diacritized version of its verses (letters inscribed with their short vowels); diacritics are often not inscribed in Arabic texts. This work proposes solutions to classify input Arabic text into the 16 poetry meters and prose. It also investigates the automatic diacritization of Arabic poetry. We adopt machine learning approach using a large dataset of 1657 k verses of poems and prose to develop neural networks to classify and diacritize Arabic poetry. We propose deep and narrow recurrent neural networks with bidirectional long short-term memory cells for solving these problems. The proposed model classifies the input text with an average accuracy of 97.27%, which is significantly higher than previous work. We also propose a solution that achieves an accuracy that approaches 100% when multiple verses of the same poem are available through predicting the class from the aggregate probabilities of the multiple verses. Diacritizing poetry is much harder than diacritizing prose due to the poet's meticulous selection of phrases and relaxation of some diacritization rules. [ABSTRACT FROM AUTHOR]

In this paper, the recurrent neural network structure of a bidirectional long short-term memory network (Bi-LSTM) with special memory cells that store information is used to characterize the deep features of the variation pattern between logging and seismic data. A mapping relationship model between high-frequency logging data and low-frequency seismic data is established via nonlinear mapping. The seismic waveform is infinitely approximated using the logging curve in the low-frequency band to obtain a nonlinear mapping model of this scale, which then stepwise approach the logging curve in the high-frequency band. Finally, a seismic-inversion method of nonlinear mapping multilevel well-seismic matching based on the Bi-LSTM network is developed. The characteristic of this method is that by applying the multilevel well-seismic matching process, the seismic data are stepwise matched to the scale range that is consistent with the logging curve. Further, the matching operator at each level can be stably obtained to effectively overcome the problems that occur in the well-seismic matching process, such as the inconsistency in the scale of two types of data, accuracy in extracting the seismic wavelet of the well-side seismic traces, and multiplicity of solutions. Model test and practical application demonstrate that this method improves the vertical resolution of inversion results, and at the same time, the boundary and the lateral characteristics of the sand body are well maintained to improve the accuracy of thin-layer sand body prediction and achieve an improved practical application effect. [ABSTRACT FROM AUTHOR]

Since the early days of spacecraft development, accurate and simple vibro-acoustic prediction of equipment on the spacecraft panels subjected to acoustic excitation has been conducted in order to mitigate the over-conservative environmental test conditions. The conventional prediction methods are based on numerical solution of equation of motion, such as FEM/BEM and SEA, the so-called deductive approach. However, in a spacecraft with complex structures, there are many structural and non-structural objects, such as wiring harnesses, connecting cables and electronic boards in the equipment, which are usually difficult to be modelled into these methods. These un-modeled objects are usually treated by uncertainty of models, which always results in overly conservative prediction. In order to mitigate this uncertainty caused by model limitation of deductive approach, this study proposes a more accurate and simple inductive approach for vibro-acoustic prediction using Gradient Boosting Decision Trees (GBDT), which is one of the machine learning algorithms based on measured data. In addition, in order to take into account the vibration modes of the structural panels and waveform trends of vibration response spectrum in the creation of the learning model, explanatory variables based on the design drawing information were added, and the concept of bidirectional recurrent neural networks (BRNN), which is used for predicting time histories waveforms, was incorporated. This approach was applied to the vibro-acoustic prediction using the measured data of the equipment on the spacecraft panels in the acoustic tests of 7 spacecrafts developed by JAXA, and the results showed that this approach can make a reasonable prediction with the uncertainty margin mitigated by about 2 to 4 dB compared with the conventional approach.

Recently, toxicity identification has become the most serious problem in online communities and social networking sites. Therefore, an automatic toxic identification system needs to be developed for preventing and limiting users from these online environments. In this paper, we present a multichannel convolutional bidirectional gated recurrent unit (MCBiGRU) for detecting toxic comments in a multilabel environment. The proposed model generates word vectors using pre-trained word embeddings. Moreover, this hybrid model extracts local features with many filters and different kernel sizes to model input words with long term dependency. We then integrate multiple channels with a fully connected layer, normalization layer, and an output layer with a sigmoid activation function for predicting multilabel categories. The experimental results indicate that the proposed MCBiGRU model outperforms in terms of multilabel metrics. [ABSTRACT FROM AUTHOR]

• This work did unprecedented analysis on the memory property of RNN cells. • The work develops extended LSTM that retains longer memory than previous RNN cells. • A new robust RNN model called dependent BRNN is proposed. • The superiority of the proposed solution is demonstrated experimentally. In this work, we first analyze the memory behavior in three recurrent neural networks (RNN) cells; namely, the simple RNN (SRN), the long short-term memory (LSTM) and the gated recurrent unit (GRU), where the memory is defined as a function that maps previous elements in a sequence to the current output. Our study shows that all three of them suffer rapid memory decay. Then, to alleviate this effect, we introduce trainable scaling factors that act like an attention mechanism to adjust memory decay adaptively. The new design is called the extended LSTM (ELSTM). Finally, to design a system that is robust to previous erroneous predictions, we propose a dependent bidirectional recurrent neural network (DBRNN). Extensive experiments are conducted on different language tasks to demonstrate the superiority of the proposed ELSTM and DBRNN solutions. The ELTSM has achieved up to 30% increase in the labeled attachment score (LAS) as compared to LSTM and GRU in the dependency parsing (DP) task. Our models also outperform other state-of-the-art models such as bi-attention [1] and convolutional sequence to sequence (convseq2seq) [2] by close to 10% in the LAS. [ABSTRACT FROM AUTHOR]

Copyright of Procesamiento del Lenguaje Natural is the property of Sociedad Espanola para el Procesamiento del Lenguaje Natural and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

This paper describes a Deep Belief Neural Network (DBNN) and Bidirectional Long-Short Term Memory (LSTM) hybrid used as an acoustic model for Speech Recognition. It was demonstrated by many independent researchers that DBNNs exhibit superior performance to other known machine learning frameworks in terms of speech recognition accuracy. Their superiority comes from the fact that these are deep learning networks. However, a trained DBNN is simply a feed-forward network with no internal memory, unlike Recurrent Neural Networks (RNNs) which are Turing complete and do posses internal memory, thus allowing them to make use of longer context. In this paper, an experiment is performed to make a hybrid of a DBNN with an advanced bidirectional RNN used to process its output. Results show that the use of the new DBNN-BLSTM hybrid as the acoustic model for the Large Vocabulary Continuous Speech Recognition (LVCSR) increases word recognition accuracy. However, the new model has many parameters and in some cases it may suffer performance issues in real-time applications. [ABSTRACT FROM AUTHOR]

Abstract: The supervised learning of recurrent neural networks well-suited for prediction of protein secondary structures from the underlying amino acids sequence is studied. Modular reciprocal recurrent neural networks (MRR-NN) are proposed to model the strong correlations between adjacent secondary structure elements. Besides, a multilayer bidirectional recurrent neural network (MBR-NN) is introduced to capture the long-range intramolecular interactions between amino acids in formation of the secondary structure. The final modular prediction system is devised based on the interactive integration of the MRR-NN and the MBR-NN structures to arbitrarily engage the neighboring effects of the secondary structure types concurrent with memorizing the sequential dependencies of amino acids along the protein chain. The advanced combined network augments the percentage accuracy (Q 3) to 79.36% and boosts the segment overlap (SOV) up to 70.09% when tested on the PSIPRED dataset in three-fold cross-validation. [ABSTRACT FROM AUTHOR]

Abstract: Protein domains contain information about the prediction of protein structure, function, evolution and design since the protein sequence may contain several domains with different or the same copies of the protein domain. In this study, we proposed an algorithm named SplitSSI-SVM that works with the following steps. First, the training and testing datasets are generated to test the SplitSSI-SVM. Second, the protein sequence is split into subsequence based on order and disorder regions. The protein sequence that is more than 600 residues is split into subsequences to investigate the effectiveness of the protein domain prediction based on subsequence. Third, multiple sequence alignment is performed to predict the secondary structure using bidirectional recurrent neural networks (BRNN) where BRNN considers the interaction between amino acids. The information of about protein secondary structure is used to increase the protein domain boundaries signal. Lastly, support vector machines (SVM) are used to classify the protein domain into single-domain, two-domain and multiple-domain. The SplitSSI-SVM is developed to reduce misleading signal, lower protein domain signal caused by primary structure of protein sequence and to provide accurate classification of the protein domain. The performance of SplitSSI-SVM is evaluated using sensitivity and specificity on single-domain, two-domain and multiple-domain. The evaluation shows that the SplitSSI-SVM achieved better results compared with other protein domain predictors such as DOMpro, GlobPlot, Dompred-DPS, Mateo, Biozon, Armadillo, KemaDom, SBASE, HMMPfam and HMMSMART especially in two-domain and multiple-domain. [Copyright &y& Elsevier]

• A new technique to analyze ECG signals named ECG language processing (ELP) is introduced. • ECG language processing (ELP) empowers computers to understand ECG signals in a way physicians do. • The proposed technique can be applied to applied to a variety of biomed- ical applications (e.g., classi cation and prediction). • The proposed technique facilitates developing interpretable algorithms. Background: A language is constructed of a finite/infinite set of sentences composing of words. Similar to natural languages, the Electrocardiogram (ECG) signal, the most common noninvasive tool to study the functionality of the heart and diagnose several abnormal arrhythmias, is made up of sequences of three or four distinct waves, including the P-wave, QRS complex, T-wave, and U-wave. An ECG signal may contain several different varieties of each wave (e.g., the QRS complex can have various appearances). For this reason, the ECG signal is a sequence of heartbeats similar to sentences in natural languages) and each heartbeat is composed of a set of waves (similar to words in a sentence) of different morphologies. Methods: Analogous to natural language processing (NLP), which is used to help computers understand and interpret the human's natural language, it is possible to develop methods inspired by NLP to aid computers to gain a deeper understanding of Electrocardiogram signals. In this work, our goal is to propose a novel ECG analysis technique, ECG language processing (ELP) , focusing on empowering computers to understand ECG signals in a way physicians do. Results: We evaluated the proposed approach on two tasks, including the classification of heartbeats and the detection of atrial fibrillation in the ECG signals. Overall, our technique resulted in better performance or comparable performance with smaller neural networks compared to other deep neural networks and existing algorithms. Conclusion: Experimental results on three databases (i.e., PhysioNet's MIT-BIH, MIT-BIH AFIB, and PhysioNet Challenge 2017 AFIB Dataset databases) reveal that the proposed approach as a general idea can be applied to a variety of biomedical applications and can achieve remarkable performance. [ABSTRACT FROM AUTHOR]

Atrial fibrillation (AF) is one of the most prevalent cardiac arrhythmias that affects the lives of many people around the world and is associated with a five-fold increased risk of stroke and mortality. Like other problems in the healthcare domain, artificial intelligence (AI)-based models have been used to detect AF from patients' ECG signals. The cardiologist level performance in detecting this arrhythmia is often achieved by deep learning-based methods, however, they suffer from the lack of interpretability. In other words, these approaches are unable to explain the reasons behind their decisions. The lack of interpretability is a common challenge toward a wide application of machine learning (ML)-based approaches in the healthcare which limits the trust of clinicians in such methods. To address this challenge, we propose HAN-ECG, an interpretable bidirectional-recurrent-neural-network-based approach for the AF detection task. The HAN-ECG employs three attention mechanism levels to provide a multi-resolution analysis of the patterns in ECG leading to AF. The detected patterns by this hierarchical attention model facilitate the interpretation of the neural network decision process in identifying the patterns in the signal which contributed the most to the final detection. Experimental results on two AF databases demonstrate that our proposed model performs better than the existing algorithms. Visualization of these attention layers illustrates that our proposed model decides upon the important waves and heartbeats which are clinically meaningful in the detection task (e.g., absence of P-waves, and irregular R-R intervals for the AF detection task)., (Copyright © 2020 Elsevier Ltd. All rights reserved.)

Filtering of Protein Secondary Structure Prediction (PSSP) aims to provide physicochemically realistic results, while it usually improves the predictive performance. We performed a comparative study on this challenging problem, utilizing both machine learning techniques and empirical rules and we found that combinations of the two lead to the highest improvement. [ABSTRACT FROM PUBLISHER]