Search

Background: Biomedical text mining is a technique that extracts essential information from scientific articles using named entity recognition (NER). Traditional NER methods rely on dictionaries, rules, or curated corpora, which may not always be accessible. To overcome these challenges, deep learning (DL) methods have emerged. However, DL-based NER methods may need help identifying long-distance relationships within text and require significant annotated datasets. Results: This research has proposed a novel model to address the challenges in natural language processing. The Improved Green anaconda-assisted Bi-GRU based Hierarchical ResNet BNER model (IGa-BiHR BNERM) is the model. IGa-BiHR BNERM model has shown promising results in accurately identifying named entities. The MACCROBAT dataset was obtained from Kaggle and underwent several pre-processing steps such as Stop Word Filtering, WordNet processing, Removal of non-alphanumeric characters, stemming Segmentation, and Tokenization, which is standardized and improves its quality. The pre-processed text was fed into a feature extraction model like the Robustly Optimized BERT –Whole Word Masking model. This model provides word embeddings with semantic information. Then, the BNER process utilized an Improved Green Anaconda-assisted Bi-GRU-based Hierarchical ResNet BNER model (IGa-BiHR BNERM). Conclusion: To improve the training phase of the IGa-BiHR BNERM, the Improved Green Anaconda Optimization technique was used to select optimal weight parameter coefficients for training the model parameters. After the model was tested using the MACCROBAT dataset, it outperformed previous models with a tremendous accuracy rate of 99.11%. This model effectively and accurately identifies biomedical names within the text, significantly advancing this field. [ABSTRACT FROM AUTHOR]

Sentiment analysis, a crucial aspect of Natural Language Processing (NLP), aims to extract subjective information from textual data. With the proliferation of social media platforms like Twitter, accurately determining public sentiment has become increasingly important for businesses, policymakers, and researchers. This study introduces the Dual Joint Classifier (DJC), which integrates the strengths of RoBERTa and BERT architectures. The DJC model leverages Bidirectional Gated Recurrent Units (BiGRU) and Bidirectional Long Short-Term Memory (BiLSTM) layers to capture complex sequential dependencies and nuanced sentiment expressions. Advanced training techniques such as Focal Loss and Hard Sample Mining address class imbalance and improve model robustness. To further validate the DJC model's robustness, the larger TweetEval Sentiment dataset was also included, on which DJC outperformed conventional models despite increased training time. Evaluations were conducted on the Twitter US Airlines and Apple Twitter Sentiment datasets to verify experiments. The DJC model achieved 87.22% and 93.87% accuracies, respectively, and demonstrated improvement over other models like RoBERTa-GLG, BiLSTM(P), and SVM. These results highlight the DJC model's effectiveness in handling diverse sentiment analysis tasks and its potential for real-world applications. [ABSTRACT FROM AUTHOR]

Copyright of Baghdad Science Journal is the property of Republic of Iraq Ministry of Higher Education & Scientific Research (MOHESR) 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.)

Natural ecosystems have been facing a major threat due to deforestation and forest fires for the past decade. These environmental challenges have led to significant biodiversity loss, disruption of natural habitats, and adverse effects on climate change. The integration of Artificial Intelligence (AI) and Optimization techniques has made a revolutionary impact in disaster management, offering new avenues for early detection and prevention strategies. Therefore, to prevent the outbreak of a forest fire, an efficient forest fire diagnosis and aversion system is needed. To address this problem, an IoT-based Artificial Intelligence (AI) technique for forest fire detection has been proposed. This system leverages the Internet of Things (IoT) to collect real-time data from various sensors deployed in forest areas, providing continuous monitoring and early warning capabilities. Several researchers have contributed different techniques to predict forest fires at various remote locations, highlighting the importance of innovative approaches in this field. The proposed work involves object detection, which is facilitated by EfficientDet, a state-of-the-art object detection model known for its accuracy and efficiency. EfficientDet enables the system to accurately identify potential fire outbreaks by analyzing visual data from the sensors. To facilitate efficient detection at the outbreak of forest fires, a bi-directional gated recurrent neural network (Bi-GRU-NN) is needed. This neural network architecture is capable of processing sequential data from multiple directions, enhancing the system's ability to predict the spread and intensity of fires. Crow Search Optimization (CSO) and fractional calculus are used to create an optimal solution in the proposed crow search fractional calculus optimization (CSFCO) algorithm for deep learning. CSO is inspired by the intelligent foraging behavior of crows, and when combined with fractional calculus, it provides a robust optimization framework that improves the accuracy and efficiency of the AI model. Experimental analysis shows that the proposed technique outperformed the other existing traditional approaches with an accuracy of 99.32% and an error rate of 0.12%. These results demonstrate the effectiveness of the integrated AI and optimization techniques in enhancing forest fire detection and prevention. The high accuracy and low error rate underscore the potential of this system to be a valuable tool in mitigating the risks associated with forest fires, ultimately contributing to the preservation of natural ecosystems. [ABSTRACT FROM AUTHOR]

ABSTRACT Accurate forecasts of renewable and nonrenewable energy output are essential for meeting global energy needs and resolving environmental issues. Energy sources like the sun and wind are variable, making forecasting difficult. Changes in weather, demand, and energy policy exacerbate this unpredictability. These challenges will be addressed by the bidirectional gated recurrent unit (Bi‐GRU) model, which forecasts power‐generating outcomes more efficiently. The investigation is done over a health data set from 2000 to 2023, including the energy states of the United Kingdom, Finland, Germany, and Switzerland. The comparison of our model (Bi‐GRU) performance with other popular models, including bidirectional long short‐term memory (Bi‐LSTM), ensemble techniques combining convolutional neural networks (CNN) and Bi‐LSTM, and CNNs, make the study more interesting. The performance remains better with a mean absolute percentage error (MAPE) of 2.75%, root mean square error (RMSE) of 0.0414, mean squared error (MSE) of 0.0017, and authentify that Bi‐GRU performs much better than others. This model's superior prediction accuracy significantly enhances our ability to forecast renewable and nonrenewable energy outputs in European states, contributing to more effective energy management strategies.

Many areas of image identification and classification for medical imaging diagnostics have greatly benefited from deep learning (DL). Diabetic retinopathy (DR) will become the most common cause of blindness worldwide, making diabetes a major threat to public health. This research proposes an automated identification system using deep recurrent neural networks (RNNs) to identify and classify four categories of diabetic eye diseases: DR, cataract, glaucoma, and diabetic macular edema (DME). We use three different model architectures based on RNN and their types, we called our proposed system RNN Diabetic framework. These models are combined with one of the commonly used architectures that support sufficient accuracy and speed for the model which is residual network (ResNet)152V2. The three model architectures are RNN+ResNet152V2, gated recurrent unit (GRU)+ResNet152V2, and bidirectional GRU (Bi-GRU)+ResNet152V2. The proposed models were assessed as collected datasets: DIARETDB0, DIARETDB1, messidor, HEI-MED, ocular, and retina. A full analysis and evaluation of these three deep RNN architectures are presented. The experiments showed that the Bi-GRU+ResNet152V2 model worked better than the other two proposed models. In addition, we compare these three proposed models with the previous studies and find that the proposed Bi-GRU+ResNet152V2 model achieves the highest results with accuracy equal to 99.8%, 98.1% sensitivity, 98.6% specificity, 99.8% precision, 99.8% F1 score, and 99.8% areas under the curve (AUC). [ABSTRACT FROM AUTHOR]

In marine ranching aquaculture, dissolved oxygen (DO) is a crucial parameter that directly impacts the survival, growth, and profitability of cultured organisms. To effectively guide the early warning and regulation of DO in aquaculture waters, this study proposes a hybrid model for spatiotemporal DO prediction named PCA-ISSA-DAM-Bi-GRU. Firstly, principal component analysis (PCA) is applied to reduce the dimensionality of the input data and eliminate data redundancy. Secondly, an improved sparrow search algorithm (ISSA) based on multi strategy fusion is proposed to enhance the optimization ability and convergence speed of the standard SSA by optimizing the population initialization method, improving the location update strategies for discoverers and followers, and introducing a Cauchy-Gaussian mutation strategy. Thirdly, a feature and temporal dual attention mechanism (DAM) is incorporated to the baseline temporal prediction model Bi-GRU to construct a feature extraction network DAM-Bi-GRU. Fourthly, the ISSA is utilized to optimize the hyperparameters of DAM-Bi-GRU. Finally, the proposed model is trained, validated, and tested using water quality and meteorological parameter data collected from a self-built LoRa+5G-based marine ranching aquaculture monitoring system. The results show that: (1) Compared with the baseline model Bi-GRU, the addition of PCA, ISSA and DAM module can effectively improve the prediction performance of the model, and their fusion is effective; (2) ISSA demonstrates superior capability in optimizing model hyperparameters and convergence speed compared to traditional methods such as standard SSA, genetic algorithm (GA), and particle swarm optimization (PSO); (3) The proposed hybrid model achieves a root mean square error (RMSE) of 0.2136, a mean absolute percentage error (MAPE) of 0.0232, and a Nash efficient (NSE) of 0.9427 for DO prediction, outperforming other similar data-driven models such as IBAS-LSTM and IDA-GRU. The prediction performance of the model meets the practical needs of precise DO prediction in aquaculture. [ABSTRACT FROM AUTHOR]

Simple Summary: Phosphorylation is a crucial process that regulates various cellular activities. Detecting phosphorylation sites, especially in cells infected by the SARS-CoV-2 virus, is challenging due to technical limitations. To address this, we developed GBMPhos, an advanced tool combining deep learning techniques, to accurately identify these sites. GBMPhos outperformed traditional methods and current state-of-the-art approaches in identifying phosphorylation sites. We have developed a free web server, which helps researchers gain a better understanding of protein modifications during a SARS-CoV-2 infection, potentially aiding in the development of therapeutic strategies and contributing to the fight against COVID-19. Phosphorylation, a reversible and widespread post-translational modification of proteins, is essential for numerous cellular processes. However, due to technical limitations, large-scale detection of phosphorylation sites, especially those infected by SARS-CoV-2, remains a challenging task. To address this gap, we propose a method called GBMPhos, a novel method that combines convolutional neural networks (CNNs) for extracting local features, gating mechanisms to selectively focus on relevant information, and a bi-directional gated recurrent unit (Bi-GRU) to capture long-range dependencies within protein sequences. GBMPhos leverages a comprehensive set of features, including sequence encoding, physicochemical properties, and structural information, to provide an in-depth analysis of phosphorylation sites. We conducted an extensive comparison of GBMPhos with traditional machine learning algorithms and state-of-the-art methods. Experimental results demonstrate the superiority of GBMPhos over existing methods. The visualization analysis further highlights its effectiveness and efficiency. Additionally, we have established a free web server platform to help researchers explore phosphorylation in SARS-CoV-2 infections. The source code of GBMPhos is publicly available on GitHub. [ABSTRACT FROM AUTHOR]

Deoxyribonucleic acid (DNA) profiling is an important branch of forensic science that aids in the identification of missing people, particularly in mass disasters. This study presents an artificial intelligence system that utilizes DNA-Short Tandem Repeat (STR) data to identify victims using Deep Neural Network (DNN), Gated Recurrent Unit (GRU), and Bidirectional GRU (Bi-GRU) deep learning models. The identification of STR information for living family members, such as parents or brothers, poses a significant challenge in victim identification. Familial data are artificially generated based on the actual data of distinct Iraqi individuals from the province of Al-Najaf. Two people are selected as male and female to create a family of 10 members. As a result of this action, 151,580 individuals were generated from 106 different people, which helps to overcome the lack of datasets caused by restrictive policies and the confidentiality of familial datasets in Iraq. These datasets are prepared and formatted for training deep learning models. Based on various reference datasets, the models are built to handle five different scenarios where both parents are alive, only one parent is alive, or the siblings are available for reference. The three models' performances were compared: Bi-GRU performed the best, with a loss of 0.0063 and an accuracy of 0.9979, followed by GRU with a loss of 0.0102 and an accuracy of 0.9964, and DNN with a loss of 0.2276 and an accuracy of 0.9174. The evaluation makes use of a confusion matrix and receiver operating characteristic curve. Based on the literature, this is the first attempt to introduce deep learning in DNA profiling, which reduces both time and effort. Despite the fact that the proposed deep learning models have good results in identifying missing persons according to their families, these models have limitations that can be confined to the availability of familial DNA profiles. The system doesn't work well if no relative samples are available as references, such as a father, mother, or brother. In the future, DNN, GRU, and Bi-GRU models will be applied to mini-STR sequences that are used in cases of degraded victims of incomplete STR sequences. [ABSTRACT FROM AUTHOR]

Sentiment analysis, a crucial aspect of Natural Language Processing (NLP), aims to extract subjective information from textual data. With the proliferation of social media platforms like Twitter, accurately determining public sentiment has become increasingly important for businesses, policymakers, and researchers. This study introduces the Dual Joint Classifier (DJC), which integrates the strengths of RoBERTa and BERT architectures. The DJC model leverages Bidirectional Gated Recurrent Units (BiGRU) and Bidirectional Long Short-Term Memory (BiLSTM) layers to capture complex sequential dependencies and nuanced sentiment expressions. Advanced training techniques such as Focal Loss and Hard Sample Mining address class imbalance and improve model robustness. To further validate the DJC model’s robustness, the larger TweetEval Sentiment dataset was also included, on which DJC outperformed conventional models despite increased training time. Evaluations were conducted on the Twitter US Airlines and Apple Twitter Sentiment datasets to verify experiments. The DJC model achieved 87.22% and 93.87% accuracies, respectively, and demonstrated improvement over other models like RoBERTa-GLG, BiLSTM(P), and SVM. These results highlight the DJC model’s effectiveness in handling diverse sentiment analysis tasks and its potential for real-world applications.

In various domains, including everyday activities, agricultural practices, and medical treatments, the escalating challenge of antibiotic resistance poses a significant concern. Traditional approaches to studying antibiotic resistance genes (ARGs) often require substantial time and effort and are limited in accuracy. Moreover, the decentralized nature of existing data repositories complicates comprehensive analysis of antibiotic resistance gene sequences. In this study, we introduce a novel computational framework named TGC-ARG designed to predict potential ARGs. This framework takes protein sequences as input, utilizes SCRATCH-1D for protein secondary structure prediction, and employs feature extraction techniques to derive distinctive features from both sequence and structural data. Subsequently, a Siamese network is employed to foster a contrastive learning environment, enhancing the model's ability to effectively represent the data. Finally, a multi-layer perceptron (MLP) integrates and processes sequence embeddings alongside predicted secondary structure embeddings to forecast ARG presence. To evaluate our approach, we curated a pioneering open dataset termed ARSS (Antibiotic Resistance Sequence Statistics). Comprehensive comparative experiments demonstrate that our method surpasses current state-of-the-art methodologies. Additionally, through detailed case studies, we illustrate the efficacy of our approach in predicting potential ARGs. [ABSTRACT FROM AUTHOR]

Automatically detecting unusual behavior in a crowded environment greatly improves public safety. Unusual behaviors are those that depend on the rules established in the environment under consideration and cannot be properly described. This paper proposes a new deep feature-based crowd anomaly detection method. Priorly, the input image is preprocessed using the Weiner filtering method. Subsequently, AlexNet, VGGnet, and ResNet-based deep features are extracted. During this process, all three models were optimally tuned. For optimization, a new hybrid optimization method called Hybrid COOT and Bald Eagle with Bernoulli Map Evaluation (HCBEBME) is introduced in this work. This improves the performance of extracting features from the input. Finally, based on the proposed feature set, anomalies are detected by the hybrid detection model that combines LSTM and Bi-GRU models, respectively. Finally, the performance of the proposed model is validated over the conventional models. The detection accuracy of the suggested approach is 96.59%, whereas the minimal accuracy scores for the other methods BRCASO, GNNN, CNN-BILSTM, LSTM, BIGRU, BILSTM, CNN, and RNN are 93.63%, 79.16%, 87.57%, 73.85%, 70.59%, 77.08%, 81.44%, and 84.66% respectively. [ABSTRACT FROM AUTHOR]

ObjectiveThe widespread prevalence of sclerotinia disease poses a significant challenge to the cultivation and supply of oilseed rape, not only results in substantial yield losses and decreased oil content in infected plant seeds but also severely impacts crop productivity and quality, leading to significant economic losses. To solve the problems of complex operation, environmental pollution, sample destruction and low detection efficiency of traditional chemical detection methods, a Bi-directional Gate Recurrent Unit (Bi-GRU) model based on space-spectrum feature fusion was constructed to achieve hyperspectral images (HSIs) segmentation of oilseed rape sclerotinia infected area.MethodsThe spectral characteristics of sclerotinia disease from a spectral perspective was initially explored. Significantly varying spectral reflectance was notably observed around 550 nm and within the wavelength range of 750-1 000 nm at different locations on rapeseed leaves. As the severity of sclerotinia infection increased, the differences in reflectance at these wavelengths became more pronounced. Subsequently, a rapeseed leaf sclerotinia disease dataset comprising 400 HSIs was curated using an intelligent data annotation tool. This dataset was divided into three subsets: a training set with 280 HSIs, a validation set with 40 HSIs, and a test set with 80 HSIs. Expanding on this, a 7×7 pixel neighborhood was extracted as the spatial feature of the target pixel, incorporating both spatial and spectral features effectively. Leveraging the Bi-GRU model enabled simultaneous feature extraction at any point within the sequence data, eliminating the impact of the order of spatial-spectral data fusion on the model's performance. The model comprises four key components: an input layer, hidden layers, fully connected layers, and an output layer. The Bi-GRU model in this study consisted of two hidden layers, each housing 512 GRU neurons. The forward hidden layer computed sequence information at the current time step, while the backward hidden layer retrieves the sequence in reverse, incorporating reversed-order information. These two hidden layers were linked to a fully connected layer, providing both forward and reversed-order information to all neurons during training. The Bi-GRU model included two fully connected layers, each with 1 000 neurons, and an output layer with two neurons representing the healthy and diseased classes, respectively.Results and DiscussionsTo thoroughly validate the comprehensive performance of the proposed Bi-GRU model and assess the effectiveness of the spatial-spectral information fusion mechanism, relevant comparative analysis experiments were conducted. These experiments primarily focused on five key parameters—ClassAP(1), ClassAP(2), mean average precision (mAP), mean intersection over union (mIoU), and Kappa coefficient—to provide a comprehensive evaluation of the Bi-GRU model's performance. The comprehensive performance analysis revealed that the Bi-GRU model, when compared to mainstream convolutional neural network (CNN) and long short-term memory (LSTM) models, demonstrated superior overall performance in detecting rapeseed sclerotinia disease. Notably, the proposed Bi-GRU model achieved an mAP of 93.7%, showcasing a 7.1% precision improvement over the CNN model. The bidirectional architecture, coupled with spatial-spectral fusion data, effectively enhanced detection accuracy. Furthermore, the study visually presented the segmentation results of sclerotinia disease-infected areas using CNN, Bi-LSTM, and Bi-GRU models. A comparison with the Ground-Truth data revealed that the Bi-GRU model outperformed the CNN and Bi-LSTM models in detecting sclerotinia disease at various infection stages. Additionally, the Dice coefficient was employed to comprehensively assess the actual detection performance of different models at early, middle, and late infection stages. The dice coefficients for the Bi-GRU model at these stages were 83.8%, 89.4% and 89.2%, respectively. While early infection detection accuracy was relatively lower, the spatial-spectral data fusion mechanism significantly enhanced the effectiveness of detecting early sclerotinia infections in oilseed rape.ConclusionsThis study introduces a Bi-GRU model that integrates spatial and spectral information to accurately and efficiently identify the infected areas of oilseed rape sclerotinia disease. This approach not only addresses the challenge of detecting early stages of sclerotinia infection but also establishes a basis for high-throughput non-destructive detection of the disease.

Copyright of Journal of Data Acquisition & Processing / Shu Ju Cai Ji Yu Chu Li is the property of Editorial Department of Journal of Nanjing University of Aeronautics & Astronautics 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.)

Deep learning promotes the breakthrough of emotion recognition in many fields, especially speech emotion recognition (SER). As an important part of speech emotion recognition, the most relevant acoustic feature extraction has always attracted the attention of existing researchers. Aiming at the problem that the emotional information contained in the current speech signals is distributed dispersedly and cannot comprehensively integrate local and global information, this paper presents a network model based on a gated recurrent unit (GRU) and multi-head attention. We evaluate our proposed emotion model on the IEMOCAP and Emo-DB corpora. The experimental results show that the network model based on Bi-GRU and multi-head attention is significantly better than the traditional network model at detecting multiple evaluation indicators. At the same time, we also apply the model to a speech sentiment analysis task. On the CH-SIMS and MOSI datasets, the model shows excellent generalization performance. [ABSTRACT FROM AUTHOR]

Copyright of Smart Agriculture is the property of Smart Agriculture Editorial Office 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.)

Amidst the ongoing wave of economic globalization, the societal demand for English proficiency is escalating, particularly for individuals adept in practical applications of the language. Recognizing the pivotal role of English reading as a cornerstone in language acquisition, there arises a need for personalized approaches tailored to individual interests, thereby necessitating an in-depth analysis of text emotions. Addressing the challenges in text classification within English reading courses, this study presents a novel method for text emotion analysis. Integrating sentiment dictionaries with BI-GRU networks, the proposed approach significantly enhances the efficiency of text emotion recognition while simultaneously fostering students’ engagement. By segmenting the emotion dictionary based on polarity and extracting pertinent features, the study amalgamates these with BI-GRU features at the feature level. This fusion facilitates emotion classification within reading texts through sophisticated activation functions. Notably, the precision of recognizing positive, negative, and neutral emotions reaches an impressive 92.5%, marking a notable improvement over methods devoid of dictionary feature integration. This framework offers novel insights for future English reading material development and intelligent learning strategies to bolster student enthusiasm and chart a promising trajectory for cultivating practical English talents.

The widespread use of the Internet and the tremendous growth of social media have enabled people to connect with each other worldwide. Individuals are free to express themselves online, sharing their photos, videos, and text messages globally. However, such freedom sometimes leads to misuse, as some individuals exploit this platform by posting hateful and abusive comments on forums. The proliferation of abusive language on social media negatively impacts individuals and groups, leading to emotional distress and affecting mental health. It is crucial to automatically detect and filter such abusive content in order to effectively tackle this challenging issue. Detecting abusive language in text messages is challenging due to intentional word concealment and contextual complexity. To counter abusive speech on social media, we need to explore the potential of machine learning (ML) and deep learning (DL) models, particularly those equipped with attention mechanisms. In this study, we utilized popular ML and DL models integrated with attention mechanism to detect abusive language in Urdu text. Our methodology involved employing Count Vectorizer and Term Frequency-Inverse Document Frequency (TF/IDF) to extract n-grams at the word level: Unigrams (Uni), Bigrams (Bi), Trigrams (Tri), and their combination (Uni + Bi + Tri). Initially, we evaluated four traditional ML models—Logistic Regression (LR), Gaussian Naïve Bayes (NB), Support Vector Machine (SVM), and Random Forest (RF)—on both proposed and established datasets. The results highlighted that RF model outperformed other conventional models in terms of accuracy, precision, recall, and F1-measure on both datasets. In our implementation of deep learning models, we employed various models integrated with custom fastText and Word2Vec embeddings, each equipped with an attention layer, except for the Convolutional Neural Network (CNN). Our findings indicated that the Bidirectional Long Short-Term Memory (Bi-LSTM) + attention model, utilizing custom Word2Vec embeddings, exhibited improved performance in detecting abusive language on both datasets.

In the rumor detection based on crowd intelligence, the crowd behavior is constructed as a graph model or probability mode. The detection of rumors is achieved through the collaborative utilization of data and knowledge. Aiming at the problems of insufficient feature extraction ability and data redundancy of current rumor detection methods based on deep learning model, a social network rumor detection method based on bidirectional gated recurrent unit (Bi-GRU) and double self-attention (DSA) mechanism is suggested. First, a combination of application program interface and third-party crawler approach is used to obtain microblogging data from publicly available fake microblogging information pages, including both rumor and non-rumor information. Second, Bi-GRU is used to capture the tendency of medium- and long-term dependence of data and is flexible enough to deal with variable length input. Finally, the DSA mechanism is introduced to help reduce the redundant information in the dataset, thereby enhancing the model’s efficacy. The results of the experiments indicate that the proposed method outperforms existing advanced methods by at least 0.114, 0.108, 0.064, and 0.085 in terms of accuracy, precision, recall, and F1-scores, respectively. Therefore, the proposed method can significantly enhance the ability of social network rumor detection.

There are many methods to diagnose heart disease; the most effective way is to analyze electrocardiogram (ECG) signals. Generally, the automatic classification techniques based on ECG analysis consist of three steps: data preprocessing, feature extraction, and classification. This study designed eight hybrid model architectures using several types of deep neural networks, including Convolution Neural Network (CNN), Gated Recurrent Unit (GRU), and Bidirectional GRU (Bi-GRU), four of them without Fast Fourier Transform (FFT) and the rest using FFT. Firstly, the MIT-BIH arrhythmia database is cleaned using the wavelet (WT) thresholding method that separates the combined noise and signal frequencies, making it ideal for processing nonstationary ECG signals. Additionally, the imbalance problem in this database was addressed using the synthetic minority over-sampling technique (SMOTE), which is more suitable for medical data than random synthesis methods. Secondly, hybrid models FFT-CNN, FFT-GRU, FFT-CNN-GRU, and FFT-CNN-Bi-GRU are constructed using the new proposed architecture by concatenating resultant features from two paths, the first path using ECG in the time domain and the second path using the resultant spectrum of ECG from FFT as input. A comparative study of the performance of all models was created in terms of accuracy, training time, number of trainable parameters, and robustness against noise. The results show that the proposed CNN, GRU, CNN-GRU, and CNN-Bi-GRU models without WT and FFT achieved 90%, 93%, 95%, and 96% accuracies, while the proposed FFT-CNN, FFT-GRU, FFT-CNN-GRU, and FFT-CNN-Bi-GRU models achieved 97%, 95%, 96%, and 97% accuracies with WT. So, the proposed FFT-CNN model was the best, with less training time and parameters than other models, which significantly impacts designing a high-efficiency model with less complexity for a practical medical diagnosis system. On the other hand, using FFT improved all models' performance, accuracy and robustness against noise. [ABSTRACT FROM AUTHOR]

Time series data sensed by underwater wireless sensor networks (UWSNs) play a crucial role in prediction and decision-making in marine applications. Unfortunately, equipment and environmental precision and interference problems in UWSNs may lead to a large amount of missing data in a specific time period. In this work, we propose a multi-head attention-based sequence-to-sequence model (MSSM) for reconstructing continuous missing data. It can reduce the negative impact of missing data due to the harsh underwater communication environment. MSSM has a dual encoder architecture that can process known data on both sides of missing values. Multi-head self-attention mechanism and bidirectional gate recurrent unit (Bi-GRU) can thoroughly learn the temporal patterns and the inter-sequence dependencies; moreover, soft thresholding can also reduce noise interference. Datasets are used to test the performance, and experimental results show that metrics are lower than other relevant alternatives, demonstrating that MSSM is an effective model with solid generalization ability. [ABSTRACT FROM AUTHOR]

The relationships among different subregions in facial images and their varying contributions to facial expression recognition indicate that using a fixed subregion weighting scheme would result in a substantial loss of valuable information. To address this issue, we propose a facial expression recognition network called BGA-Net, which combines bidirectional gated recurrent units (BiGRUs) with an attention mechanism. Firstly, a convolutional neural network (CNN) is employed to extract feature maps from facial images. Then, a sliding window cropping strategy is applied to divide the feature maps into multiple subregions. The BiGRUs are utilized to capture the dependencies among these subregions. Finally, an attention mechanism is employed to adaptively focus on the most discriminative regions. When evaluated on CK+, FER2013, and JAFFE datasets, our proposed method achieves promising results. [ABSTRACT FROM AUTHOR]

The technology for human activity recognition has diverse applications within the Internet of Things spectrum, including medical sensing, security measures, smart home systems, and more. Predominantly, human activity recognition methods have relied on contact sensors, and some research uses inertial sensors embedded in smartphones or other devices, which present several limitations. Additionally, most research has concentrated on recognizing discrete activities, even though activities in real-life scenarios tend to be continuous. In this paper, we introduce a method to classify continuous human activities, such as walking, running, squatting, standing, and jumping. Our approach hinges on the micro-Doppler (MD) features derived from continuous-wave radar signals. We first process the radar echo signals generated from human activities to produce MD spectrograms. Subsequently, a bidirectional gate recurrent unit (Bi-GRU) network is employed to train and test these extracted features. Preliminary results highlight the efficacy of our approach, with an average recognition accuracy exceeding 90%. [ABSTRACT FROM AUTHOR]

Offensive language in social media affects the social experience of individuals and groups and hurts social harmony and moral values. Therefore, in recent years, the problem of offensive language detection has attracted the attention of many researchers. However, the primary research currently focuses on detecting English offensive language, while few studies on the Chinese language exist. In this paper, we propose an innovative approach to detect Chinese offensive language. First, unlike previous approaches, we utilized both RoBERTa's sentence-level and word-level embedding, combining the sentence embedding and word embedding of RoBERTa's model, bidirectional GRU, and multi-head self-attention mechanism. This feature fusion allows the model to consider sentence-level and word-level semantic information at the same time so as to capture the semantic information of Chinese text more comprehensively. Second, by concatenating the output results of multi-head attention with RoBERTa's sentence embedding, we achieved an efficient fusion of local and global information and improved the representation ability of the model. The experiments showed that the proposed model achieved 82.931% accuracy and 82.842% F1-score in Chinese offensive language detection tasks, delivering high performance and broad application potential. [ABSTRACT FROM AUTHOR]

Cultural communication plays a vital role in social development and human interaction. Red culture, as an integral part of China’s revolutionary history and socialist construction, holds significant meaning and exerts a wide influence. However, in the era of information technology, effectively disseminating red culture and stimulating public interest and participation has become an urgent challenge. In this study, we use the advanced deep learning tech to explore the use of multimodal data fusion for enhancing the effectiveness and impact of red culture communication. Specifically, we extract text features and image features from users’ browsing information using BI-GRU and CNN, respectively. These features are then fused with user portraits to create a multi-source information fusion vector. Subsequently, we employ a BPNN (Backpropagation Neural Network) to perform user interest classification based on the fused features. Experimental results demonstrate that our proposed user recognition framework achieves an average recognition rate of 95.4% across three types of users, indicating high accuracy. Therefore, the user interest classification model, incorporating fused multi-features, presented in this paper offers a promising approach for future red culture communication, as well as user intelligent recommendation and analysis.

Wireless energy transfer technology (WET)-enabled mobile charging provides an innovative strategy for energy replenishment in wireless rechargeable sensor networks (WRSNs), where the mobile charger (MC) can charge the sensors sequentially by WET according to the mobile charging scheduling scheme. Although there have been fruitful studies, they usually assume that all sensors will be charged fully once scheduled or charged to a fixed percentage determined by a charging upper threshold, resulting in low charging performance as they cannot adjust the charging operation on each sensor adaptively according to the real-time charging demands. To tackle this challenge, we first formulate the mobile charging scheduling as a joint mobile charging sequence scheduling and charging upper threshold control problem (JSSTC), where the charging upper threshold of each sensor can adjust adaptively. Then, we propose a novel multi-discrete action space deep Q-network approach for JSSTC (MDDRL-JSSTC), where MC is regarded as an agent exploring the environment. The state information observed by MC at each time step is encoded to construct a high-dimensional vector. Furthermore, a two-dimensional action is mapped to the charging destination of MC and the corresponding charging upper threshold at the next time step, using bidirectional gated recurrent units (Bi-GRU). Finally, we conduct a series of experiments to verify the superior performance of the proposed approach in prolonging the lifetime compared with the state-of-the-art approaches. [ABSTRACT FROM AUTHOR]

Sentiment analysis of social media text containing opinions about the product, event, or service is used in various applications like election results prediction, product endorsement, and many more. Sarcasm is a form of sentiment in which people use positive words to express negative feelings. While communicating verbally, people express sarcasm using hand gestures, facial expressions, and eye movements. These clues are missing in text data, making sarcasm detection challenging. Because of these challenges, scholars are interested in detecting sarcasm in social media texts. The feature extraction technique is an important component in a sarcasm detection model. Most solutions use GloVe, word2vec, or general-purpose pre-trained models for feature extraction. The GloVe/word2vec techniques ignore words that are not present in their vocabulary leading to information loss, require more extensive data for training and generating exact vectors, and ignore contextual information. A general-purpose pre-trained model overcomes the limitations of GloVe/word2vec models but cannot learn features from the social media text due to informal grammar, abbreviations, and irregular vocabulary. In this view, the BERTweet model (trained on social media text) is applied to generate sentence-level semantics and contextual features. The Bi-GRU model processes these features to learn long-distance dependencies from both directions (forward and backward), and the self-attention layer is applied on top of the Bi-GRU model to remove redundant and irrelevant information. This work presents a hybrid method called B2GRUA that combines the strengths of the BERTweet pre-trained model, bi-directional gated recurrent unit and attention mechanism (Bi-GRUAM) for classifying text into sarcastic/non-sarcastic. The efficacy of the proposed model is evaluated on three benchmark datasets, namely SemEval 2018 Task 3.A, iSarcasm, and 2020 shared sarcasm detection task (Twitter data). It is observed from the results that the proposed model out-performed state-of-the-art models on all the datasets (24% better on the iSarcasm dataset and around 2% on both the 2020 shared sarcasm detection task and SemEval 2018 Task 3.A dataset). ANOVA one-way test is applied to validate the results statistically. [ABSTRACT FROM AUTHOR]

As the monitor probes are used more and more widely these days, the task of detecting abnormal behaviors in surveillance videos has gained widespread attention. The generalization ability and parameter overhead of the model affect how accurate the detection result is. To deal with the poor generalization ability and high parameter overhead of the model in existing anomaly detection methods, we propose a three-dimensional multi-branch convolutional fusion network, named "Branch-Fusion Net". The network is designed with a multi-branch structure not only to significantly reduce parameter overhead but also to improve the generalization ability by understanding the input feature map from different perspectives. To ignore useless features during the model training, we propose a simple yet effective Channel Spatial Attention Module (CSAM), which sequentially focuses attention on key channels and spatial feature regions to suppress useless features and enhance important features. We combine the Branch-Fusion Net and the CSAM as a local feature extraction network and use the Bi-Directional Gated Recurrent Unit (Bi-GRU) to extract global feature information. The experiments are validated on a self-built Crimes-mini dataset, and the accuracy of anomaly detection in surveillance videos reaches 93.55% on the test set. The result shows that the model proposed in the paper significantly improves the accuracy of anomaly detection in surveillance videos with low parameter overhead. [ABSTRACT FROM AUTHOR]

Traffic flow prediction plays an important role in intelligent transportation systems. To accurately capture the complex non-linear temporal characteristics of traffic flow, this paper adopts a Bi-directional Gated Recurrent Unit (Bi-GRU) model in traffic flow prediction. Compared to Gated Recurrent Unit (GRU), which can memorize information from the previous sequence, this model can memorize the traffic flow information in both previous and subsequent sequence. To demonstrate the model's performance, a set of real case data at 1-hour intervals from 5 working days was used, wherein the dataset was separated into training and validation. To improve data quality, an augmented dickey-fuller unit root test and differential processing were performed before model training. Four benchmark models were used, including the Autoregressive Integrated Moving Average (ARIMA), Long Short-Term Memory (LSTM), Bidirectional Long Short-Term Memory (Bi-LSTM), and GRU. The prediction results show the superior performance of Bi-GRU. The Root Mean Square Error (RMSE), Mean Absolute Percentage Error (MAPE), and Mean Absolute Error (MAE) of the Bi-GRU model are 30.38, 9.88%, and 23.35, respectively. The prediction accuracy of LSTM, Bi-LSTM, GRU, and Bi-GRU, which belong to deep learning methods, is significantly higher than that of the traditional ARIMA model. The MAPE difference of Bi-GRU and GRU is 0.48% which is a small prediction error value. The results show that the prediction accuracy of the peak period is higher than that of the low peak. The Bi-GRU model has a certain lag on traffic flow prediction. [ABSTRACT FROM AUTHOR]

Abstract Background Compound–protein interaction prediction is necessary to investigate health regulatory functions and promotes drug discovery. Machine learning is becoming increasingly important in bioinformatics for applications such as analyzing protein-related data to achieve successful solutions. Modeling the properties and functions of proteins is important but challenging, especially when dealing with predictions of the sequence type. Result We propose a method to model compounds and proteins for compound–protein interaction prediction. A graph neural network is used to represent the compounds, and a convolutional layer extended with a bidirectional recurrent neural network framework, Long Short-Term Memory, and Gate Recurrent unit is used for protein sequence vectorization. The convolutional layer captures regulatory protein functions, while the recurrent layer captures long-term dependencies between protein functions, thus improving the accuracy of interaction prediction with compounds. A database of 7000 sets of annotated compound protein interaction, containing 1000 base length proteins is taken into consideration for the implementation. The results indicate that the proposed model performs effectively and can yield satisfactory accuracy regarding compound protein interaction prediction. Conclusion The performance of GCRNN is based on the classification accordiong to a binary class of interactions between proteins and compounds The architectural design of GCRNN model comes with the integration of the Bi-Recurrent layer on top of CNN to learn dependencies of motifs on protein sequences and improve the accuracy of the predictions.

Deep learning methods have shown early progress in analyzing complicated ECG signals, especially in heartbeat classification and arrhythmia detection. However, there is still a long way to go in terms of health-related data analysis. This research provides a duel structured and bidirectional Recurrent Neural Network(RNN) method for arrhythmia classification that addresses the issues with multilayered dilated convolution neural network (CNN) models. Initially, the data is preprocessed by Chebyshev Type II filtering that is faster and do not use statistical characteristics. Noise from the preprocesed filter is aslo removed by using Daubechies wavelet that can able to solve fractal problems and signal discontinuities. An then Z-normalization is done using Pan-Tompkins normalization technique for handling of different normally distributed samples. Finally, a generative adversarial network (GAN)-based synthetic signal is generated for recreation of signal to handle imbalanced signal class. The proposed Bidirectional RNN with Dilated CNN (BRDC) appears to take advantage of the potentiality of multilayered dilated CNN and bidirectional RNN unit (bidirectional gated recurrent Units, BiGRU - bidirectional long short-term memory, BiLSTM) architecture to generate fusion features. Finally, the signals are classified by fully connected layer and Rectified Linear Unit (ReLU) activation function. The PhysioNet 2017 challenge dataset is used to train and validate the proposed model. By combining fusion features with dilated CNN, the learned model significantly improves the classification performance and interpretability. The experimental findings show that, for MIT-BIH provided ECG (electrocardiogram) data to identify arrhythmia, the proposed BRDC model outperforms existing models with 99.90 % accuracy, 98.41 % F1, 97.96 % precision, and 99.90 % recall during training. One of the significant findings of this study is that the proposed approach can significantly reduce time length when employing RNN networks with multilayered dilated CNN. Overall, our hybrid model using BiGRU-BiLSTM and multi-layered dilated CNN provides a cost-effective ECG signal reduction and high-performance automated recognition technique to identify arrhythmia. Our future improvement will focus on the classification of numerous arrhythmia signal-based data, automatic and cloud based ECG classification.

Aiming at the problems of lack of potential implicit relation mining between entities and insufficient head entity information extraction in existing methods, a head entity-enhanced multi-hop attention implicit relations joint mining model Multi-Air (multi-hop attention implicit relations joint mining method) was proposed. The method first uses the BERT (bidirectional encoder representations from transformers) model to encode the features of the input sentence and predicts the position of the head entity through the Sigmoid function, and then uses the bidirectional gated recurrent unit (Bi-GRU) to enhance the feature of the head entity. After making full use of the high-level information of the head entities, the model can output the possible starting and ending positions of the tail entities with multiple relationships. Then the model continues to use the tail entities as the head entities of the next hop and iteratively perform the prediction of the multi-hop tail entities. At the same time, the model uses the attention weight to dynamically adjust the features of the learned entities and relationships so as to realize the mining of potential relationship triples in the plain text. The Multi-Air model has made good improvements in both the public dataset NYT and the civil aviation emergency dataset.

The relationships among different subregions in facial images and their varying contributions to facial expression recognition indicate that using a fixed subregion weighting scheme would result in a substantial loss of valuable information. To address this issue, we propose a facial expression recognition network called BGA-Net, which combines bidirectional gated recurrent units (BiGRUs) with an attention mechanism. Firstly, a convolutional neural network (CNN) is employed to extract feature maps from facial images. Then, a sliding window cropping strategy is applied to divide the feature maps into multiple subregions. The BiGRUs are utilized to capture the dependencies among these subregions. Finally, an attention mechanism is employed to adaptively focus on the most discriminative regions. When evaluated on CK+, FER2013, and JAFFE datasets, our proposed method achieves promising results.

Offensive language in social media affects the social experience of individuals and groups and hurts social harmony and moral values. Therefore, in recent years, the problem of offensive language detection has attracted the attention of many researchers. However, the primary research currently focuses on detecting English offensive language, while few studies on the Chinese language exist. In this paper, we propose an innovative approach to detect Chinese offensive language. First, unlike previous approaches, we utilized both RoBERTa’s sentence-level and word-level embedding, combining the sentence embedding and word embedding of RoBERTa’s model, bidirectional GRU, and multi-head self-attention mechanism. This feature fusion allows the model to consider sentence-level and word-level semantic information at the same time so as to capture the semantic information of Chinese text more comprehensively. Second, by concatenating the output results of multi-head attention with RoBERTa’s sentence embedding, we achieved an efficient fusion of local and global information and improved the representation ability of the model. The experiments showed that the proposed model achieved 82.931% accuracy and 82.842% F1-score in Chinese offensive language detection tasks, delivering high performance and broad application potential.

This work aims to provide a novel multimodal sarcasm detection model that includes four stages: pre-processing, feature extraction, feature level fusion, and classification. The pre-processing uses multimodal data that includes text, video, and audio. Here, text is pre-processed using tokenization and stemming, video is pre-processed during the face detection phase, and audio is pre-processed using the filtering technique. During the feature extraction stage, such text features as TF-IDF, improved bag of visual words, n-gram, and emojis as well on the video features using improved SLBT, and constraint local model (CLM) are extraction. Similarly the audio features like MFCC, chroma, spectral features, and jitter are extracted. Then, the extracted features are transferred to the feature level fusion stage, wherein an improved multilevel canonical correlation analysis (CCA) fusion technique is performed. The classification is performed using a hybrid classifier (HC), e.g. bidirectional gated recurrent unit (Bi-GRU) and LSTM. The outcomes of Bi-GRU and LSTM are averaged to obtain an effective output. To make the detection results more accurate, the weight of LSTM will be optimally tuned by the proposed opposition learning-based aquila optimization (OLAO) model. The MUStARD dataset is a multimodal video corpus used for automated sarcasm discovery studies. Finally, the effectiveness of the proposed approach is proved based on various metrics.

Wireless energy transfer technology (WET)-enabled mobile charging provides an innovative strategy for energy replenishment in wireless rechargeable sensor networks (WRSNs), where the mobile charger (MC) can charge the sensors sequentially by WET according to the mobile charging scheduling scheme. Although there have been fruitful studies, they usually assume that all sensors will be charged fully once scheduled or charged to a fixed percentage determined by a charging upper threshold, resulting in low charging performance as they cannot adjust the charging operation on each sensor adaptively according to the real-time charging demands. To tackle this challenge, we first formulate the mobile charging scheduling as a joint mobile charging sequence scheduling and charging upper threshold control problem (JSSTC), where the charging upper threshold of each sensor can adjust adaptively. Then, we propose a novel multi-discrete action space deep Q-network approach for JSSTC (MDDRL-JSSTC), where MC is regarded as an agent exploring the environment. The state information observed by MC at each time step is encoded to construct a high-dimensional vector. Furthermore, a two-dimensional action is mapped to the charging destination of MC and the corresponding charging upper threshold at the next time step, using bidirectional gated recurrent units (Bi-GRU). Finally, we conduct a series of experiments to verify the superior performance of the proposed approach in prolonging the lifetime compared with the state-of-the-art approaches.

This work aims to provide a novel multimodal sarcasm detection model that includes four stages: pre-processing, feature extraction, feature level fusion, and classification. The pre-processing uses multimodal data that includes text, video, and audio. Here, text is pre-processed using tokenization and stemming, video is pre-processed during the face detection phase, and audio is pre-processed using the filtering technique. During the feature extraction stage, such text features as TF-IDF, improved bag of visual words, n-gram, and emojis as well on the video features using improved SLBT, and constraint local model (CLM) are extraction. Similarly the audio features like MFCC, chroma, spectral features, and jitter are extracted. Then, the extracted features are transferred to the feature level fusion stage, wherein an improved multilevel canonical correlation analysis (CCA) fusion technique is performed. The classification is performed using a hybrid classifier (HC), e.g. bidirectional gated recurrent unit (Bi-GRU) and LSTM. The outcomes of Bi-GRU and LSTM are averaged to obtain an effective output. To make the detection results more accurate, the weight of LSTM will be optimally tuned by the proposed opposition learning-based aquila optimization (OLAO) model. The MUStARD dataset is a multimodal video corpus used for automated sarcasm discovery studies. Finally, the effectiveness of the proposed approach is proved based on various metrics. [ABSTRACT FROM AUTHOR]

The problem of low recognition accuracy of emotion recognition models is easily caused by interference such as data redundancy and irrelevant features. In this paper, we propose a speech emotion recognition (SER) method based on an attentional convolutional neural network (CNN) bidirectional gated recurrent unit (Bi-GRU) fusing visual information. First, we pretrained the log-mel spectrograms in a ResNet-based attentional convolutional neural network (RACNN) to extract speech features. Second, the CNN-extracted facial static appearance features are fused with speech features using a deep Bi-GRU to obtain speech appearance features. A series of gated recurrent units with attention mechanisms (AGRUs) are used to extract facial geometric features. Then, the hybrid features are obtained by further combining the integrated speech appearance features with facial geometric features, and kernel linear discriminant analysis (KLDA) is used to discriminate them. Finally, the proposed method in this paper obtained accuracies of 87.92% and 89.65% on the RAVDESS and eNTERFACE'05 emotion databases, respectively. The experimental results demonstrate that the method in this paper effectively improved the accuracy and robustness of SER. [ABSTRACT FROM AUTHOR]

Sepsis is a common and deadly condition that must be treated eloquently within 19 hours. Numerous deep learning techniques, including Recurrent Neural Networks, Convolution Neural Networks, Long Short-Term Memory, and Gated Recurrent Units, have been suggested for diagnosing long-term sepsis. Regardless, a sizable portion of them are computationally risky and have precision problems. The primary issue described is that output will degrade, and resource utilization will expand proportionately as the volume of dependencies grows. To overcome these issues, we propose a G-Sep technique utilizing Bidirectional Gated Recurrent Unit Algorithm, which consumes much less resource to detect the disease and in a short time with better accuracy than the existing methods to diagnose the sepsis. AI models could assist with distinguishing potential clinical factors and give better than existing conventional low-execution models. The proposed model is implemented utilizing Conda and Tensorflow Framework using the California Inpatient Severe Sepsis (CISS) Patient Dataset. The comparative simulation of the various existing models and the proposed G-Sep model is done using Conda and Tensor frameworks. The simulation results revealed that the proposed model had outperformed other frameworks in terms of mean average precision (mAP), receiver operating characteristic curve (ROC), and Area under the ROC Curve (AUROC) metrics linearly. [ABSTRACT FROM AUTHOR]

A question answering (QA) system is a research direction in the field of artificial intelligence and natural language processing (NLP) that has attracted much attention and has broad development prospects. As one of the main components in the QA system, the accuracy of question classification plays a key role in the entire QA task. Therefore, not only the traditional machine learning methods but also today's deep learning methods are widely used and deeply studied in question classification tasks. This paper mainly introduces our work on two aspects of Chinese question classification. The first is to use an answer-driven method to build a richer Chinese question classification dataset for the small-scale problems of the existing experimental dataset, which has a certain reference value for the expansion of the dataset, especially for the construction of those low-resource language datasets. The second is to propose a deep learning model of problem classification with a Transformer + Bi-GRU + Attention structure. Transformer has strong learning and coding ability, but it adopts the scheme of fixed coding length, which divides the long text into multiple segments, and each segment is coded separately; there is no interaction that occurs between segments. Here, we achieve the information interaction between segments through Bi-GRU so as to improve the coding effect of long sentences. Our purpose of adding the Attention mechanism is to highlight the key semantics in questions that contain answers. The experimental results show that the model proposed in this paper has significantly improved the accuracy of question classification. [ABSTRACT FROM AUTHOR]

The rapid development of artificial intelligence (AI) technology has enabled large-scale AI applications to land in the market and practice. However, plenty of security issues have been exposed to society while AI technology has brought many conveniences to humankind, especially for the chatbot with online learning. This paper proposes a speech censorship chatbot system with reinforcement learning, which is mainly composed of two parts: the aggressive speech censorship model and the speech purification model. The aggressive speech censorship can combine the context of user input sentences to detect aggressive speech and respond to the rapid evolution of aggressive speech. According to the situation of the chatbot that is polluted by large numbers of aggressive speech, the speech purification model has the capacity to "forget" the learned malicious data through reinforcement learning rather than rolling back to the early versions. In addition, by integrating few-shot learning, the speed of speech purification is accelerated while reducing the influence on the quality of replies. The experimental results show that our proposed method reduces the probability of generating aggressive speeches and that the integration of the few-shot learning improves the training speed rapidly while effectively slowing down the decline in BLEU values. [ABSTRACT FROM AUTHOR]

Background: Compound–protein interaction prediction is necessary to investigate health regulatory functions and promotes drug discovery. Machine learning is becoming increasingly important in bioinformatics for applications such as analyzing protein-related data to achieve successful solutions. Modeling the properties and functions of proteins is important but challenging, especially when dealing with predictions of the sequence type. Result: We propose a method to model compounds and proteins for compound–protein interaction prediction. A graph neural network is used to represent the compounds, and a convolutional layer extended with a bidirectional recurrent neural network framework, Long Short-Term Memory, and Gate Recurrent unit is used for protein sequence vectorization. The convolutional layer captures regulatory protein functions, while the recurrent layer captures long-term dependencies between protein functions, thus improving the accuracy of interaction prediction with compounds. A database of 7000 sets of annotated compound protein interaction, containing 1000 base length proteins is taken into consideration for the implementation. The results indicate that the proposed model performs effectively and can yield satisfactory accuracy regarding compound protein interaction prediction. Conclusion: The performance of GCRNN is based on the classification accordiong to a binary class of interactions between proteins and compounds The architectural design of GCRNN model comes with the integration of the Bi-Recurrent layer on top of CNN to learn dependencies of motifs on protein sequences and improve the accuracy of the predictions. [ABSTRACT FROM AUTHOR]

In chest and abdomen robotic radiosurgery, due to the motion delay of the robotic manipulator, the tumor position tracking process has a period of delay. This delay ultimately affects the accuracy of radiosurgery treatment. To address the influence of the delay in robotic radiosurgery, a Long-and-Short-Term Memory (LSTM) network as a deep Recurrent Neural Network (RNN) has been applied in a prediction network model for respiratory motion tracking in recent years. However, patients' respiratory state may change in the process of treatment, which may influence the accuracy of prediction. Therefore, it is necessary to update the prediction network through additional data, such as the actual position of the tumor obtained by X-ray imaging. However, the LSTM network has a long update time, and it may not be able to complete the prediction model update in a cycle of X-ray acquisition. To solve this problem, a fast prediction model based on Bidirectional Gated Recurrent Unit (Bi-GRU), is proposed in this paper. This method can reduce the average updating time of the network model by 30%.

Event extraction is an important research direction in the field of natural language processing (NLP) applications including information retrieval (IR). Traditional event extraction is realized with two methods: the pipeline and the joint extraction methods. The pipeline method determines the event by triggering word recognition to further implement event extraction and is prone to error cascading. The joint extraction method applies deep learning to implement the completion of the trigger word and the argument role classification task. Most studies with the joint extraction method adopt the CNN or RNN network structure. However, in the case of event extraction, deeper understanding of complex contexts is required. Existing studies do not make full use of syntactic relations. This paper proposes a novel event extraction model, which is built upon a Tree-LSTM network and a Bi-GRU network and carries syntactically related information. It is illustrated that this method simultaneously uses Tree-LSTM and Bi-GRU to obtain a representation of the candidate event sentence and identify the event type, which results in a better performance compared to the ones that use chain structured LSTM, CNN or only Tree-LSTM. Finally, the hidden state of each node is used in Tree-LSTM to predict a label for candidate arguments and identify/classify all arguments of an event. Lab results show that the proposed event extraction model achieves competitive results compared to previous works.

Fault location is one of the most essential techniques to maintain the stable operation of power systems. A fast and accurate fault location allows operators to restore power grids faster and avoid economic losses. Conventional methods rely on expert knowledge to extract the necessary features (e.g. DWT, DFT). For large systems, more coupling effects of transmission lines require more complex feature engineering, and incomplete features can easily introduce large errors. To overcome this, a deep learning approach without manual feature extraction is introduced to the fault location model under big data application. Towards this end, in the proposed method, the attention mechanism, the Bi-GRU and a dual structure network are applied to analyze the current data from different perspectives. Complete information for the fault features is extracted for the fault location. Based on a time series model and benefit from the ability to internally acquire the information architecture of faulty line, the established model is adaptive to the power grids with very complex topologies. Simulation results indicate that the proposed double-structure model reduces the maximum error and is less affected by noise. In comparison with different structures and different models, the proposed method shows better performance in IEEE 39-bus system.

Research of document classification is ongoing to employ the attention based-deep learning algorithms and achieves impressive results. Owing to the complexity of the document, classical models, as well as single attention mechanism, fail to meet the demand of high-accuracy classification. This paper proposes a method that classifies the document via the hierarchical multi-attention networks, which describes the document from the word-sentence level and the sentence-document level. Further, different attention strategies are performed on different levels, which enables accurate assigning of the attention weight. Specifically, the soft attention mechanism is applied to the word-sentence level while the CNN-attention to the sentence-document level. Due to the distinctiveness of the model, the proposed method delivers the highest accuracy compared to other state-of-the-art methods. In addition, the attention weight visualization outcomes present the effectiveness of attention mechanism in distinguishing the importance. [ABSTRACT FROM AUTHOR]

Sentiment classification is a fundamental task in NLP, and its aim to predict the sentiment polarities of the given texts. Recent researches show great interest in modeling Chinese sentiment classification. However, the complexity of Chinese grammar makes the performance of the existing Chinese sentiment classification model not perform well. In order to address the above problem, we propose a sentiment classification method based on grammar rules and bidirectional encoder representation from transformers (BERT). We first preprocess data through BERT model. Then we combine the Chinese grammar rules with Bi-gated recurrent neural network (GRU) in the form of constraints, and simulate the linguistic functions at the sentence by standardizing the output of adjacent positions. Extensive experiments on two public datasets demonstrate the effectiveness of our proposed method, and our findings in the experiment provide new insights for the future development of Chinese sentiment classification.