Search

In this contribution, we employ a recurrent neural network (RNN) architecture in a variational optimization to obtain the ground state of linear chains of planar, dipolar rotors. We test different local basis sets and discuss their impact on the sign structure of the many-body ground state wavefunction. It is demonstrated that the RNN ansatz we employ is able to treat systems with and without a sign problem in the ground state. For larger chains with up to 50 rotors, accurate properties, such as correlation functions and Binder parameters, are calculated. By employing quantum annealing, we show that precise entanglement properties can be obtained. All these properties allow one to identify a quantum phase transition between a paraelectric and a ferroelectric quantum phase. [ABSTRACT FROM AUTHOR]

This literature review explores the role of attractor neural networks (ANNs) in modeling psychological processes in artificial and biological systems. By synthesizing research from dynamical systems theory, psychology, and computational neuroscience, the review provides an overview of the current understanding of ANN function in memory formation, memory reinforcement, retrieval, and forgetting. Key mathematical foundations of ANNs, including dynamical systems theory and energy functions, are discussed to explain the behavior and stability of these networks. The review also examines empirical applications of ANNs in cognitive processes such as semantic memory and episodic recall, as well as highlighting the hippocampus' role in pattern separation and completion. The review addresses challenges like catastrophic forgetting and noise effects on memory retrieval. By identifying gaps between theoretical models and empirical findings, it highlights the interdisciplinary nature of ANN research and suggests future areas for exploration.

Throughput analysis plays an important role in the operations and management of automotive manufacturing. Predicting how the system throughput changes over time helps the plant managers to make timely operational decisions to meet the daily production requirement. In today's automotive production systems, the availability of sensing data reflecting process variables and machine status across the plant floor raises new opportunities for improving the throughput prediction accuracy. However, challenges exist in extracting the most important features from the high-dimensional data and capturing the complicated time-vary interdependency among different assets in the system. To overcome such challenges, in this paper we propose a hierarchical Recurrent Neural Network (RNN)-based framework that is composed of clustering, dimension reduction and feature selection, regression, and prediction pruning/adjustment. In addition to predicting end-of-line throughput, our framework then identifies the associations for low throughput to facilitate downtime prevention and maintenance decision-making. The proposed framework is applied to an automotive production system, and its effectiveness is demonstrated by comparison with conventional methods. [ABSTRACT FROM AUTHOR]

Image captioning, which combines natural language processing and computer vision, has developed rapidly in recent years. It tends to be applied in data retrieval, blind navigation, intelligent transportation, smart home, medical assistance, news media and other domains. In order to elevate the consistency and abundance of image captioning languages and express people's subjective emotions effectively, a Generative Adversarial Network (GAN) is applied in this paper to obtain multi-stylized image emotion captions and generate two captions containing positive and negative emotions, respectively. Among them, Residual Network (ResNet) and Gate Recurrent Unit (GRU) are integrated into the generator, while the capsule neural network is applied to the discriminator. We conduct experiments on the popular MSCOCO and Senticap datasets to validate the model and demonstrate its satisfied performance in comparison to current advanced image captioning approaches. [ABSTRACT FROM AUTHOR]

The ability to accurately predict the stock market is a crucial financial topic. The basic assumption is that future stock returns can be somewhat predicted based on publicly available data from the past. Economic variables like interest rates and currency rates, information related to political influences, and company-specific data like statements of income and dividends are some examples of this information. Yet, attempting to forecast stock returns goes against the idea that markets are generally efficient. A prediction model has been created that leverages machine learning, big data analytics, and social media analytics to predict stock market trends regularly. The main objective of this work is to propose a novel approach for stock market prediction for handling big data. This research develops a new SP-PDA (Stock Price Prediction with Political Data Analysis) which includes preprocessing, feature extraction, and prediction. In this work, we employed three types of data (Stock data, News data and Political data). For the stock data, preprocessing was done using Z-score normalization, while tokenization and lemmatization were used for the news and political data. Next, to handle the big data, we used a Map-Reduce architecture. Here, feature extraction from preprocessed stock, news, and political data occurs in the mapper function. The reduction face yields the final extracted feature set. Commodity channel index (CCI), Chaikin Volatility (CV), and Donchian Channel (DC) are examples of technical indicator-based features that are extracted for stock data during the feature extraction phase. Mutual Information (MI), Improved Pointwise mutual information (IPMI), Term Frequency-Inverse Document Frequency (TF-IDF), and correlation features are retrieved for news and political data. Stock market prediction is based on the features extracted. To predict stock prices, we employed an ensemble classification model that incorporates classifiers such as Convolutional Neural Network (CNN), Recurrent Neural Network (RNN), Bidirectional Gated Recurrent Unit (Bi-GRU), Long-short term memory (LSTM), and Deep Maxout. An improved score level fusion is carried out to define the final prediction outcome from the obtained intermediate prediction results. The suggested Dwarf Mongoose Updated War Strategy-based Generalized Normal Distribution (DMUWS-based GND) Optimization Algorithm is used to optimizing the weights of ensemble classifiers. Finally, the performance of the proposed model is evaluated over existing models in terms of error measures like Mean Absolute Error (MAE), Mean Squared Error (MSE), Mean Absolute Percentage Error (MAPE) and Mean Squared Log Error (MSLE). The SP-PDA achieved the desired MAE of 0.012, which is low when compared to the other traditional systems for accurate prediction of the stock market. [ABSTRACT FROM AUTHOR]

Image captioning generates textual description from the corresponding input image with the help of computer vision and natural language processing. In recent years, deep learning approaches have shown promise in image captioning. This research introduces a novel image captioning architecture comprising a dual self-attention fused encoder-decoder framework. The VGG16 Hybrid Places 1365 (V16HP1365) encoder captures diverse visual features from images, enhancing the quality of image representations. In this article, the Gated Recurrent Unit (GRU) is considered as a decoder for conducting word-level language modeling. Additionally, the dual self-attention network embedded in the architecture allows for capturing contextual image information to provide accurate content descriptions and relationship understanding. Experimental evaluations on the COCO dataset showcase superior performance, surpassing existing methods in terms of captioning quality metrics. This approach holds potential for applications such as aiding the visually impaired and advancing content retrieval. Future work aims to extend the model to support multilingual captioning. [ABSTRACT FROM AUTHOR]

Nowadays, unmanned aerial vehicles (UAVs) becomes more prominent because of their benefits namely versatile, inexpensive, scalable, and autonomous. The aerial image classification procedure has involved further attention to the UAV-allowed surveillance model facilitating objective detection, and tracking procedure and significantly enhancing the visual surveillance outcomes. Earlier methods to resolve the scene classification method dependent upon feature coding systems with unsupervised feature learning or low-level hand-engineered features can make mid-level image features with limited representational capabilities that usually limit them from the achievement of effectual outcomes. The development of convolutional neural networks (CNNs) helps to attain an efficient classification process in the area of image classification. This paper introduces a Pelican Optimization Algorithm with Convolutional-Recurrent Hop-Field Neural Network (POA-CRHFNN) model for the UAV image classification model. The presented POA-CRHFNN method initially involves preprocessing in two stages namely Gaussian Filter (GF) based noise elimination and contrast enhancement. In addition, the POA-CRHFNN method exploits the ShuffleNetv2 CNN model for feature vector generation. For aerial image classification, the CRHFNN model is used, which allocates different types of classes to the aerial images. The combination of Convolutional Neural Networks (CNNs) and Recurrent Neural Networks (RNNs) in the CRHFNN can capture both spatial and temporal dependencies in aerial images. The integration of the Hopfield Neural Network (HNN) can recognize and classify aerial images based on learned patterns and associations. Finally, the POA is applied for fine-tuning the hyperparameters of the CRHFNN model. By leveraging the behavioural patterns of pelicans, the POA may explore the search space effectively and potentially find better hyperparameter configurations for improved classification performance. The effectiveness of the POA-SCNN model can be investigated on a benchmark aerial image database and the simulation outcomes reported the higher outcome of the POA- CRHFNN method over existing techniques with maximum accuracy of 99.77%, precision of 98.01%, recall of 97.52%, and F-score of 97.69%. [ABSTRACT FROM AUTHOR]

ABSTRACT In this article, a neuroadaptive sliding mode control (NSMC) strategy based on recurrent neural network (RNN) for robustly and adaptively tracking the desired position and attitude of an uncertain tilting quadrotor unmanned aerial vehicle (TQUAV) with unknown controllers is presented. The main contribution of this article is the real‐time adjustment of unknown flight controllers using the approximation characteristics of RNN, in which the derived approximation errors of RNN are sufficiently estimated by adaptive control method that can reduce or eliminate the impact of error terms on the evolution of closed‐loop systems. Especially, Lyapunov stability analysis is greatly simplified compared to existing methods and does not require amplification or reduction. Finally, the superior performance of the NSMC strategy was verified by comparing simulation results. [ABSTRACT FROM AUTHOR]

ABSTRACT To fulfil the commitments of the Paris Agreement, China will strive to achieve carbon peak (CP) by 2030. It is necessary to identify the evolution characteristics of China's carbon emissions and provide a scientific path prediction for the formulation of reasonable emission reduction policies and measures. This study summarises and predicts the pathway of China's carbon peak (CP) using carbon emission intensity (CEI) and the percentage of non‐fossil energy consumption (NEC) as indicators, and combining MSIH(3)‐AR(2) model and recurrent neural network. The results show that: (1) China's CEI experiences a ‘low decline regime’ (LDR), a ‘medium decline regime’ (MDR) and a ‘high decline regime’ (HDR), while the share of NEC goes through a ‘low fluctuation regime’ (LFR), a ‘medium growth regime’ (MGR) and a ‘high growth regime’ (HGR). (2) For CEI, the switching probability from MDR to the HDR is 74.88%, illustrating a substantial improvement. For NEC, the switching probability from MGR to HGR is 28.92%, but the probability of returning to MGR is 61.76%, indicating an adjustment. (3) By 2030, CEI will reach 0.9896 tons/100 million CNY, decreased by 66.35% compared with 2005. While the percentage of NEC will rise to 26.61%. Based on these, policy suggestions such as strengthening the top‐level design, upgrading energy mix and accelerating green technological changes are proposed to break the bottlenecks of reaching CP and further zero carbon goal. This study is expected to provide theoretical support and empirical evidence for the achievement CP in China, and to provide empirical references for promoting the ‘dual carbon’ process in other countries. [ABSTRACT FROM AUTHOR]

The integration of IoT and cloud services enhances communication and quality of life, while predictive analytics powered by AI and deep learning enables proactive healthcare. Deep learning, a subset of machine learning, efficiently analyzes vast datasets, offering rapid disease prediction. Leveraging recurrent neural networks on electronic health records improves accuracy for timely intervention and preventative care. In this manuscript, Internet of Things and Cloud Computing-based Disease Diagnosis using Optimized Improved Generative Adversarial Network in Smart Healthcare System (IOT-CC-DD-OICAN-SHS) is proposed. Initially, an Internet of Things (IoT) device collects diabetes, chronic kidney disease, and heart disease data from patients via wearable devices and intelligent sensors and then saves the patient’s large data in the cloud. These cloud data are pre-processed to turn them into a suitable format. The pre-processed dataset is sent into the Improved Generative Adversarial Network (IGAN), which reliably classifies the data as disease-free or diseased. Then, IGAN was optimized using the Flamingo Search optimization algorithm (FSOA). The proposed technique is implemented in Java using Cloud Sim and examined utilizing several performance metrics. The proposed method attains greater accuracy and specificity with lower execution time compared to existing methodologies, IoT-C-SHMS-HDP-DL, PPEDL‐MDTC and CSO-CLSTM-DD-SHS respectively. [ABSTRACT FROM AUTHOR]

Lithology classification is a fundamental aspect of reservoir classification. Due to the limited availability of core samples, computational modelling methods for lithology classification based on indirect measurements are required. The main challenge for standard clustering methods is the complex vertical dependency of sedimentological sequences as well as the spatial coupling of well logs. Machine learning methods, such as recurrent neural networks, long short‐term memory and bidirectional long short‐term memory, can account for the spatial correlation of the measured data and the predicted model. Based on these developments, we propose a novel approach using two distinct models: a self‐attention‐assisted bidirectional long short‐term memory model and a multi‐head attention‐based bidirectional long short‐term memory model. These models consider spatial continuity and adaptively adjust the weight in each step to improve the classification using the attention mechanism. The proposed method is tested on a set of real well logs with limited training data obtained from core samples. The prediction results from the proposed models and the benchmark one are compared in terms of the accuracy of lithology classification. Additionally, the weight matrices from both attention mechanisms are visualized to elucidate the correlations between depth steps and to help analyse how these mechanisms contribute to improved prediction accuracy. The study shows that the proposed multi‐head attention‐based bidirectional long short‐term memory model improves classification, especially for thin layers. [ABSTRACT FROM AUTHOR]

This paper explores the application of machine learning techniques, specifically deep learning, to the inverse problem of marine controlled‐source electromagnetic data. A novel approach is proposed that combines the convolutional neural network and recurrent neural network architectures to reconstruct layered electrical resistivity variation beneath the seafloor from marine controlled‐source electromagnetic data. The approach leverages the strengths of both convolutional neural network and recurrent neural network, where convolutional neural network is used for recognizing and classifying features in the data, and recurrent neural network is used to capture the contextual information in the sequential data. We have built a large synthetic dataset based on one‐dimensional forward modelling of a large number of resistivity models with different levels of electromagnetic structural complexity. The combined learning of convolutional neural network and recurrent neural network is used to construct the mapping relationship between the marine controlled‐source electromagnetic data and the resistivity model. The trained network is then used to predict the distribution of resistivity in the model by feeding it with marine controlled‐source electromagnetic responses. The accuracy of the proposed approach is examined using several synthetic scenarios and applied to a field dataset. We explore the sensitivity of deep learning inversion to different electromagnetic responses produced by resistive targets distributed at different depths and with varying levels of noise. Results from both numerical simulations and field data processing consistently demonstrate that deep learning inversions reliably reconstruct the subsurface resistivity structures. Moreover, the proposed method significantly improves the efficiency of electromagnetic inversion and offers significant performance advantages over traditional electromagnetic inversion methods. [ABSTRACT FROM AUTHOR]

Conversational Artificial Intelligence (AI) has emerged as a promising technology in the healthcare domain, facilitating interactive and personalized interactions between patients, healthcare professionals, and virtual assistants. This abstract presents an overview of the development process for Conversational AI in healthcare, focusing on utilizing Recurrent Neural Networks (RNNs). RNNs are well-suited for sequence modeling tasks and enable context-aware responses. Conversations can be complex, and emotions expressed within them may not always be clear-cut. It can be challenging for sentiment analysis models to interpret accurately. To overcome these issues, create a novel technique called Generative Pretrained based Recurrent Neural Network (GPbRNN). The developed model is to increase the efficiency of the model and also improve the emotional predictions. Conversational AI in healthcare, empowered by RNNs, can revolutionize the field by providing personalized and accessible information to patients, supporting healthcare professionals in decision-making, and enhancing overall healthcare delivery. Further research and development in this area promise to improve patient outcomes and transform healthcare. [ABSTRACT FROM AUTHOR]

Motorcycle crashes remain a significant public safety concern, requiring diverse analytical approaches to inform countermeasures. This study uses machine learning to analyze injury severity in crashes in Thailand from 2018 to 2020. Traditional and advanced models, including including random forest (RF), support vector machine (SVM), deep neural network (DNN), recurrent neural network (RNN), long short-term memory (LSTM), and eXtreme gradient boosting (XGBoost), were compared. Hyperparameter tuning via GridSearchCV optimized performance. XGBoost, with a tradeoff score of 105.65%, outperformed other models in predicting severe and fatal injuries. SHapley Additive exPlanations (SHAPs) identified significant risk factors including speeding, drunk driving, two-lane roads, unlit conditions, head-on and truck collisions, and nighttime crashes. Conversely, factors such as barrier medians, flashing traffic signals, sideswipes, rear-end crashes, and wet roads were associated with reduced severity. These findings suggest opportunities for integrated infrastructure improvements and expanded rider training and education programs to address behavioral risks. [ABSTRACT FROM AUTHOR]

As the extraction of mineral resources intensifies, ground subsidence in mining areas has escalated, posing substantial challenges to sustainable development and operational safety. This subsidence, resulting directly from mining activities, significantly compromises the safety of nearby residents by damaging residential structures and infrastructure. Thus, developing precise and dependable methods for predicting ground subsidence is crucial. This study introduces an innovative Cabs-Unet model, which enhances the U-Net architecture by integrating a Convolutional Block Attention Module (CBAM) and Depthwise Separable Convolutions (DSC). This model aims to predict the spatiotemporal dynamics of the Interferometric Synthetic Aperture Radar (InSAR) time series. Employing Small Baseline Subset Interferometric Synthetic Aperture Radar (SBAS InSAR) technology, we gathered and validated data on ground subsidence at the Pengzhuang coal mine from May 2017 to November 2021, covering 130 scenes, with its accuracy corroborated by levelling survey results. An empirical evaluation of the Cabs-Unet model in two distinct subsidence zones demonstrated superior performance over conventional methods like Convolutional Long Short-Term Memory (ConvLSTM) and Predictive Recurrent Neural Network (PredRNN), with Root Mean Square Error (RMSE) values of 1.44 and 1.70, respectively. These findings highlight the model’s efficacy in accurately predicting spatiotemporal InSAR ground subsidence. Further predictive analysis using InSAR data indicated an expected increase in subsidence, projecting cumulative declines of −457 mm in Area A and −1278 mm in Area B by 17 July 2022. Our model proves effective in assessing subsidence, promptly detecting potential risks and facilitating the rapid implementation of risk mitigation strategies. [ABSTRACT FROM AUTHOR]

Background: Epilepsy is a prevalent neurological disorder in which seizures cause recurrent episodes of unconsciousness or muscle convulsions, seriously affecting the patient's work, quality of life, and health and safety. Timely prediction of seizures is critical for patients to take appropriate therapeutic measures. Accurate prediction of seizures remains a challenge due to the complex and variable nature of EEG signals. The study proposes an epileptic seizure model based on a multidimensional Transformer with recurrent neural network(LSTM-GRU) fusion for seizure classification of EEG signals. Methodology: Firstly, a short-time Fourier transform was employed in the extraction of time-frequency features from EEG signals. Second, the extracted time-frequency features are learned using the Multidimensional Transformer model. Then, LSTM and GRU are then used for further learning of the time and frequency characteristics of the EEG signals. Next, the output features of LSTM and GRU are spliced and categorized using the gating mechanism. Subsequently, seizure prediction is conducted. Results: The model was tested on two datasets: the Bonn EEG dataset and the CHB-MIT dataset. On the CHB-MIT dataset, the average sensitivity and average specificity of the model were 98.24% and 97.27%, respectively. On the Bonn dataset, the model obtained about 99% and about 98% accuracy on the binary classification task and the tertiary upper classification task, respectively. Conclusion: The findings of the experimental investigation demonstrate that our model is capable of exploiting the temporal and frequency characteristics present within EEG signals. [ABSTRACT FROM AUTHOR]

To address the challenge of constructing spatial and temporal dependencies in existing traffic flow prediction methods, this paper proposed a new method called spatial temporal position attention graph neural network (ST-PAGNN), which utilized spatiotemporal location attention. Firstly, the graph neural network contained a location attention mechanism, which could better capture the spatial dependence of traffic nodes in the urban road network. Then, it used a gated recurrent neural network with trend adaptive transformer (Trendformer) to capture the local and global information of the traffic flow sequence in the time dimension. Finally, it used the improved grid search optimization method to optimize the introduced para-meters of the model, obtaining the global optimal solution with high time efficiency. The experimental results show that in the dataset PEMS-BAY, the evaluation indexes RMSE, MAE and MAPE of the ST-PAGNN method are 1.37, 2.57, 2.67%, 1.55, 3.64, 3.37%, 1.97, 4. 37 and 4.43%, respectively, when the prediction step size is 15 min, 30 min and 60 min, respectively. In the dataset METR-LA, when the prediction step size is 15 min, 30 min and 60 min, the evaluation indexes RMSE, MAE and MAPE of the ST-PAGNN method are 2. 73, 5. 16, 7.13%, 2.99, 5.97, 7.86%, 3.53, 7. 16 and 9.96%, respectively. The results show that the proposed ST-PAGNN method is higher than the existing models in the evaluation indexes under different granularities, which illustrates the effectiveness and superiority of ST-PAGNN in solving traffic prediction problems. [ABSTRACT FROM AUTHOR]

Radiation parameterization schemes are crucial components of weather and climate models, but they are also known to be computationally intensive. In recent decades, researchers have attempted to emulate these schemes using neural networks, with more attention to convolutional neural networks. However, in this paper, we explore the potential of recurrent neural networks (RNNs) for predicting solar heating rates. Our architecture was trained and tested using long-term hindcast data from the Pechora Sea region, with the conventional RRTMG scheme serving as a shortwave parameterization. Our findings show that the RNN offers rapid learning, fast inference, and excellent data fitting. We also present preliminary results demonstrating the use of RNNs for operational weather forecasting, which achieved a significant speedup in parameterization and reduced the overall forecast time by 40 %. [ABSTRACT FROM AUTHOR]

AbstractObjectiveDesignStudy sampleResultsConclusionsTo compare the performance of a selection of machine learning algorithms, trained to label peaks I, III, and V of the auditory brainstem response (ABR) waveform. An additional algorithm was trained to provide a confidence measure related to the ABR wave latency estimates.Secondary data analysis of a previously published ABR dataset. Five types of machine learning algorithm were compared within a nested k-fold cross-validation procedure.A set of 482 suprathreshold ABR waveforms were used. These were recorded from 81 participants with audiometric thresholds within normal limits.A convolutional recurrent neural network (CRNN) outperformed the other algorithms evaluated. The algorithm labelled 95.9% of ABR waves within ±0.1 ms of the target. The mean absolute error was 0.025 ms, averaged across the outer validation folds of the nested cross-validation procedure. High confidence levels were generally associated with greater wave-labelling accuracy.Machine learning algorithms have the potential to assist clinicians with ABR interpretation. The present work identifies a promising machine learning approach, but any algorithm to be used in clinical practice would need to be trained on a large, accurately labelled, heterogeneous dataset and evaluated in clinical settings in follow-on work. [ABSTRACT FROM AUTHOR]

Accurate machine learning streamflow prediction often requires coupling data-driven models with preprocessing techniques. This study aims to improve the performance of deep learning (DL) models, including long short-term memory, recurrent neural network (RNN), and gated recurrent unit (GRU) by incorporating maximal overlap discrete wavelet entropy transform (MODWET) techniques for streamflow forecasting. The merit of MODWET over maximal overlap discrete wavelet transform (MODWT) is that MODWET utilizes Entropy to determine the optimal decomposition level and suitable wavelet function, which was an unaddressed problem in wavelet-based decomposition models. Suitable decomposition level prevents providing unnecessary information or missing essential information. In this study we show that a unique decomposition level and wavelet filter is not suitable for any dataset. The research focuses on monthly streamflow data from three case studies in the CAMEL dataset in the USA. The accuracy of the models is evaluated using statistical measures such as Nash–Sutcliffe efficiency (NSE), root-mean-squared error, percent bias, and correlation coefficient (r). To determine the optimal model, a Taylor diagram is utilized. The results demonstrate the effectiveness of coupling MODWET with DL models in flood forecasting. Furthermore, genetic programming (GP) and partial correlation index (PCI) are employed for predictor selection. Hybrid models, namely MODWET-GP-GRU (NSE of 0.83), MODWET-GP-RNN (NSE of 0.95), and MODWET-PCI-GRU (NSE of 0.95), outperform simple DL models in terms of NSE and Taylor diagram evaluation. This study emphasizes the potential of hybrid models that combine DL algorithms with the recently proposed MODWET technique for streamflow prediction. [ABSTRACT FROM AUTHOR]

Railway bridges are essential components of any transportation system and are typically subjected to several environmental and operational actions that can cause damage. Furthermore, they are not easily replaced, and their failure can have catastrophic consequences. Considering the expected lifespan of bridges, it is essential to guarantee their adequate serviceability and safety. In this scenario, emerges the Structural Health Monitoring (SHM), which allows the early identification of damage before it becomes critical. Damage identification is usually performed by the comparison between the damaged and undamaged responses obtained from monitoring data. Among the several features extracted from the responses, the time-series models exhibit a better performance, capability of early damage detection, and may also be applied within online damage detection strategies using unsupervised machine learning frameworks. In this paper, a review of advanced time-series methodologies for damage detection is presented. Initially, several time-series models often used in SHM are described, such as Autoregressive Models (AR), Recurrent Neural Networks (RNN), Gated Recurrent Unit (GRU), and Long Short-Term Memory (LSTM). Later, the framework where these models are usually applied is also detailed, including the latest upgrades and most relevant results. Finally, the conclusions summarize and elucidate the current perspectives and research gaps on the time-series models. [ABSTRACT FROM AUTHOR]

This paper focuses on the computation issue of portfolio optimisation with scenario-based mean-Average Value at Risk (AVaR) in credibilistic environment. The portfolio optimisation problem is designed in two cases: risk taker model and risk-averse model. The main idea is to replace the portfolio selection models with linear programming (LP) problems. Since the computing time required for solving LP greatly depends on the dimension and the structure of the problem, the conventional numerical methods are usually less effective in real-time applications. One promising approach to handle online applications is to employ recurrent neural networks based on circuit implementation. Hence, according to the convex optimisation theory and some concepts of ordinary differential equations, a neural network model for solving the LP problems related to portfolio selection problems is presented. The equilibrium point of the proposed model is proved to be equivalent to the optimal solution of the original problem. It is also shown that the proposed neural network model is stable in the sense of Lyapunov and it is globally convergent to an exact optimal solution of the portfolio selection problem with fuzzy returns. Some illustrative examples are provided to show the feasibility and the efficiency of the proposed method in this paper. [ABSTRACT FROM AUTHOR]

Bone illnesses arise at a young age itself, so bone age assessment (BAA) is primarily utilized in paediatrics to identify their growth. Several BAA-related methods are employed to determine bone maturity; however, they do not give accuracy and the rate of error increases. To overcome this issue, in this manuscript, TW3-based Region of Interest (ROI) identification and classification with Chaotic ANN (CANN), as well as Deep Neural Network optimized with Evolved Gradient Direction optimizer (EVGO) architecture, is proposed for an Automated BAA classification. In this, Hand X-Ray images are taken from the Digital Hand Atlas (DHA) dataset and the Radiological Society of North America (RSNA) database. Here, Tanner-Whitehouse (TW3) and Shallow convolution neural network (SCNN) is utilized for extracting the ROI regions such as radius, ulna, and short bones (RUS) from the wrist region of the left hand. After that, the extracted ROI feature data are given to CANN for predicting BAA. Then, the predicted data are classified using the DNN-EVGO network. The bone age evaluation is then classified depending on maturity phases. The proposed approach is activated in MATLAB. Simulation outcomes of TW3- SCNN- CANN- DNN-EVGO- BAA methods for DHA database attain higher accuracy of 45.75%, 37.64%, 24.64% when analysed to the existing models, such as Faster Region-Convolutional Neural network oriented feature learning with optimal trained Recurrent Neural Network for BAA for paediatrics (TW3-F-R-CNN-AF-SFO), Multi-objective segmentation approach for BAA under parameter tuning- U-net architecture, and TW3-based fully automated bone age assessment system using deep neural networks (TW3-F-R-CNN-VGGNet-CNN-BAA). [ABSTRACT FROM AUTHOR]

High-pressure die casting (HPDC) is a permanent mold-based production technology that facilitates the casting of near net shape components from nonferrous alloys. The pressure and temperature conditions within the cavity impact the cast product quality during and after the conclusion of the die filling process. Die surface cavity sensors can deliver information describing the conditions at the die-casting interface. They are associated with high costs and limited service lifetimes below the achievable total cycle count of the die inserts and therefore ill-suited for industrial use cases. In this work, the suitability of long short-term memory (LSTM) recurrent neural networks (RNN) for substituting physical cavity temperature and pressure sensors virtually after the production ramp-up or at the end of the sensor service life is investigated. Training LSTMs with data of 233 casting cycles with different process parameters provides networks which are then applied to 99 further cycles. The prediction accuracy is investigated for different time interval lengths in the solidification and cooling phase. For longer time intervals, the cavity pressure prediction deteriorates, potentially due to a highly individual and hardly ascertainable buildup of casting distortion and internal stresses. Overall, however, the accuracy of the developed LSTMs is excellent for the cavity temperatures and good for the cavity pressures. [ABSTRACT FROM AUTHOR]

With the surge in mobile data traffic, leveraging Wi-Fi has become pivotal to alleviate the strain on LTE networks. An essential challenge lies in selecting the optimal Wi-Fi access point during Handover (HO), impacting Quality of Experience (QoE) and connection stability. This study introduces an innovative approach to enhance VoIP call Quality of Service (QoS) during HOs from LTE to Wi-Fi. Our contribution integrates two Mean Opinion Score (MOS) prediction models, multilayer perceptron (MLP) and recurrent neural networks (LSTMs) into the software-defined network controller (SDN), optimized by the Grey Wolf Optimizer (GWO). This solution provides accurate predictions for the SDN to forward mobile devices (MEs) and traffic to the nearest optimal Wi-Fi access points. Simulation results affirm the effectiveness of our approach. The LSTM-based approach performs well in the long run and provides better stability in terms of QoS parameters, especially when the metric of choice for mobile users is based on the Signal-to-Noise and Interference Ratio (SNIR), rather than the Received Signal Strength (RSSI) indicator. These results explore opportunities for the real-world deployment of the implemented models, particularly in urban areas experiencing cellular network congestion. [ABSTRACT FROM AUTHOR]

Airborne diseases have contributed in an incredible way to increase worldwide morbidity and disability during the last few years. The omicron is an airborne disease that spreads rapidly all over the world, with many infected persons. As a result, detecting and identifying the positive cases of omicron for treatment on time, and preventing the disease from spreading is a constant challenge and difficult process. Several medical tests have been used and analyzed just to identify omicron but all are ineffective and not properly accurate at an early stage. The current study examines the enormous potential of the latest progress of IoT after incorporating a framework using fog-cloud computing for omicron viruses. This paper presented a novel framework based on IoT technology for diagnosis such disease on time. The proposed model uses the weighted naive bayes algorithm that has ability to analyze the health condition and predict probabilistic health state vulnerability of an individual. In addition, the porposed framework indulges the utilization of Long Short-Term-Memory and Enhanced temporal data-based Re-current neural network (LSTM-ERNN) for the predictivity of the spread of such virus. The experimental outcomes indicate that the proposed system achieves an accuracy of 98.7%, a Recall of 97.8%, a precision of 92.8%, and the f-measure consists of 97.4%. The experimental results depicted that the proposed approach has performed better outcomes as compared to other state-of-the-art approaches. [ABSTRACT FROM AUTHOR]

Premature babies scream to make contact with their mothers or other people. Infants communicate via their screams in different ways based on the motivation behind their cries. A considerable amount of work and focus is required these days to preprocess, extract features, and classify audio signals. This research aims to propose a novel Elephant Herding Optimized Deep Convolutional Gated Recurrent Neural Network (EHO-DCGR net) for classifying cry signals from premature babies. Cry signals are first preprocessed to remove distortion caused by short sample times. MFCC (Mel-frequency cepstral coefficient), Power Normalized Cepstral Coefficients (PNCC), BFCC (Bark-frequency cepstral coefficient), and LPCC (Linear Prediction cepstral coefficient) are used to identify abnormal weeping through their prosodic aspects. The Elephant Herding optimization (EHO) algorithm is utilized for choosing the best features from the extracted set to form a fused feature matrix. These characteristics are then used to categorize premature baby cry sounds using the DCGR net. The proposed EHO-DCGR net effectiveness is measured by precision, specificity, recall, and F1-score, accuracy. According to experimental fallouts, the proposed EHO-DCGR net detects baby cry signals with an astounding 98.45% classification accuracy. From the experimental analysis, the EHO-DCGR Net increases the overall accuracy by 12.64%, 3.18%, 9.71% and 3.50% better than MFCC-SVM, DFFNN, SVM-RBF and SGDM respectively. [ABSTRACT FROM AUTHOR]

Cybersecurity has become increasingly important because of the widespread use of data and its enormous global storage. Hackers and other invaders always want to breach data security by interfering with network traffic. The breaches must be stopped by several tools, such as firewalls. Other solutions, such as intrusion detection systems (IDSs), may detect network intrusions effectively. In this paper, we introduce a hybrid technique (CNN-LSTM) that combines the convolutional neural network (CNN) with long short-term memory (LSTM), a modified version of the recurrent neural network (RNN). The model is tested using the CSE-CIC-IDS2018 dataset. Both CNN and LSTM were individually applied to the datasets, and the results are compared with our hybrid CNN-LSTM model. The hybrid CNN-LSTM model demonstrated higher accuracy (99%) during both training and validation processes compared to individual models; the accuracy of the CNN model is 92% and the accuracy of the LSTM is 93.5%. The outcomes validate the usefulness and effectiveness of the hybridizing model. [ABSTRACT FROM AUTHOR]

This paper addresses the challenges of exponentially growing traffic in cellular networks by proposing a novel predictive model, Hybrid Graph Convolutional Recurrent Network (HGCRN), which combines static graph convolutional recurrent neural network and meta-graph learning. The model is designed to effectively capture the complex spatio-temporal dependencies in network traffic, enhancing prediction accuracy and operational efficiency. By constructing graph adjacency matrices that go beyond mere geographical proximity, HGCRN offers a deeper understanding of the dynamic interactions within the network. Tested on real-world datasets from Telecom Italia and China Mobile, the model demonstrates significant improvements over traditional and state-of-the-art methods in terms of predictive accuracy and reliability. HGCRN outperforms other models in terms of MAE, MAPE, and RMSE, specifically, our model achieves a lower RMSE of 6.53 compared to the state-of-the-art on a publicly available Telecom dataset, and an RMSE of 11.79, outperforming existing methods in the step-12 setting on the China Mobile dataset. [ABSTRACT FROM AUTHOR]

In recent years, solar irradiance forecasting has become essential to managing, developing, and effectively integrating photovoltaic (PV) systems properly into the smart grid. The foundation of a conventional variational autoencoder (VAE) is an entirely coupled layer that includes both decoder and encoder components. In this study, a novel deep attention-driven model for forecasting named bidirectional long short-term memory (BiLSTM) which is combined with the VAE model is introduced as an enhanced version of the VAE. BiLSTM is integrated at the encoder side of VAE to effectively extract and learn temporal dependencies that are embedded in the panel irradiance data. Additionally, a self-attention mechanism (SAM) is added to bilateral variational autoencoder (BiVAE) which is known as BiVAE-SAM that highlights the important characteristics. The proposed BiVAE-SAM permits the VAE's capacity to design the temporal dependency. The examined models are assessed using sun irradiance measurements from New York City, Turkey, Canopy, Los Angeles, California, and Florida. The outcomes exhibit that the proposed BiVAESAM model performs better mean absolute percentage error (MAPE) with values of 1.7935, 0.7828, 1.3491 and 2.8346 respectively for California, Los Angeles, New York City, and Florida, over existing stacked denoising autoencoders (SDA) model at projecting solar irradiance. [ABSTRACT FROM AUTHOR]

The early prediction of mental illnesses reduces the severity of the disease. The symptoms like poor concentration, unstable energy of the body, pressure, and loss of interest cause depression. A large-scale group decision making (LSGDM) method is proposed in this paper along with the ensemble classifier model by combining convolutional neural network (CNN), recurrent neural network (RNN), and long short-term memory (LSTM) for effective classification of depression, anxiety and stress (DAS) levels. The data is collected from the depression anxiety stress scale-42 (DASS-42) dataset for efficient classification and predictions of mental health problems. The min-max normalization is used to pre-process the data, and the feature selection is done for extracting informative features. The extracted features are provided as input to the ensemble classifier. The proposed LSGDM model maximizes the classification accuracy with the help of grid search CV hyperparameter tuning, and results in an accuracy of 98.88%, precision of 98.21%, recall of 99.62%, F1-Score of 98.90%, and MCC of 99.41%. The proposed LSGDM method gives superior results when compared to the existing machine learning (ML) based ensemble model, a principal component analysis along with modified fast correlation based filtering (PCA-mFCBF), and LSTM based RNN (LSTM-RNN). [ABSTRACT FROM AUTHOR]

This research evaluates a number of hybrid recurrent neural network (RNN) architectures for classifying sequential sentences in biomedical abstracts. The architectures include long short-term memory (LSTM), bidirectional LSTM (BI-LSTM), gated recurrent unit (GRU), and bidirectional GRU (BIGRU) models, all of which are combined with the universal sentence encoder (USE). The investigation assesses their efficacy in categorizing sentences into predefined classes: background, objective, method, result, and conclusion. Each RNN variant is used with the pre-trained USE as word embeddings to find complex sequential relationships in biomedical text. Results demonstrate the adaptability and effectiveness of these hybrid architectures in discerning diverse sentence functions. This research addresses the need for improved literature comprehension in biomedicine by employing automated sentence classification techniques, highlighting the significance of advanced hybrid algorithms in enhancing text classification methodologies within biomedical research. [ABSTRACT FROM AUTHOR]

The importance of an effective cross-site scripting (XSS) detection system cannot be overstated in web security. XSS attacks continue to be a prevalent and severe threat to web applications, making the need for robust detection systems more crucial than ever. This paper introduced a hybrid model that leverages deep learning algorithms, combining recurrent neural network (RNN) and convolutional neural network (CNN) architectures. Our hybrid RNN-CNN model emerged as the top performer in our evaluation, demonstrating outstanding performance across key metrics. It achieved an impressive accuracy of 96.74%, excelling inaccurate predictions. Notably, the precision score reached an impressive 97.78%, highlighting its precision in identifying positive instances while minimizing false positives. Furthermore, the model's recall score of 95.65% showcased its ability to capture a substantial portion of true positive instances. This resulted in an exceptional F1-Score of 96.70, underlining the model's remarkable balance between precision and recall. Compared to other models in the evaluation, our proposed model unequivocally demonstrated its leadership, emphasizing its excellence in detecting potential XSS vulnerabilities within web content. [ABSTRACT FROM AUTHOR]

Orthogonal frequency division multiplexing (OFDM) is a famous multi-carrier modulation technique as it has a vast range of features like robustness against multi-path fading, higher bandwidth efficiency, and higher data rates. Though, OFDM has its own challenges. Among them, high peak power to average power ratio (PAPR) of the transmitted signal is the major problem in OFDM. In recent years, deep learning has drastically enhanced the performance of PAPR. In addition, the excessive training data and high computational complexity lead to a considerable issue in OFDM system. Thus, this paper implements a new PAPR reduction scheme in OFDM Systems through hybrid deep learning algorithms. A new optimized hybrid deep learning termed O-DNN + RNN is implemented by integrating the deep neural networks (DNN) and recurrent neural networks (RNN), where the parameters of both DNN and RNN are optimized using Hybrid Reptile Dragonfly Search Algorithm (HR-DSA). The new deep learning model is adopted for determining the constellation mapping and demapping of symbols on each subcarrier. This new optimized hybrid deep learning helps in reducing the PAPR and maximizes the performance. [ABSTRACT FROM AUTHOR]

Vehicular Ad Hoc Networks (VANETs) are becoming important with the advancement of connected technologies, yet they grapple with the pivotal challenge of ensuring quality of service (QoS) and quality of experience (QoE). This stems from the inherently erratic nature and huge data volumes. While reliability and efficiency remain paramount, the need to address QoS and QoE emerges as a critical motivation. Most current algorithms that address streaming data fall short in terms of QoS performance metrics. Thus this work strives to improve upon QoS metrics and further improve QoE. Our proposed method uses deep learning to address these problems in a realistic software-defined vehicular network (SDVN) based on the QoS and QoE. Our research aims to combine SDVN with a recurrent neural network (RNN) in Ryu SDVN controller. The RNN model encompasses a sophisticated architecture comprising multiple layers of recurrent units designed to capture temporal dependencies in data. Through a meticulously crafted training methodology utilizing techniques such as backpropagation through time, it learns to predict future network states based on historical data. Fine-tuning hyperparameters such as the number of epochs and the batch size enables optimal model convergence. We examine this method in a realistic simulation and compare its effectiveness with conventional approaches. The results show significant gains, i.e., marginal to 28% better than the nearest rival and far better than VANET. We also evaluate the network's resilience by varying transmission rate and packet size. Our method functions well in high-density situations, suggesting that real-world deployments can benefit from it. [ABSTRACT FROM AUTHOR]

This paper proposes a new approach to counter cyberattacks using the increasingly diverse malware in cyber security. Traditional signature detection methods that utilize static and dynamic features face limitations due to the continuous evolution and diversity of new malware. Recently, machine learning-based malware detection techniques, such as Convolutional Neural Networks (CNN) and Recurrent Neural Networks (RNN), have gained attention. While these methods demonstrate high performance by leveraging static and dynamic features, they are limited in detecting new malware or variants because they learn based on the characteristics of existing malware. To overcome these limitations, malware detection techniques employing One-Shot Learning and Few-Shot Learning have been introduced. Based on this, the Siamese Network, which can effectively learn from a small number of samples and perform predictions based on similarity rather than learning the characteristics of the input data, enables the detection of new malware or variants. We propose a dual Siamese network-based detection framework that utilizes byte images converted from malware binary data to grayscale, and opcode frequency-based images generated after extracting opcodes and converting them into 2-gram frequencies. The proposed framework integrates two independent Siamese network models, one learning from byte images and the other from opcode frequency-based images. The detection models trained on the different kinds of images generated separately apply the L1 distance measure to the output vectors the models generate, calculate the similarity, and then apply different weights to each model. Our proposed framework achieved a malware detection accuracy of 95.9% and 99.83% in the experiments using different malware datasets. The experimental results demonstrate that our malware detection model can effectively detect malware by utilizing two different types of features and employing the dual Siamese network-based model. [ABSTRACT FROM AUTHOR]

Natural answer generation is in a very clear practical significance and strong application background, which can be widely used in the field of knowledge services such as community question answering and intelligent customer service. Traditional knowledge question answering is to provide precise answer entities and neglect the defects; namely, users hope to receive a complete natural answer. In this research, we propose a novel attention-based recurrent neural network for natural answer generation, which is enhanced with multi-level copying mechanisms and question-aware loss. To generate natural answers that conform to grammar, we leverage multi-level copying mechanisms and the prediction mechanism which can copy semantic units and predict common words. Moreover, considering the problem that the generated natural answer does not match the user question, question-aware loss is introduced to make the generated target answer sequences correspond to the question. Experiments on three response generation tasks show our model to be superior in quality while being more parallelizable and requiring significantly less time to train. Our model achieves 0.727 BLEU on the SimpleQuestions response generation task, improving over the existing best results by over 0.007 BLEU. Our model has scored a significant enhancement on naturalness with up to 0.05 more than best performing baseline. The simulation results show that our method can generate grammatical and contextual natural answers according to user needs. [ABSTRACT FROM AUTHOR]

An efficient transportation system is conducive to maintaining traffic flow and safety. Passenger flow forecasting (PFF), an area of traffic forecasting, is a key part of the efficient transportation system. In recent years, deep-learning-based models have led to extensive research on the different conditions in this field. Hence, model determination is the most suitable for a specific application would be a key advantage. To address this issue, a comparative analysis of nine typical deep network approaches, including recurrent neural network (long short-term memory (LSTM) and gated recurrent unit (GRU)), bidirectional-based RNN (BiLSTM and BiGRU), convolutional neural network (CNN) (CNN1D and CNN2D), convolutional LSTM, and hybrid network (CNN1D-LSTM and CNN1D-GRU), for hourly PFF has been conducted. This comparison utilized two datasets with hourly records of bus lines from Guangzhou, China. The comparison results show that the bidirectional-based models were slightly better than other candidates in terms of the values of the root-mean-square error, determination coefficient, and Theil coefficient, and had a lower individual error distribution than the others, both numerically and proportionally. Furthermore, the bidirectional-based models were different from the other models in terms of the Friedman test (a special case is that the BiLSTM and LSTM had no significant difference for Line 10 application). Besides, the results from model structure aspect indicated that the bidirectional-based models achieved better performance with stable and reliable model structure (measurement index: posterior error distribution) and less computational complexity (measurement index: number of floating-point operations and parameters). It is concluded that the bidirectional-based deep learning models are the preferential choice for hourly PFF. [ABSTRACT FROM AUTHOR]

Urban air pollution, a combination of industry, traffic, forest burning, and agriculture pollutants, significantly impacts human health, plants, and economic growth. Ozone exposure can lead to mortality, heart attacks, and lung damage, necessitating the creation of complex environmental safety regulations by forecasting ozone concentrations and associated pollutants. This study proposes a hybrid method, RFNN-GOA, combining recurrent fuzzy neural network (RFNN) and grasshopper optimization algorithm (GOA) to estimate and forecast the daily ozone (O 3 ) in specific urban areas, specifically Kopački Rit and Osijek city in Croatia, aiming to improve air quality, human health, and ecosystems. Due to the intricate structure of atmospheric particles, modeling of O 3 likely poses the biggest challenge in air pollution today. The dataset used by the proposed RFNN-GOA model for the prediction of O 3 concentrations in each explored area consists of the following air pollutants, NO, NO 2 , CO, SO 2 , O 3 , PM 10 , and PM 2.5 ; and five meteorological elements, including temperature, relative humidity, wind direction, speed, and pressure. The RFNN-GOA method optimizes membership functions' parameters and the rule premise, demonstrating robustness and reliability compared to other identifiers and indicating its superiority over competing methods. The RFNN-GOA method demonstrated superior accuracy in Osijek city and Kopački Rit area, with variance-accounted for (VAF) values of 91.135%, 83.676%, 87.807%, 79.673% compared to the RFNN method's corresponding values of 85.682%, 80.687%, 80.808%, 74.202% in both training and testing phases, respectively. This reveals that RFNN-GOA increased the average VAF in Osijek city and Kopački Rit area by over 5% and 8%, respectively. [ABSTRACT FROM AUTHOR]

Due to the increasing digitization of city processes, there has been a significant shift in how cities are governed and how people make their living. However, several types of attacks could target smart cities, and Flooding Attacks (FA) are the most dangerous type. It is also a major issue for many people and programs using the Internet nowadays. Security in smart cities refers to preventative measures necessary to shield the city and its residents from direct or indirect harm by attackers who try to crash the system and deny legitimate users the use of the services. Smart city security, in contrast to standard security mechanisms, necessitates new and creative approaches to protecting the systems and applications while considering characteristics like resource limitations, distributed architecture nature, and geographic distribution. Smart cities are vulnerable to several particular issues, including faulty communication, insufficient data, and privilege protection. Therefore, a hybrid CRNN model that consists of Convolutional Neural Network (CNN) and Recurrent Neural Network (RNN) algorithms is employed for the detection of Flood Attacks based on the classification of traffic data. Subsequently, the performance of the CRNN is tested and evaluated using the CIC-Bell-DNS-EXF-2021 dataset. The obtained accuracy results of the proposed CRNN model achieved in FA detection is 99.2%. [ABSTRACT FROM AUTHOR]

In new technologies like fog computing, edge computing, cloud computing, and the Internet of Things (IoT), cybersecurity concerns and cyber-attacks have surged. The demand for better threat detection and prevention systems has increased due to the present global uptick in phishing and computer network attacks. In order to identify irregularities and attacks on the network, which have increased in scale and prevalence, threat identification is essential. However, the community is forced to investigate and create novel threat detection approaches that are capable of detecting threats using anomalies due to the increase in network threats, the growth of new methods of attack and computations, and the requirement to ensure security measures. A novel mechanism is employed to identify threats in a data based on optimized deep learning. The main aim of this paper is the usage of data classification system based on Deep Learning (DL). The proposed mechanism employed the TCP (Transmission Control Protocol) communication protocol to extract data from loud IoT (Internet of Things) networks for the purpose of threat detection. To perform feature extraction an Ant Colony Optimization (ACO) is utilised, through Recurrent Neural Network (RNN), the attacks in data are classified and detected. Additionally, the suggested approach has been evaluated and trained using the BOUN DDoS contemporary dataset, which comprises a variety of attack types and allows for the effectiveness of the framework to be determined to compare it to previous approaches. The Findings indicate that the suggested approach achieved higher accuracy in DDoS attack identification in comparison with Traditional deep learning methods. The existing method detects the generic attack with lower efficiency however; the proposed mechanism achieves better accuracy in both the detection of the DDoS attack and the detection of regular traffic. [ABSTRACT FROM AUTHOR]

Herbal leaves, known for their efficacy in treating a range of infectious diseases including cancer, asthma, and heart conditions, are still widely used by medical professionals. Traditionally, villagers have identified these plants visually, but given the similarity in appearance among various species, this method is prone to human error. Accurate identification of these plant species is critical for effective treatment. Hence, the development of an intelligent plant classification system is crucial to reduce the risk of misidentification and enhance treatment accuracy. This paper introduces the deep HybridNet with hybrid optimization module (DeepHybrid-OptNet) a novel deep learning framework for medicinal plant identification and classification. Merging convolutional and recurrent neural network architectures, deep HybridNet excels in extracting complex botanical features through channel-wise feature extraction modules in convolutional neural network (CNN) and feedback loop in recurrent neural network (RNN). The incorporation of a DeepHybrid-OptNet module enhances the model's learning efficiency and accuracy. Empirical results on the Mendley and folio dataset demonstrate the framework's superiority over existing methods in accuracy, precision, and recall making it a valuable asset for botany and herbal medicine research. [ABSTRACT FROM AUTHOR]

The article focuses on enhancing stock market price prediction through artificial neural networks and machine learning. It underscores the significance of improving forecast accuracy by incorporating historical stock prices, macroeconomic indicators, news events, and technical indicators. Exploring deep learning principles, it delves into convolutional neural networks (CNN), recurrent neural networks (RNN), including long shortterm memory (LSTM), and gated recurrent unit (GRU) modifications. This financial time series processing study covers data preprocessing, creating training/test sets, and selecting evaluation metrics. Results suggest promising applications for the developed forecasting models in stock markets, stressing the importance of considering various factors for precise forecasts in dynamic financial environments. Historical reserve data serves as the model foundation. Integration of macroeconomic, news, and technical indicators offers a holistic approach, aiding trend and anomaly identification for enhanced forecasts. The article recommends suitable deep learning architectures, highlighting LSTM and GRU's effectiveness in adapting to intricate data dependencies. Experimental outcomes showcase these architectures' benefits in predicting stock market prices, offering valuable insights for finance and asset management professionals in financial analysis and machine learning realms. [ABSTRACT FROM AUTHOR]

Gene fusions are vital biomarkers for tumor diagnosis and drug development, with precise detection becoming increasingly important. This review explores the links between gene fusions and common tumors, systematically evaluating detection technologies like fluorescence in situ hybridization (FISH), polymerase chain reaction (PCR), immunohistochemistry (IHC), electrochemiluminescence (ECL), and next-generation sequencing (NGS). FISH is the gold standard for DNA-level rearrangements, while PCR and NGS are widely used, with PCR confirming known fusions and NGS offering comprehensive genome-wide detection. Bioinformatic tools like STAR-Fusion, FusionCatcher, and Arriba are assessed for diagnostic accuracy. The review highlights how artificial intelligence (AI), particularly deep learning (DL) technologies like convolutional neural networks (CNNs) and recurrent neural networks (RNNs), is transforming gene fusion research by accurately detecting and annotating genes from genomic data, eliminating biases. Finally, we present an overview of advanced technologies for gene fusion analysis, emphasizing their potential to uncover unknown gene fusions. Highlights: • This review explores the intricate relationship between gene fusions and prevalent tumors, comprehensively reviewing techniques for gene fusion testing. • It evaluates the strengths and limitations of key assays like fluorescence in situ hybridization, polymerase chain reaction, and next-generation sequencing, and explores the emerging use of electrochemiluminescence for high-sensitivity gene fusion detection. • It emphasizes the diagnostic accuracy of bioinformatics tools such as STAR-Fusion, FusionCatcher, Arriba, and underscores the transformative role of artificial intelligence, particularly deep learning technologies, in revolutionizing gene fusion research. [ABSTRACT FROM AUTHOR]

A fricative sound is produced by the close proximity of two articulators, resulting in a partially obstructed airstream and turbulent airflow. The frequency spectrum of the majority of fricatives is similar to noise. This peculiarity presents a significant challenge in their numerical processing. Nonetheless, the literature contains a large number of works for fricatives, particularly those of Latin languages. However, very few are related to the Arabic language. The objective of this article is to create and validate a dataset of 14 Arabic fricatives using an acoustic characterization approach. We recorded a speech corpus from thirty speakers for this purpose, yielding 25 200 units of analysis. The data was then analyzed and validated using deep learning-based classification methods. As a result, two distinct approaches were proposed. The first approach is based on the use of three deep neural networks (Convolutional Recurrent Neural Network (CRNN), Vision Transformers (ViT), and ResNet50) whereas their input is one of three feature extraction techniques (spectrogram, Mel-spectrogram, and Mel-frequency Cepstral Coefficient (MFCC)) from speech units /Vowel-Consonant-Vowel/ (or /VCV/). The best classification rate observed when using the Mel spectrogram with the CRNN model was 94.07%, while the ViT model surpassed this with a rate of 95.64%. This rate was improved to 97.5% by proposing a second approach based on a hierarchical H-CRNN model with fricatives organized into categories (sibilance, voicing, and place of articulation). The obtained results surpassed the traditional approaches proposed in the literature for the classification of Arabic fricatives due to the hierarchical approach. [ABSTRACT FROM AUTHOR]

During delayed ballistic reaches, motor areas consistently display movement-specific activity patterns prior to movement onset. It is unclear why these patterns arise: while they have been proposed to seed an initial neural state from which the movement unfolds, recent experiments have uncovered the presence and necessity of ongoing inputs during movement, which may lessen the need for careful initialization. Here, we modeled the motor cortex as an input-driven dynamical system, and we asked what the optimal way to control this system to perform fast delayed reaches is. We find that delay-period inputs consistently arise in an optimally controlled model of M1. By studying a variety of network architectures, we could dissect and predict the situations in which it is beneficial for a network to prepare. Finally, we show that optimal input-driven control of neural dynamics gives rise to multiple phases of preparation during reach sequences, providing a novel explanation for experimentally observed features of monkey M1 activity in double reaching. [ABSTRACT FROM AUTHOR]

In an era of consumer electronics, effective consumption recommendation contributes a lot to improving benefits for online shopping operators. However, consumers and products constitute a complex heterogeneous information network, in which extraction of structural features is essential. To construct more fine-grained feature space for modeling, we introduce the graph neural network (GNN) theory for this purpose. Accordingly, a multi-scale GNN-based personalized recommender system for online consumption decision is proposed in this paper. Firstly, we set up an encoder for consumption decision based on a multi-scale GNN structure, and define the loss function. Two realistic datasets are utilized as the research scenario, in order to complete the feature combination of online consumption. Then, personalized recommendation settings are completed based on a recurrent neural network structure. And a graph learning module is embedded in it to integrate cross-attention mechanism to establish the consumption decision algorithm. Finally, the proposed method is tested by comparing with relevant research methods under several metrics: adaptability, success rate, and stability. The results show that the proposed method has achieved some improvement in accuracy, coverage, and recommendation speed. [ABSTRACT FROM AUTHOR]