Advanced Deep Learning Models Research Articles

High-accuracy gearbox fault detection using deep learning on vibrational data

Abstract In the face of escalating demands for reliability in industrial machinery, this study introduces an advanced deep-learning model aimed at the early detection of faults in rotating gear systems. Addressing the need for timely fault diagnosis to prevent operational disruptions and financial setbacks, our work utilizes a novel neural network approach applied to vibrational data from Kaggle’s Gearbox Fault Diagnosis Dataset. The methodology involves data preprocessing steps, including normalization and one-hot encoding, followed by training a sequential neural network with multiple dense layers and dropout for regularization. The model, trained using the Adam optimizer and evaluated over 50 epochs, significantly outperforms traditional machine learning techniques, achieving a remarkable 98.63% accuracy in distinguishing between healthy and compromised gear conditions. This research marks a significant stride in predictive maintenance, providing a robust foundation for future development toward prognostic health management systems that anticipate maintenance needs and ensure uninterrupted industrial productivity.

ADVANCED FRAUD DETECTION IN CARD-BASED FINANCIAL SYSTEMS USING A BIDIRECTIONAL LSTM-GRU ENSEMBLE MODEL

This article addresses the challenges of fraud in card-based financial systems and proposes effective detection and prevention strategies. By leveraging recent data analytics and real-time monitoring, the study aims to enhance transaction security and integrity. The authors review existing fraud detection methodologies, emerging trends, and the evolving tactics of fraudsters, emphasizing the importance of collaboration among financial institutions, regulatory agencies, and technology providers. Our proposed solution is an ensemble model combining Bidirectional Gated Recurrent Unit (BiGRU) and Bidirectional Long Short-Term Memory (BiLSTM) networks, designed to capture complex transactional patterns more effectively. Comparative analysis of six machine learning classifiers—AdaBoost, Naïve Bayes, Decision Tree, Logistic Regression, Random Forest, and Voting—demonstrates that our BiLSTM-BiGRU ensemble model outperforms traditional methods, achieving a fraud detection performance score of 89.22%. This highlights the advanced deep learning model's superior ability to enhance the robustness and reliability of fraud detection systems.

A dataset of manually annotated filaments from H-alpha observations.

We present the Manually Annotated GONG Filaments in H-alpha Observations (MAGFiLO v1.0) dataset. This dataset contains 10,244 annotated filaments from 1,593 observations captured by the Global Oscillation Network Group (GONG), spanning the years 2011 through 2022. Each annotation details one filament's segmentation, minimum bounding box, spine, and magnetic field chirality. With a total of over one thousand person-hours of annotation, and a double-blind review process, we ensured high-quality ground-truth data. Our inter-annotator agreement reaches a Kappa score of 0.66. We also verified that the hemispheric preference of filaments as annotated in MAGFiLO aligns with the findings from similar datasets of much smaller sample sizes. MAGFiLO is the first dataset of its size, enabling advanced deep learning models to identify filaments and their features with unprecedented precision. It also provides a testbed for solar physicists interested in large-scale analysis of filaments. In this report, we document the details of the annotation and the post-processing phases that were applied.

DEEP LEARNING MODEL FOR DETECTING TERROR FINANCING PATTERNS IN FINANCIAL TRANSACTIONS

Terror financing remains a critical threat to global security, with illicit actors continually adapting their methods to evade detection. Traditional financial monitoring systems often struggle to identify the complex and covert patterns associated with terror-related transactions due to their reliance on predefined rules and statistical thresholds. This study introduces advanced deep learning models designed to detect terror financing patterns within vast datasets of financial transactions. We developed and evaluated several neural network architectures, including Convolutional Neural Networks (CNNs) and Long Short-Term Memory networks (LSTMs), to capture both spatial and temporal transaction features. The models were trained and tested on a dataset comprising anonymized financial transactions labeled for suspicious activities related to terror financing. Our deep learning models demonstrated superior performance over conventional machine learning approaches, achieving higher accuracy, precision, and recall in identifying suspicious transactions. Notably, the LSTM-based model excelled in detecting sequential transaction patterns indicative of layering and integration stages commonly used in terror financing. The results underscore the potential of deep learning techniques in enhancing the capabilities of financial institutions and regulatory bodies to combat terror financing. Implementing such models can lead to more proactive and effective monitoring systems that adapt to evolving illicit financing strategies.

A novel energy management of public charging stations using attention-based deep learning model

Electricity grids are complex systems that must balance the supply and demand of electricity in real-time. However, with the increasing adoption of electric vehicles (EVs), managing the grid’s stability has become more challenging. EV charging can cause spikes in electricity demand, leading to peak demand periods that strain the power grid’s infrastructure. With the help of load forecasting, this effect on the grid can be mitigated by predicting the charging demand of electric vehicles in advance. This will help utilities adjust their energy supply in real-time, ensuring enough energy is available to meet demand, and preventing overloads or under utilization of the grid. Moreover, the EV charging demand is influenced by a wide range of factors, including charging station locations, weather, and time of day. Therefore, advanced deep learning models are required to learn these complex relationships and identify patterns in EV charging demand, enabling utilities to make more informed decisions. In this research, an attention-based deep learning approach is proposed for more accurate prediction of EV load demand. This novel approach integrates attention mechanisms with traditional deep learning models like LSTM and GRU, allowing the model to dynamically weight the importance of different features and focus on the most relevant information. The outcomes are compared to conventional deep learning and machine learning algorithms. To test the efficacy of the proposed framework, an actual ACN dataset for public EV charging stations is utilized.

Drone-based inspection of wind turbine blades: a comparative study of deep learning models

Maintaining wind turbine blades is a challenging task, often marked by high costs, safety risks, time inefficiency, and the possibility of incorrect diagnosis. A promising approach to support preventive maintenance involves the use of drones and deep learning for inspection and early fault detection. This paper offers a comparative study of five advanced deep learning models: Residual Networks (ResNet), Inception Networks (InceptionV3), YOLO (You Only Look Once), EfficientNet, and Vision Transformers (ViTs) applied to the DTU Drone Inspection Images of Wind Turbine dataset. The objective is to determine the most effective model for detecting defects in wind turbine blades, which is critical for ensuring the efficiency and safety of wind energy systems. The models were assessed using key performance metrics, such as accuracy, precision, recall, F1-score, and inference time. EfficientNet emerged as the top performer with the highest accuracy and balanced efficiency, while YOLO demonstrated the fastest inference time, making it ideal for real-time applications. ResNet and InceptionV3 also performed well, particularly in detecting subtle defects, whereas ViTs showed strong capabilities in capturing complex global features despite longer processing times. The findings provide valuable insights for selecting appropriate deep learning models in drone-based wind turbine inspections, contributing to more effective and reliable maintenance practices.

Real-Time Fire Classification Models Based on Deep Learning for Building an Intelligent Multi-Sensor System

Fire detection systems are critical for mitigating the damage caused by fires, which can result in significant annual property losses and fatalities. This paper presents a deep learning-based fire classification model for an intelligent multi-sensor system aimed at early and reliable fire detection. The model processes data from multiple sensors that detect various parameters, such as temperature, humidity, and gas concentrations. Several deep learning architectures were evaluated, including LSTM, GRU, Bi-LSTM, LSTM-FCN, InceptionTime, and Transformer. The models were trained on data collected from controlled fire scenarios and validated for classification accuracy, loss, and real-time performance. The results indicated that the LSTM-based models (particularly Bi-LSTM and LSTM) could achieve high classification accuracy and low false alarm rates, demonstrating their effectiveness for real-time fire detection. The findings highlight the potential of advanced deep-learning models to enhance the reliability of sensor-based fire detection systems.

A Comprehensive Study of Ocular Disease Prediction Through Multi-Modal Fusion Using Deep Learning Approach

In recent years, the prevalence of ocular diseases has surged globally, necessitating advanced diagnostic methods to enhance early detection and treatment outcomes. This study presents a comprehensive analysis of ocular disease prediction through using deep learning techniques, aiming to address the limitations of traditional diagnostic methods by leveraging the strengths of multiple data sources. A key aspect of the study is the application of advanced deep learning models, specifically convolutional neural networks (CNNs) for image analysis and recurrent neural networks (RNNs) for temporal and sequential data processing. In conclusion, this research provides valuable insights into the application of multi-modal fusion in ocular disease prediction and sets the stage for future developments in integrating diverse data types for more accurate and timely diagnosis.

Combined ResNet Attention Multi-Head Net (CRAMNet): A Novel Approach to Fault Diagnosis of Rolling Bearings Using Acoustic Radiation Signals and Advanced Deep Learning Techniques

The fault diagnosis of rolling bearing acoustic radiation signals holds significant importance in industrial equipment maintenance. It effectively prevents equipment failures and downtime, ensuring the smooth operation of the production process. Compared with traditional vibration signals, acoustic radiation signals have the advantage of non-contact measurement. They can diagnose faults in special conditions where sensors cannot be installed and provide more comprehensive equipment status information. Therefore, to extract the fault characteristic information of rolling bearings from complex acoustic signals, this paper proposes an advanced deep learning model combining Gramian Angular Field (GAF), ResNet1D, ResNet2D, and multi-head attention mechanism, named CRAMNet (Combined ResNet Attention Multi-Head Net), to diagnose the faults of rolling bearing acoustic radiation signals. Firstly, this method includes converting one-dimensional signals into GAF images and performing data standardization and segmentation. Then, the method utilizes ResNet1D to extract features from one-dimensional signals and ResNet2D to extract features from GAF images. Further, it combines the multi-head attention mechanism to enhance feature representation and capture dependencies between different channels. Finally, this paper compares the proposed method with several traditional models (including CNN, LSTM, DenseNet, and CNN-Transformers). Experimental results show that the proposed method performs outstandingly in terms of accuracy and robustness. The combination of residual networks and multi-head attention mechanism in the model significantly enhances its ability to accurately diagnose rolling bearing faults, proving the superiority of the algorithm.

COMPUTER VISION METHODS FOR CONDUCTING OSINT INVESTIGATIONS

This paper researches the application of computer vision techniques in concern with OSINT (Open Source Intelligence) investigations. It explores how state-ofthe- art algorithms and models in computer vision can be used to automate and enhance the process of gathering, analyzing, and interpreting visual data from open sources. The research focuses on the critical steps of image scraping, data preprocessing, and embedding generation using advanced deep learning models such as CLIP. Additionally, the study examines the challenges of managing large-scale visual data and implementing efficient search mechanisms through vector databases like Faiss and Weaviate. By applying these technologies, the paper illustrates how investigators can improve the accuracy and efficiency of imagebased searches, which are later used for uncovering hidden connections and verifying information in OSINT investigations. The findings contribute to the growing field of computer vision and intelligence gathering, offering practical recommendations for enhancing investigative processes through the integration of computer vision methodologies.

A hybrid carbon price forecasting model combining time series clustering and data augmentation

Stable and accurate carbon price forecasts are crucial for carbon asset management and policy design. Therefore, various advanced models have been proposed to forecast carbon price, and decomposition–reconstruction models occupy a major position. On this basis, this study proposes an improved decomposition–reconstruction framework and constructs a novel hybrid forecasting model Informer-DA-BOHB-TVFEMD-CL. The hybrid model decomposes carbon price by TVFEMD to obtain different volatility characteristics and classifies these features by time series clustering, where the high-frequency components are retained and the low-frequency components are reconstructed. Then, the training set was expanded by the generative data augment technique to improve the model's generalization ability and avoid overfitting. Finally, the advanced deep learning model Informer and the hyperparameter optimization technique BOHB are used to model the training set, and the prediction results of each subcomponent are integrated to obtain the final prediction results. Experiments are conducted in EU-ETS, Guangdong, Hubei, and Chinese national carbon market. Experimental results show that the stability and prediction performance of the proposed hybrid model are better than those of the benchmark models, proving the applicability of the improved decomposition–reconstruction idea and the validity of the hybrid model. Meanwhile, the proposed model has demonstrated good profitability in carbon trading.

Utilizing artificial intelligence for precision exploration of N protein targeting phenanthridine sars-cov-2 inhibitors: A novel approach

The persistent mutation of the novel coronavirus presents a continual threat of infections and associated illnesses. While considerable research efforts have concentrated on the functional proteins of SARS-CoV-2 in the development of anti-COVID-19 therapeutics, the structural proteins, particularly the N protein, have received comparatively less attention. This study focuses on the N protein, a critical structural component of the virus, and employs advanced deep learning models, including EMPIRE and DeepFrag, to optimize the structures of phenanthridine-based compounds. More than 10,000 small molecules, derived through deep learning, underwent high-throughput virtual screening, resulting in the synthesis of 44 compounds. Compound 38 showed a binding potential energy of −8.2 kcal/mol in molecular docking. Surface Plasmon Resonance (SPR) and Microscale Thermophoresis (MST) validation yielded dissociation constants of 353 nM and 726 nM, confirming strong binding to the N protein. Compound 38 demonstrated antiviral activity in vitro and exhibited anti-COVID-19 effects by interfering with the binding of N proteins to RNA. This research underscores the potential of targeting the SARS-CoV-2 N protein for therapeutic intervention and illustrates the efficacy of deep learning model in the design of lead compounds. The application of these deep learning models represents a promising approach for accelerating the discovery and development of antiviral agents.

Classifying Anemia Diseases based on VGG-16 and CBAM Attention Mechanism Models

Anemia is a major public health concern and has a significant impact on women and children worldwide. This condition is caused by a lack of sufficient Red Blood Cells (RBCs). Traditionally, anemia diagnosis has relied on invasive techniques that require blood samples, leading to pain and discomfort. This study explores deep learning techniques for automating and enhancing anemia detection accuracy using the Blood Cell Count and Detection (BCCD) dataset. We employ the VGG16 convolutional neural network architecture augmented with the Convolutional Block Attention Module (CBAM) to boost feature representation and performance. The BCCD dataset, consisting of 364 images with 4888 labeled blood cells, was used for training and evaluation. Our enhanced VGG16 model achieved accuracy for RBCs, WBCs, and platelets with values of 0.84, 0.93, and 0.92, respectively, effectively identifying various blood cell types and detecting anemia. Precision-recall analysis and confusion matrix metrics confirmed the robustness of the model, with high precision and recall rates and minimal false positives and negatives. These results suggest that advanced deep learning models with attention mechanisms, such as CBAM, can significantly improve medical diagnostics by providing reliable, efficient, and accurate early disease detection tools. Future development will be directed towards fine-tuning and validating the proposed model on other medical imaging datasets for clinical applications.

Comparing CNN-LSTM and BERTimbau: an analysis of AI models in legal document classification

This study addresses the urgent need of the Attorney General's Office of the State of Bahia (PGE) to automate the classification of initial petitions, a challenge exacerbated by the lack of standardization in file naming. To tackle this issue, the work proposes the implementation of advanced Deep Learning models, aiming to overcome the limitations of the currently used approach based on regular expressions (Regex), which shows an average accuracy of 80%. The research compares the efficacy of a hybrid model, integrating Convolutional Neural Network (CNN) with Long Short-Term Memory (LSTM), and the BERTimbau model, with the goal of not only enhancing the precision in identifying these essential documents but also promoting procedural efficiency through automation. Preliminary results reveal that the CNN-LSTM model achieved an accuracy of 99.34%, while BERTimbau obtained 98.51%, demonstrating the great potential of both techniques in optimizing the judicial workflow in the digital era.

Framework for multivariate carbon price forecasting: A novel hybrid model

The complex characteristics of volatility and non-linearity of carbon price pose a serious challenge to accurately predict carbon price. Therefore, this study proposes a new hybrid model for multivariate carbon price forecasting, including feature selection, deep learning, intelligent optimization algorithms, model combination and evaluation indicators. First, this study collects and organizes the historical carbon price series of Hubei and Shanghai as well as the influencing factors in five dimensions including structured and unstructured data, totaling twenty variables. Second, data dimensionality reduction is performed and input variables are obtained using the least absolute shrinkage and selection operator, followed by the introduction of nine advanced deep learning models to predict carbon price and compare the prediction effects. Then, through the combination of models, three models with the best performance are combined with Pelican optimization algorithm to construct a hybrid forecasting model. Finally, the experimental results show that the developed forecasting model outperforms other comparation models in terms of prediction accuracy, stability and statistical hypothesis testing, and exhibits excellent prediction performance. Furthermore, this study also applies the developed model to European carbon market price prediction and uses the Hubei carbon market as an example for quantitative trading simulation, and the empirical results further verify its robust prediction performance and investment application value. In conclusion, the proposed hybrid prediction model can not only provide high-precision carbon market price prediction for the government and corporate decision makers, but also help investors optimize their trading strategies and improve their returns.

Anterior Cruciate Ligament Tear Detection Based on T-Distribution Slice Attention Framework with Penalty Weight Loss Optimisation.

Anterior cruciate ligament (ACL) plays an important role in stabilising the knee joint, prevents excessive anterior translation of the tibia, and provides rotational stability. ACL injuries commonly occur as a result of rapid deceleration, sudden change in direction, or direct impact to the knee during sports activities. Although several deep learning techniques have recently been applied in the detection of ACL tears, challenges such as effective slice filtering and the nuanced relationship between varying tear grades still remain underexplored. This study used an advanced deep learning model that integrated a T-distribution-based slice attention filtering mechanism with a penalty weight loss function to improve the performance for detection of ACL tears. A T-distribution slice attention module was effectively utilised to develop a robust slice filtering system of the deep learning model. By incorporating class relationships and substituting the conventional cross-entropy loss with a penalty weight loss function, the classification accuracy of our model is markedly increased. The combination of slice filtering and penalty weight loss shows significant improvements in diagnostic performance across six different backbone networks. In particular, the VGG-Slice-Weight model provided an area score of 0.9590 under the receiver operating characteristic curve (AUC). The deep learning framework used in this study offers an effective diagnostic tool that supports better ACL injury detection in clinical diagnosis practice.

Enhancing Jujube Forest Growth Estimation and Disease Detection Using a Novel Diffusion-Transformer Architecture.

This paper proposes an advanced deep learning model that integrates the Diffusion-Transformer structure and parallel attention mechanism for the tasks of growth estimation and disease detection in jujube forests. Existing methods in forestry monitoring often fall short in meeting the practical needs of large-scale and highly complex forest areas due to limitations in data processing capabilities and feature extraction precision. In response to this challenge, this paper designs and conducts a series of benchmark tests and ablation experiments to systematically evaluate and verify the performance of the proposed model across key performance metrics such as precision, recall, accuracy, and F1-score. Experimental results demonstrate that compared to traditional machine learning models like Support Vector Machines and Random Forests, as well as common deep learning models such as AlexNet and ResNet, the model proposed in this paper achieves a precision of 95%, a recall of 92%, an accuracy of 93%, and an F1-score of 94% in the task of disease detection in jujube forests, showing similarly superior performance in growth estimation tasks as well. Furthermore, ablation experiments with different attention mechanisms and loss functions further validate the effectiveness of parallel attention and parallel loss function in enhancing the overall performance of the model. These research findings not only provide a new technical path for forestry disease monitoring and health assessment but also contribute rich theoretical and experimental foundations for related fields.

Developing a Dynamic Feature Selection System (DFSS) for Stock Market Prediction: Application to the Korean Industry Sectors

For several years, a growing interest among numerous researchers and investors in predicting stock price movements has spurred extensive exploration into employing advanced deep learning models. These models aim to develop systems capable of comprehending the stock market’s complex nature. Despite the immense challenge posed by the diverse factors influencing stock price forecasting, there remains a notable lack of research focused on identifying the essential feature set for accurate predictions. In this study, we propose a Dynamic Feature Selection System (DFSS) to predict stock prices across the 10 major industries, as classified by the FnGuide Industry Classification Standard (FICS) in South Korea. We apply 16 feature selection algorithms from filter, wrapper, embedded, and ensemble categories. Subsequently, we adjust the settings of industry-specific index data to evaluate the model’s performance and robustness over time. Our comprehensive results identify the optimal feature sets that significantly impact stock prices within each sector at specific points in time. By analyzing the inclusion ratios and significance of the optimal feature set by category, we gain insights into the proportion of feature classes and their importance. This analysis ensures the interpretability and reliability of our model. The proposed methodology complements existing methods that do not consider changes in the types of variables significantly affecting stock prices over time by dynamically adjusting the input variables used for learning. The primary goal of this study is to enhance active investment strategies by facilitating the creation of diversified portfolios for individual stocks across various sectors, offering robust models and feature sets that consistently demonstrate high performance across industries over time.

Enhancing Melanoma Diagnosis with Advanced Deep Learning Models Focusing on Vision Transformer, Swin Transformer, and ConvNeXt.

Skin tumors, especially melanoma, which is highly aggressive and progresses quickly to other sites, are an issue in various parts of the world. Nevertheless, the one and only way to save lives is to detect it at its initial stages. This study explores the application of advanced deep learning models for classifying benign and malignant melanoma using dermoscopic images. The aim of the study is to enhance the accuracy and efficiency of melanoma diagnosis with the ConvNeXt, Vision Transformer (ViT) Base-16, and Swin Transformer V2 Small (Swin V2 S) deep learning models. The ConvNeXt model, which integrates principles of both convolutional neural networks and transformers, demonstrated superior performance, with balanced precision and recall metrics. The dataset, sourced from Kaggle, comprises 13,900 uniformly sized images, preprocessed to standardize the inputs for the models. Experimental results revealed that ConvNeXt achieved the highest diagnostic accuracy among the tested models. Experimental results revealed that ConvNeXt achieved an accuracy of 91.5%, with balanced precision and recall rates of 90.45% and 92.8% for benign cases, and 92.61% and 90.2% for malignant cases, respectively. The F1-scores for ConvNeXt were 91.61% for benign cases and 91.39% for malignant cases. This research points out the potential of hybrid deep learning architectures in medical image analysis, particularly for early melanoma detection.

RS-Xception: A Lightweight Network for Facial Expression Recognition

Facial expression recognition (FER) utilizes artificial intelligence for the detection and analysis of human faces, with significant applications across various scenarios. Our objective is to deploy the facial emotion recognition network on mobile devices and extend its application to diverse areas, including classroom effect monitoring, human–computer interaction, specialized training for athletes (such as in figure skating and rhythmic gymnastics), and actor emotion training. Recent studies have employed advanced deep learning models to address this task, though these models often encounter challenges like subpar performance and an excessive number of parameters that do not align with the requirements of FER for embedded devices. To tackle this issue, we have devised a lightweight network structure named RS-Xception, which is straightforward yet highly effective. Drawing on the strengths of ResNet and SENet, this network integrates elements from the Xception architecture. Our models have been trained on FER2013 datasets and demonstrate superior efficiency compared to conventional network models. Furthermore, we have assessed the model’s performance on the CK+, FER2013, and Bigfer2013 datasets, achieving accuracy rates of 97.13%, 69.02%, and 72.06%, respectively. Evaluation on the complex RAF-DB dataset yielded an accuracy rate of 82.98%. The incorporation of transfer learning notably enhanced the model’s accuracy, with a performance of 75.38% on the Bigfer2013 dataset, underscoring its significance in our research. In conclusion, our proposed model proves to be a viable solution for precise sentiment detection and estimation. In the future, our lightweight model may be deployed on embedded devices for research purposes.