Deep Model Research Articles

Utilizing deep learning-based causal inference to explore vancomycin's impact on continuous kidney replacement therapy necessity in blood culture-positive intensive care unit patients.

Patients with positive blood cultures in the intensive care unit (ICU) are at high risk for septic acute kidney injury requiring continuous kidney replacement therapy (CKRT), especially when treated with vancomycin. This study developed a machine learning model to predict CKRT and examined vancomycin's impact using deep learning-based causal inference. We analyzed ICU patients with positive blood cultures, utilizing the Medical Information Mart for Intensive Care III data set. The primary outcome was defined as the initiation of CKRT during the ICU stay. The machine learning models were developed to predict the outcome. The deep learning-based causal inference model was utilized to quantitatively demonstrate the impact of vancomycin on the probability of CKRT initiation. Logistic regression was performed to analyze the relationship between the variables and the susceptibility of vancomycin. A total of 1,318 patients were included in the analysis, with 41 requiring CKRT. The Random Forest and Light Gradient Boosting Machine exhibited the best performance, with Area Under Curve of Receiver Operating Characteristic Curve values of 0.905 and 0.886, respectively. The deep learning-based causal inference model demonstrated an average 7.7% increase in the probability of CKRT occurrence when administrating vancomycin in total data set. Additionally, that younger age, lower diastolic blood pressure, higher heart rate, higher baseline creatinine, and lower bicarbonate levels sensitized the probability of CKRT application in response to vancomycin treatment. Deep learning-based causal inference models showed that vancomycin administration increases CKRT risk, identifying specific patient characteristics associated with higher susceptibility.IMPORTANCEThis study assesses the impact of vancomycin on the risk of continuous kidney replacement therapy (CKRT) in intensive care unit (ICU) patients with blood culture-positive infections. Utilizing deep learning-based causal inference and machine learning models, the research quantifies how vancomycin administration increases CKRT risk by an average of 7.7%. Key variables influencing susceptibility include baseline creatinine, diastolic blood pressure, heart rate, and bicarbonate levels. These findings offer insights into managing vancomycin-induced kidney risk and may inform patient-specific treatment strategies in ICU settings.

Automated segmentation of dental restorations using deep learning: exploring data augmentation techniques.

Deep learning has revolutionized image analysis for dentistry. Automated segmentation of dental radiographs is of great importance towards digital dentistry. The performance of deep learning models heavily relies on the quality and diversity of the training data. Data augmentation is a widely used technique implemented in machine learning and deep learning to artificially increase the size and diversity of a training dataset by applying various transformations to the original data. This work aims to automatically segment implants, prostheses, and fillings in panoramic images using 9 different deep learning segmentation models. Later, it explores the effect of data augmentation methods on segmentation performance of the models. Eight different data augmentation techniques are examined. Performance is evaluated by well-accepted metrics such as intersection over union (IoU) and Dice coefficient. While averaging the segmentation results for the three classes, IoU varies between 0.62 and 0.82 while Dice score is between 0.75 and 0.9 among deep learning models used. Augmentation techniques provided performance improvements of up to 3.37%, 5.75% and 8.75% for implant, prosthesis and filling classes, respectively. Findings reveal that choosing optimal augmentation strategies depends on both model architecture and dental structure type.

Reinforcement Learning for Intra- & Inter-Node Recommender Data Pipeline Optimization

Deep learning-based recommender models (DLRMs) have become an essential component of many modern recommender systems. Several companies are now building large compute clusters reserved for DLRM training, driving new interest in cost- & time- saving optimizations. The systems challenges faced in this setting are unique; while typical deep learning (DL) training jobs are dominated by model execution times, the most important factor in DLRM training performance is often online data ingestion. In this paper, we study real-world DLRM data processing pipelines taken from our compute cluster at Netflix to observe the performance impacts of online ingestion and identify shortfalls in existing data pipeline optimizers. Our studies lead us to design a new solution for data pipeline optimization, InTuneX . InTuneX is designed for production-scale multi-node recommender data pipelines. It unifies & tackles the challenges of both intra- and inter- node pipeline optimization. We achieve this with a multi-agent reinforcement learning (RL) design, simultaneously optimizing node assignments at the cluster level & CPU assignments within nodes. Our experiments show that InTuneX can build optimized data pipeline configurations within minutes. We apply InTuneX to our cluster, and find that it increases single-node data ingestion throughput by as much as 2.29X versus state-of-the-art optimizers, while improving the cost-efficiency of multi-node pipelines by 15-25%.

Integrating AI and Deep Learning for Efficient Drug Discovery and Target Identification

Artificial intelligence (AI) shows great potential in medical diagnosis and drug analysis. Through deep learning technology, AI can automatically extract features from large amounts of complex medical data, significantly improving diagnostic accuracy and drug development efficiency. Especially in drug target discovery and antiviral peptide classification, AI technology can accelerate data processing and prediction, helping researchers identify potential therapeutic molecules more quickly and optimize the drug development process. This study proposes and validates a Deep learn-based model, deep-Avpiden, to improve the classification and discovery efficiency of antiviral peptides (AVPs). By using sequential convolutional networks (TCNs), the model outperforms traditional recurrent neural networks (RNNs) and long short-term memory networks (LSTMs) in capturing long-term dependencies and parallel computing capabilities. In terms of dataset, we used AVPs and non-AVPS samples from multiple databases, totaling 5,414 cleaned and de-weighted peptide sequences for model training after data preprocessing and embedding. The Deep-AVPiden model has been shown to outperform existing advanced classifiers in experiments, and its effectiveness has been verified by accuracy, precision, recall rate, and area under the ROC curve (OC-ROC). In addition, to accommodate computing resource constraints, we propose an optimized version of Deep-AvPIDen (DS), which utilizes Deep separation convolution technology to significantly reduce computing resource consumption. Through the online application platform, researchers can efficiently classify antiviral proteins and discover new AVPs. Future research could further optimize the model's computational efficiency, handle larger data sets, and expand its potential for biomedical problems such as drug combination prediction and new drug discovery.

Research on a Passenger Flow Prediction Model Based on BWO-TCLS-Self-Attention

In recent years, with the rapid development of the global demand and scale for deep underground space utilization, deep space has gradually transitioned from single-purpose uses such as underground transportation, civil defense, and commerce to a comprehensive, livable, and disaster-resistant underground ecosystem. This shift has brought increasing attention to the safety of personnel flow in deep spaces. In addressing challenges in deep space passenger flow prediction, such as irregular flow patterns, surges in extreme conditions, large data dimensions, and redundant features complicating the model, this paper proposes a deep space passenger flow prediction model that integrates a Temporal Convolutional Network (TCN) and Long Short-Term Memory (LSTM) network. The model first employs a dual-layer LSTM network structure with a Dropout layer to capture complex temporal dynamics while preventing overfitting. Then, a Self-Attention mechanism and TCN network are introduced to reduce redundant feature data and enhance the model’s performance and speed. Finally, the Beluga Whale Optimization (BWO) algorithm is used to optimize hyperparameters, further improving the prediction accuracy of the network. Experimental results demonstrate that the BWO-TCLS-Self-Attention model proposed in this paper achieves an R2 value of 96.94%, with MAE and RMSE values of 118.464 and 218.118, respectively. Compared with some mainstream prediction models, the R2 value has increased, while both MAE and RMSE values have decreased, indicating its ability to accurately predict passenger flow in deep underground spaces.

Recognition of Conus species using a combined approach of supervised learning and deep learning-based feature extraction.

Cone snails are venomous marine gastropods comprising more than 950 species widely distributed across different habitats. Their conical shells are remarkably similar to those of other invertebrates in terms of color, pattern, and size. For these reasons, assigning taxonomic signatures to cone snail shells is a challenging task. In this report, we propose an ensemble learning strategy based on the combination of Random Forest (RF) and XGBoost (XGB) methods. We used 47,600 cone shell images of uniform size (224 x 224 pixels), which were split into an 80:20 train-test ratio. Prior to performing subsequent operations, these images were subjected to pre-processing and transformation. After applying a deep learning approach (Visual Geometry Group with a 16-layer deep model architecture) for feature extraction, model specificity was further assessed by including multiple related and unrelated seashell images. Both classifiers demonstrated comparable recognition ability on random test samples. The evaluation results suggested that RF outperformed XGB due to its high accuracy in recognizing Conus species, with an average precision of 95.78%. The area under the receiver operating characteristic curve was 0.99, indicating the model's optimal performance. The learning and validation curves also demonstrated a robust fit, with the training score reaching 1 and the validation score gradually increasing to 95 as more data was provided. These values indicate a well-trained model that generalizes effectively to validation data without significant overfitting. The gradual improvement in the validation score curve is crucial for ensuring model reliability and minimizing the risk of overfitting. Our findings revealed an interactive visualization. The performance of our proposed model suggests its potential for use with datasets of other mollusks, and optimal results may be achieved for their categorization and taxonomical characterization.

Generative Architectural Design: A Deep Learning Approach for Automated Space and Infrastructure Planning

Architectural design generation is a complex and time-intensive process requiring significant expertise and creativity. This research introduces a novel deep learning-based generative model that automates architectural design for rooms, buildings, plots, highways, and flyovers. The proposed model takes structured inputs such as size, area, location, and specific user prompts, including room dimensions, to produce customized architectural layouts. Leveraging a diverse dataset of architectural images and modern machine learning techniques, the model ensures adaptability and precision in design generation. This work highlights the effectiveness of integrating deep learning into architecture to streamline workflows, enhance creativity, and support urban and infrastructure planning

Diffusion‐Based Causality‐Preserving Neural Network for Dementia Recognition

ABSTRACTAnalyzing large‐scale functional brain networks for brain disorders often relies on undirected correlations in activation signals between brain regions. While focusing on co‐occurring activations, this approach overlooks the potential for directionality inherent in brain connectivity. Established research indicates the causal nature of brain networks, suggesting that activation patterns co‐occur and potentially influence one another. To this end, we propose a novel dffusion vector auto‐regressive (Diff‐VAR) method, enabling the assessment of whole‐brain effective connectivity (EC) as a directed and weighted network by integrating the search objectives into the deep neural network model as learnable parameters. The EC learned by our method identifies widespread differences in flow of influence within the brain network for individuals with impaired brain function compared to those with normal brain function. Moreover, we introduce an adaptive smoothing mechanism to enhance the stability and reliability of inferred EC. We evaluated the results of our proposed method on the Alzheimer's Disease Neuroimaging Initiative (ADNI) database. The model's performance is compared with existing correlation‐based and causality‐based methods. The results revealed that the brain networks constructed by our method achieve high classification accuracy and exhibit features consistent with physiological mechanisms. The code is available at https://github.com/SaqibMamoon/Diff‐VAR.

PanicleDet: a deep learning-based model for detection of panicle stages in paddy

PanicleDet: a deep learning-based model for detection of panicle stages in paddy

Adapting SAM2 Model from Natural Images for Tooth Segmentation in Dental Panoramic X-Ray Images

Dental panoramic X-ray imaging, due to its high cost-effectiveness and low radiation dose, has become a widely used diagnostic tool in dentistry. Accurate tooth segmentation is crucial for lesion analysis and treatment planning, helping dentists to quickly and precisely assess the condition of teeth. However, dental X-ray images often suffer from noise, low contrast, and overlapping anatomical structures, coupled with limited available datasets, leading traditional deep learning models to experience overfitting, which affects generalization ability. In addition, high-precision deep models typically require significant computational resources for inference, making deployment in real-world applications challenging. To address these challenges, this paper proposes a tooth segmentation method based on the pre-trained SAM2 model. We employ adapter modules to fine-tune the SAM2 model and introduce ScConv modules and gated attention mechanisms to enhance the model’s semantic understanding and multi-scale feature extraction capabilities for medical images. In terms of efficiency, we utilize knowledge distillation, using the fine-tuned SAM2 model as the teacher model for distilling knowledge to a smaller model named LightUNet. Experimental results on the UFBA-UESC dataset show that, in terms of performance, our model significantly outperforms the traditional UNet model in multiple metrics such as IoU, effectively improving segmentation accuracy and model robustness, particularly with limited sample datasets. In terms of efficiency, LightUNet achieves comparable performance to UNet, but with only 1.6% of its parameters and 24.0% of the inference time, demonstrating its feasibility for deployment on edge devices.

Physics-guided deep learning for skillful wind-wave modeling.

Modeling sea surface wind-waves is crucial for both scientific research and engineering applications. Nowadays, the most accurate wave models are based on numerical methods, which primarily concern the wave spectrum evolution by solving wave action balance partial differential equations. These methods are computationally expensive and limited by incomplete physical representations of wave spectral evolution. Here, we present a deep learning-based wave model trained using observation-merged wave hindcasts. Guided by the physics knowledge that waves are either generated by local current winds or by remote historical winds, this method can directly model significant wave height, bypassing the need for wave spectral information. This feature engineering effectively reduces the complexity of model inputs and outputs. The resulting artificial intelligence method can model 1 year of global significant wave heights at a 0.5° × 0.5° × 1-hour resolution within half an hour on a personal computer, achieving higher accuracy than state-of-the-art numerical wave models.

EXACT-Net: Framework for EHR-Guided Lung Tumor Auto-Segmentation for Non-Small Cell Lung Cancer Radiotherapy

Background/Objectives: Lung cancer is a devastating disease with the highest mortality rate among cancer types. Over 60% of non-small cell lung cancer (NSCLC) patients, accounting for 87% of lung cancer diagnoses, require radiation therapy. Rapid treatment initiation significantly increases the patient’s survival rate and reduces the mortality rate. Accurate tumor segmentation is a critical step in diagnosing and treating NSCLC. Manual segmentation is time- and labor-consuming and causes delays in treatment initiation. Although many lung nodule detection methods, including deep learning-based models, have been proposed. Most of these methods still have a long-standing problem of high false positives (FPs). Methods: Here, we developed an electronic health record (EHR)-guided lung tumor auto-segmentation called EXACT-Net (EHR-enhanced eXACtitude in Tumor segmentation), where the extracted information from EHRs using a pre-trained large language model (LLM) was used to remove the FPs and keep the TP nodules only. Results: The auto-segmentation model was trained on NSCLC patients’ computed tomography (CT), and the pre-trained LLM was used with the zero-shot learning approach. Our approach resulted in a 250% boost in successful nodule detection using the data from ten NSCLC patients treated in our institution. Conclusions: We demonstrated that combining vision-language information in EXACT-Net multi-modal AI framework greatly enhances the performance of vision only models, paving the road to multimodal AI framework for medical image processing.

Self-supervised learning improves robustness of deep learning lung tumor segmentation models to CT imaging differences.

Self-supervised learning (SSL) is an approach to extract useful feature representations from unlabeled data, and enable fine-tuning on downstream tasks with limited labeled examples. Self-pretraining is a SSL approach that uses curated downstream task dataset for both pretraining and fine-tuning. Availability of large, diverse, and uncurated public medical image sets presents the opportunity to potentially create foundation models by applying SSL in the "wild" that are robust to imaging variations. However, the benefit of wild- versus self-pretraining has not been studied for medical imageanalysis. Compare robustness of wild versus self-pretrained models created using convolutional neural network (CNN) and transformer (vision transformer [ViT] and hierarchical shifted window [Swin]) models for non-small cell lung cancer (NSCLC) segmentation from 3D computed tomography (CT) scans. CNN, ViT, and Swin models were wild-pretrained using unlabeled 10,412 3D CTs sourced from the cancer imaging archive and internal datasets. Self-pretraining was applied to same networks using a curated public downstream task dataset (n = 377) of patients with NSCLC. Pretext tasks introduced in self-distilled masked image transformer were used for both pretraining approaches. All models were fine-tuned to segment NSCLC (n = 377 training dataset) and tested on two separate datasets containing early (public n = 156) and advanced stage (internal n = 196) NSCLC. Models were evaluated in terms of: (a) accuracy, (b) robustness to image differences from contrast, slice thickness, and reconstruction kernels, and (c) impact of pretext tasks for pretraining. Feature reuse was evaluated using centered kernel alignment. Wild-pretrained Swin models resulted in higher feature reuse at earlier level layers and increased feature differentiation close to output. Wild-pretrained Swin outperformed self-pretrained models for analyzed imaging acquisitions. Neither ViT nor CNN showed a clear benefit of wild-pretraining compared to self-pretraining. Masked image prediction pretext task that forces networks to learn the local structure resulted in higher accuracy compared to contrastive task that models global image information. Wild-pretrained Swin networks were more robust to analyzed CT imaging differences for lung tumor segmentation than self-pretrained methods. ViT and CNN models did not show a clear benefit for wild-pretraining overself-pretraining.

Generating Synthetic Raman Spectra of DMMP and 2-CEES by Mathematical Transforms and Deep Generative Models

To build an automated system detecting toxic chemicals from Raman spectra, we have to obtain sufficient data of toxic chemicals. However, it usually costs high to gather Raman spectra of toxic chemicals in diverse situations. Tackling this problem, we develop methods to generate synthetic Raman spectra of DMMP and 2-CEES without actual experiments. First, we propose certain mathematical transforms to augment few original Raman spectra. Then, we train deep generative models to generate more realistic and diverse data. Analyzing synthetic Raman spectra of toxic chemicals generated by our methods through visualization, we qualitatively verify that the data are sufficiently similar to original data and diverse. For conclusion, we obtain a synthetic dataset of DMMP and 2-CEES with the proposed algorithm.

An enhanced deep intelligent model with feature fusion and ensemble learning for the fault diagnosis of rotating machinery

Vibration signals, serving as critical sources of information for monitoring the status of rotating machinery, demand effective extraction and rational utilization of its features to significantly enhance the accuracy and reliability of fault diagnosis. However, vibration signal features typically manifest as nonlinear and nonstationary, posing a significant challenge in industrial settings. To tackle this challenge, this article proposes an enhanced deep intelligent model based on feature fusion and ensemble learning for practical fault diagnosis of rotating machinery. First, a parallel network structure is introduced to comprehensively and accurately explore the fault characteristics of rotating machinery. This network comprises two branches: the first branch designs an improved one-dimensional convolutional neural network to extract locally robust features from raw signals; the second branch adopts variational mode decomposition to decompose raw signals into a set of intrinsic mode functions and extract comprehensive statistical features in both the time and frequency domains, significantly enhancing the signal representation capability. Subsequently, a deep neural network is used to extract more stable feature information. The features from the two branches are then fused, and the final network output is generated through a softmax regression function. Finally, ensemble learning uses a majority voting scheme to obtain more stable final outputs. To confirm the effectiveness of the proposed method, experiments are conducted on two laboratory cases and one industrial case. The experimental results demonstrate that the proposed method significantly improves fault diagnosis accuracy and reliability in controlled laboratory environments and real-world industrial applications, making it highly applicable for real-time monitoring and predictive maintenance of industrial machinery. These improvements can reduce maintenance costs and downtime, thus enhancing operational efficiency in various industrial settings.

A Deep Learning-based Predictive Analytics Model for Remote Patient Monitoring and Early Intervention in Diabetes Care

A Deep Learning-based Predictive Analytics Model for Remote Patient Monitoring and Early Intervention in Diabetes Care

A comprehensive analysis and performance evaluation for osteoporosis prediction models

Medical data analysis is an expanding area of study that holds the promise of transforming the healthcare landscape. The use of available data by researchers gives guidelines to improve health practitioners’ decision-making capacity, thus enhancing patients’ lives. The study looks at using deep learning techniques to predict the onset of osteoporosis from the NHANES 2017–2020 dataset that was preprocessed and arranged into SpineOsteo and FemurOsteo datasets. Two feature selection methods, namely mutual information (MI) and recursive feature elimination (RFE), were applied to sequential deep neural network models, convolutional neural network models, and recurrent neural network models. It can be concluded from the models that the mutual information method achieved higher accuracy than recursive feature elimination, and the MI feature selection CNN model showed better performance by showing 99.15% accuracy for the SpineOsteo dataset and 99.94% classification accuracy for the FemurOsteo dataset. Key findings of this study include family medical history, cases of fractures in patients and parental hip fractures, and regular use of medications like prednisone or cortisone. The research underscores the potential for deep learning in medical data processing, which eventually opens the way for enhanced models for diagnosis and prognosis based on non-image medical data. The implications of the study shall then be important for healthcare providers to be more informed in their decision-making processes for patients’ outcomes.

A Robust Malaria Cell Detection Framework Using Adaptive and Atrous Convolution-Based Recurrent Mobilenetv2 with Trans-MobileUNet + + -Based Abnormality Segmentation.

The highly contagious malaria disease is spread by the female Anopheles mosquito. This disease results in a patient's death or incapacity to move their muscles, if it is not appropriately identified in the early stages. A Rapid Diagnostic Test (RDT) is a frequently used approach to find malaria cells in red blood cells. However, it might not be able to identify infections with small amounts of samples. In the microscopic detection model, blood stains are placed under a microscope for diagnosing malaria. But accurate diagnosis is hard in this method, particularly in developing nations where the disease is most common. The microscopic detection processes are expensive and time-consuming due to the usage of microscopes. The quality of the blood smears and the availability of a qualified specialist, who is skilled in recognizing the disease, impact the accuracy of malaria detection results. The traditional deep learning-based malaria identification models need more processing power. Therefore, a deep learning-based adaptive method is designed to detect malaria cells through the medical image. Hence, the images are gathered from the standard sites and then fed to the segmentation process. Here, the abnormality segmentation is carried out with the help of a developed Trans-MobileUNet + + (T-MUnet + +) network. Trans-MobileUNet + + captures global context, so it is well-suited for segmentation tasks. The segmented image is applied to the adaptive detection phase where the Adaptive and Atrous Convolution-based Recurrent MobilenetV2 (AA-CRMV2) model is designed for the effective recognition of malaria cells. The efficiency of the designed approach is elevated by optimizing the parameters from the AA-CRMV2 network with the help of the Updated Random Parameter-based Fennec Fox Optimization (URP-FFO) algorithm. Several experimental analyses are evaluated in the implemented model over classical techniques to display their effectualness rate.

DEEP Learning-based air quality monitoring model via BM-KMC using seasonal images

DEEP Learning-based air quality monitoring model via BM-KMC using seasonal images

The CWA Benchmark: A Seismic Dataset from Taiwan for Seismic Research

Abstract In recent years, numerous studies have employed deep learning in seismology, with data-driven neural network models increasingly becoming the norm in emerging research paradigms. Alongside advancements in model architectures, several benchmark datasets have also been introduced. However, many of these datasets suffer from imbalanced distributions, particularly a scarcity of large-magnitude events, which can impair the performance of deep learning-based models. Taiwan, situated in a seismically active region experiences a high frequency of earthquakes annually. The Central Weather Administration (CWA) in Taiwan maintains comprehensive records of these seismic events. This study introduces a benchmark dataset, named CWA, specifically compiled from this extensive database. The CWA benchmark includes 331 events with magnitudes greater than five, collected from a high-density seismic network spanning from 2011 to 2021, making it well-suited for training deep learning models. In addition, the CWA benchmark features over 40 attributes and ∼500,000 seismograms, providing valuable data labels for various seismology-related tasks.