High Class Imbalance Research Articles

StAlK: Structural Alignment based Self Knowledge distillation for Medical Image Classification

In the realm of medical image analysis, where challenges like high class imbalance, inter-class similarity, and intra-class variance are prevalent, knowledge distillation has emerged as a powerful mechanism for model compression and regularization. Existing methodologies, including label smoothening, contrastive learning, and relational knowledge transfer, aim to address these challenges but exhibit limitations in effectively managing either class imbalance or intricate inter and intra-class relations within input samples. In response, this paper introduces StAlK (Structural Alignment based Self Knowledge distillation) for Medical Image Classification, a novel approach which leverages the alignment of complex high-order discriminative features from a mean teacher model. This alignment enhances the student model’s ability to distinguish examples across different classes. StAlK demonstrates superior performance in scenarios involving both inter and intra-class relationships and proves significantly more robust in handling class imbalance compared to baseline methods. Extensive investigations across multiple benchmark datasets reveal that StAlK achieves a substantial improvement of 6%–7% in top-1 accuracy compared to various state-of-the-art baselines. The code is available at: https://github.com/philsaurabh/StAlK_KBS.

An ensemble machine learning-based approach to predict thyroid disease using hybrid feature selection

Thyroid issues are becoming more and more common, and early detection is critical for therapy that reduces mortality and complications. Because of these factors, detecting thyroid problems has become more crucial in the medical field. Estimating the course of a disease accurately and understanding how clinical features interact are critical for medical diagnosis and treatment. All of these limits are overcome in our study by using a standard machine-learning model with proper clinical feature analysis and an ensemble-learning technique. Predicting sickness progression and the interdependence of clinical features or aspects are critical in medical diagnosis and therapy. However, machine learning has enabled us to detect the risk factors for this sickness. To select the best thyroid prediction outcome, we used five machine learning models in addition to the Ensemble ML classifier (hard voting). Class balancing approaches greatly increase classification performance. It has been demonstrated that using random oversampling improves classification results dramatically. Based on the experimental data, our suggested model outperforms existing methods by a wide margin. Using the XGBoost and SelectKBest feature selection strategies, the Ensemble ML classifier achieves the best results on hard voting on RF and DT, with 100 % sensitivity and 99.71 % accuracy. When features are decreased and the issue of high-class imbalance is addressed, the ensemble ML classifier (hard voting) performs better in dealing with classification challenges.

Machine learning based stellar classification with highly sparse photometry data.

Identifying stars belonging to different classes is vital in order to build up statistical samples of different phases and pathways of stellar evolution. In the era of surveys covering billions of stars, an automated method of identifying these classes becomes necessary. Many classes of stars are identified based on their emitted spectra. In this paper, we use a combination of the multi-class multi-label Machine Learning (ML) method XGBoost and the PySSED spectral-energy-distribution fitting algorithm to classify stars into nine different classes, based on their photometric data. The classifier is trained on subsets of the SIMBAD database. Particular challenges are the very high sparsity (large fraction of missing values) of the underlying data as well as the high class imbalance. We discuss the different variables available, such as photometric measurements on the one hand, and indirect predictors such as Galactic position on the other hand. We show the difference in performance when excluding certain variables, and discuss in which contexts which of the variables should be used. Finally, we show that increasing the number of samples of a particular type of star significantly increases the performance of the model for that particular type, while having little to no impact on other types. The accuracy of the main classifier is ∼0.7 with a macro F1 score of 0.61. While the current accuracy of the classifier is not high enough to be reliably used in stellar classification, this work is an initial proof of feasibility for using ML to classify stars based on photometry.

An ensemble machine learning-based approach to predict cervical cancer using hybrid feature selection

Cervical cancer has recently emerged as the leading cause of premature death among women. Around 85% of cervical cancer cases occur in underdeveloped countries. There are several risk factors associated with cervical cancer. This study describes a novel predictive model that uses early screening and risk trends from individual health records to forecast cervical cancer patients' prognoses. This study uses machine learning classification techniques to investigate the risk factors for cervical cancer. Additionally, use the voting method to evaluate all models and select the most appropriate model. The dataset used in this study contains missing values and shows a significant imbalance. Thus, the Random Oversampling technique was used as a sampling method. We used Principal Component Analysis (PCA) and XGBoost feature selection techniques to determine the most important features. To predict the accuracy, we used several machine learning classifiers, including Support Vector Machines (SVM), Random Forest (RF), k-nearest Neighbors (KNN), Decision Trees (DT), Naive Bayes (NB), Logistic Regression (LR), AdaBoost (AdB), Gradient Boosting (GB), Multilayer Perceptron (MLP), and Nearest Centroid Classifier (NCC). To demonstrate the efficacy of the suggested model, a comparison of its accuracy, sensitivity, and specificity was performed. We used the Random Oversampling approach along with the Ensemble ML method, hard voting on RF and MLP, and achieved 99.19% accuracy. It is demonstrated that the ensemble ML classifier (hard voting) performs better at handling classification problems when features are decreased and the high-class imbalance problem is handled.

Entropy‐based hybrid sampling (EHS) method to handle class overlap in highly imbalanced dataset

AbstractClass imbalance and class overlap create difficulties in the training phase of the standard machine learning algorithm. Its performance is not well in minority classes, especially when there is a high class imbalance and significant class overlap. Recently it has been observed by researchers that, the joint effects of class overlap and imbalance are more harmful as compared to their direct impact. To handle these problems, many methods have been proposed by researchers in past years that can be broadly categorized as data‐level, algorithm‐level, ensemble learning, and hybrid methods. Existing data‐level methods often suffer from problems like information loss and overfitting. To overcome these problems, we introduce a novel entropy‐based hybrid sampling (EHS) method to handle class overlap in highly imbalanced datasets. The EHS eliminates less informative majority instances from the overlap region during the undersampling phase and regenerates high informative synthetic minority instances in the oversampling phase near the borderline. The proposed EHS achieved significant improvement in F1‐score, G‐mean, and AUC performance metrics value by DT, NB, and SVM classifiers as compared to well‐established state‐of‐the‐art methods. Classifiers performances are tested on 28 datasets with extreme ranges in imbalance and overlap.

Multivariate data binning and examples generation to build a Diabetic Retinopathy classifier based on temporal clinical and analytical risk factors

In this paper, we explore the possibility of exploiting retrospective clinical data from Electronic Health Records (EHR) for classification tasks in chronic patients. The different intervals, short length and high class imbalance make it unfeasible to use traditional time series techniques. The first contribution of the paper is a preprocessing method to construct a multivariate time series dataset using EHR data, which infers missing data and regularizes the data frequency. The second contribution addresses class imbalance by using domain knowledge and existing short EHR series. We synthetically extrapolate patients’ data by using similar long time series and a fuzzy-based approach. The paper addresses the problem of detection of Diabetic Retinopathy (DR). Expert domain knowledge from ophthalmologists has been used in the proposed techniques to guide the processing of time series. The novelty in that case study consists in not using eye-fundus image analysis. Instead, the proposed methods are based solely on EHR data. Several multivariate multiclass time series classifiers are used to detect the four levels of DR severity from the pre-processed data sequences. Experiments prove the quality of the sequence preprocessing techniques proposed for EHR data. Results indicate that the TapNet classifier is the best one for DR grading. Despite being tested for DR detection, the proposed data preparation methods are applicable to other diseases with similar characteristics.

Improving Credit Card Fraud Detection with Data Reduction Approaches

Detecting fraudulent activities in credit card transactions can be challenging due to issues like high dimensionality and class imbalance that are often present in the datasets. To address these challenges, data reduction techniques such as data sampling and feature selection have become essential. In this study, we compare four approaches for data reduction: using data sampling alone, employing feature selection alone, applying data sampling followed by feature selection, and using feature selection followed by data sampling. Additionally, we include results using all features. We build classification models using five Decision Tree-based classifiers and Logistic Regression, and evaluate their performance using two performance metrics: the Area Under the Receiver Operating Characteristic Curve (AUC) and the Area under the Precision–Recall Curve (AUPRC). In this work, we adopt ensemble supervised feature selection (SFS) techniques and Random Undersampling (RUS) for data reduction. The experimental results demonstrate that all four data reduction techniques have the potential to improve the performance of classifiers. These results are valuable since the classifiers available are dependent upon application domains, computing environments, and licensing agreements. However, these techniques can be applied independently of all these dependencies. We recommend utilizing the ensemble SFS followed by RUS (SFS–RUS) approach as the preferred data reduction method due to its ability to run feature selection and data sampling in parallel. Additionally, we find that XGBoost and CatBoost outperform other classifiers.

Model-data-driven adversarial active learning for brain tumor segmentation

Active learning (AL) attempts to select informative samples in a dataset to minimize the number of required labels while maximizing the performance of the model. Current AL in segmentation tasks is limited to the expansion of popular classification-based methods including entropy, MC-dropout, etc. Meanwhile, most applications in the medical field are simply migrations that fail to consider the nature of medical images, such as high class imbalance, high domain difference, and data scarcity. In this study, we address these challenges and propose a novel AL framework for medical image segmentation task. Our approach introduces a pseudo-label-based filter addressing excessive blank patches in medical abnormalities segmentation tasks, e.g., lesions, and tumors, used before the AL selection. This filter helps reduce resource usage and allows the model to focus on selecting more informative samples. For the sample selection, we propose a novel query strategy that combines both model impact and data stability by employing adversarial attack. Furthermore, we harness the adversarial samples generated during the query process to enhance the robustness of the model. The experimental results verify our framework’s effectiveness over various state-of-the-art methods. Our proposed method only needs less than 14% annotated patches in 3D brain MRI multiple sclerosis (MS) segmentation tasks and 20% for Low-Grade Glioma (LGG) tumor segmentation to achieve competitive results with full supervision. These promising outcomes not only improve performance but alleviate the time burden associated with expert annotation, thereby facilitating further advancements in the field of medical image segmentation. Our code is available at https://github.com/HelenMa9998/adversarial_active_learning.

A multi-season machine learning approach to examine the training load and injury relationship in professional soccer

OBJECTIVES: The purpose of this study was to use machine learning to examine the relationship between training load and soccer injury with a multi-season dataset from one English Premier League club. METHODS: Participants were 35 male professional soccer players (aged 25.79±3.75 years, range 18–37 years; height 1.80±0.07 m, range 1.63–1.95 m; weight 80.70±6.78 kg, range 66.03–93.70 kg), with data collected from the 2014–2015 season until the 2018–2019 season. A total of 106 training loads variables (40 GPS data, 6 personal information, 14 physical data, 4 psychological data and 14 ACWR, 14 MSWR and 14 EWMA data) were examined in relation to 133 non-contact injuries, with a high imbalance ratio of 0.013. RESULTS: XGBoost and Artificial Neural Network were implemented to train the machine learning models using four and a half seasons’ data, with the developed models subsequently tested on the following half season’s data. During the first four and a half seasons, there were 341 injuries; during the next half season there were 37 injuries. To interpret and visualize the output of each model and the contribution of each feature (i.e., training load) towards the model, we used the Shapley Additive Explanations (SHAP) approach. Of 37 injuries, XGBoost correctly predicted 26 injuries, with recall and precision of 73% and 10% respectively. Artificial Neural Network correctly predicted 28 injuries, with recall and precision of 77% and 13% respectively. In the model using Artificial Neural Network (the relatively more accurate model), last injury area and weight appeared to be the most important features contributing to the prediction of injury. CONCLUSIONS: This was the first study of its kind to use Artificial Neural Network and a multi-season dataset for injury prediction. Our results demonstrate the potential to predict injuries with high recall, thereby identifying most of the injury cases, albeit, due to high class imbalance, precision suffered. This approach to using machine learning provides potentially valuable insights for soccer organizations and practitioners when monitoring load injuries.

Classification of high-dimensional imbalanced biomedical data based on spectral clustering SMOTE and marine predators algorithm

The research of biomedical data is crucial for disease diagnosis, health management, and medicine development. However, biomedical data are usually characterized by high dimensionality and class imbalance, which increase computational cost and affect the classification performance of minority class, making accurate classification difficult. In this paper, we propose a biomedical data classification method based on feature selection and data resampling. First, use the minimal-redundancy maximal-relevance (mRMR) method to select biomedical data features, reduce the feature dimension, reduce the computational cost, and improve the generalization ability; then, a new SMOTE oversampling method (Spectral-SMOTE) is proposed, which solves the noise sensitivity problem of SMOTE by an improved spectral clustering method; finally, the marine predators algorithm is improved using piecewise linear chaotic maps and random opposition-based learning strategy to improve the algorithm’s optimization seeking ability and convergence speed, and the key parameters of the spectral-SMOTE are optimized using the improved marine predators algorithm, which effectively improves the performance of the over-sampling approach. In this paper, five real biomedical datasets are selected to test and evaluate the proposed method using four classifiers, and three evaluation metrics are used to compare with seven data resampling methods. The experimental results show that the method effectively improves the classification performance of biomedical data. Statistical test results also show that the proposed PRMPA-Spectral-SMOTE method outperforms other data resampling methods.

LiFoL: An Efficient Framework for Financial Distress Prediction in High-Dimensional Unbalanced Scenario

Corporate financial distress will significantly damage the company’s and its stakeholders’ interests and even lead to a global financial crisis. Therefore, finding an efficient method for financial distress prediction (FDP) to avoid greater losses is essential. Although there is a lot of research and progress in this field, the existing methods rarely consider the problems of high dimensionality and class imbalance, which will largely limit the models to achieve satisfactory performance. To alleviate these problems, this article first proposes a novel Lightspace-SMOTE upsampling method, which can reduce the feature dimensionality and increase the signal-to-noise ratio (SNR) of the original data and then upsample it to increase the number of minor class samples. In addition, this article proposes an efficient ensemble framework (LiFoL) that combines Lightspace-SMOTE, focal loss (FL), and LightGBM, which can not only focus more on minor class and the hard-to-class samples but also obtain better performance. At the same time, the feature importance provided by the model can provide strong support for model interpretability. Experimental results show that the Lightspace-SMOTE upsampling method can help the model achieve higher scores in area under ROC curve (AUC) and recall, especially in the case of longer prediction periods. Compared with current methods, LiFoL can achieve more than 10% improvement in AUC and more than 20% in recall.

Eye-LRCN: A Long-Term Recurrent Convolutional Network for Eye Blink Completeness Detection.

Computer vision syndrome causes vision problems and discomfort mainly due to dry eye. Several studies show that dry eye in computer users is caused by a reduction in the blink rate and an increase in the prevalence of incomplete blinks. In this context, this article introduces Eye-LRCN, a new eye blink detection method that also evaluates the completeness of the blink. The method is based on a long-term recurrent convolutional network (LRCN), which combines a convolutional neural network (CNN) for feature extraction with a bidirectional recurrent neural network that performs sequence learning and classifies the blinks. A Siamese architecture is used during CNN training to overcome the high-class imbalance present in blink detection and the limited amount of data available to train blink detection models. The method was evaluated on three different tasks: blink detection, blink completeness detection, and eye state detection. We report superior performance to the state-of-the-art methods in blink detection and blink completeness detection, and remarkable results in eye state detection.

A Comparative Analysis of Combination of CNN-Based Models with Ensemble Learning on Imbalanced Data

This study investigates the usefulness of the Synthetic Minority Oversampling Technique (SMOTE) in conjunction with convolutional neural network (CNN) models, which include both single and ensemble classifiers. The objective of this research is to handle the difficulty of multi-class imbalanced image classification. The application of SMOTE in imbalanced picture datasets is still underexplored, even though CNNs have been shown to be successful in image classification and that ensemble learning approaches have improved their performance. To investigate whether or not SMOTE can increase classification accuracy and other performance measures when combined with CNN-based classifiers, our research makes use of a CIFAR-10 dataset that has been artificially step-imbalanced and has varying imbalanced ratios. We conducted experiments using five distinct models, namely AdaBoost, XGBoost, standalone CNN, CNN-AdaBoost, and CNN-XGBoost, on datasets that were either imbalanced or SMOTE-balanced. Metrics such as accuracy, precision, recall, F1-score, and the area under the receiver operating characteristic curve (AUC) were included in the evaluation process. The findings indicate that SMOTE dramatically improves the accuracy of minority classes, and that the combination of ensemble classifiers with CNNs and oversampling techniques significantly improves overall classification performance, particularly in situations when there is a high-class imbalance. When it comes to enhancing imbalanced classification tasks, this study demonstrates the potential of merging oversampling techniques with CNN-based ensemble classifiers to minimize the impacts of class imbalance in picture datasets. This suggests a promising direction for future research in this area.

Credit Card Fraud Detection Using State-of-the-Art Machine Learning and Deep Learning Algorithms

People can use credit cards for online transactions as it provides an efficient and easy-to-use facility. With the increase in usage of credit cards, the capacity of credit card misuse has also enhanced. Credit card frauds cause significant financial losses for both credit card holders and financial companies. In this research study, the main aim is to detect such frauds, including the accessibility of public data, high-class imbalance data, the changes in fraud nature, and high rates of false alarm. The relevant literature presents many machines learning based approaches for credit card detection, such as Extreme Learning Method, Decision Tree, Random Forest, Support Vector Machine, Logistic Regression and XG Boost. However, due to low accuracy, there is still a need to apply state of the art deep learning algorithms to reduce fraud losses. The main focus has been to apply the recent development of deep learning algorithms for this purpose. Comparative analysis of both machine learning and deep learning algorithms was performed to find efficient outcomes. The detailed empirical analysis is carried out using the European card benchmark dataset for fraud detection. A machine learning algorithm was first applied to the dataset, which improved the accuracy of detection of the frauds to some extent. Later, three architectures based on a convolutional neural network are applied to improve fraud detection performance. Further addition of layers further increased the accuracy of detection. A comprehensive empirical analysis has been carried out by applying variations in the number of hidden layers, epochs and applying the latest models. The evaluation of research work shows the improved results achieved, such as accuracy, f1-score, precision and AUC Curves having optimized values of 99.9%,85.71%,93%, and 98%, respectively. The proposed model outperforms the state-of-the-art machine learning and deep learning algorithms for credit card detection problems. In addition, we have performed experiments by balancing the data and applying deep learning algorithms to minimize the false negative rate. The proposed approaches can be implemented effectively for the real-world detection of credit card fraud

Synthesizing class labels for highly imbalanced credit card fraud detection data

Acquiring labeled datasets often incurs substantial costs primarily due to the requirement of expert human intervention to produce accurate and reliable class labels. In the modern data landscape, an overwhelming proportion of newly generated data is unlabeled. This paradigm is especially evident in domains such as fraud detection and datasets for credit card fraud detection. These types of data have their own difficulties associated with being highly class imbalanced, which poses its own challenges to machine learning and classification. Our research addresses these challenges by extensively evaluating a novel methodology for synthesizing class labels for highly imbalanced credit card fraud data. The methodology uses an autoencoder as its underlying learner to effectively learn from dataset features to produce an error metric for use in creating new binary class labels. The methodology aims to automatically produce new labels with minimal expert input. These class labels are then used to train supervised classifiers for fraud detection. Our empirical results show that the synthesized labels are of high enough quality to produce classifiers that significantly outperform a baseline learner comparison when using area under the precision-recall curve (AUPRC). We also present results of varying levels of positive-labeled instances and their effect on classifier performance. Results show that AUPRC performance improves as more instances are labeled positive and belong to the minority class. Our methodology thereby effectively addresses the concerns of high class imbalance in machine learning by creating new and effective class labels.

Machine learning based stellar classification with highly sparse photometry data

Background Identifying stars belonging to different classes is vital in order to build up statistical samples of different phases and pathways of stellar evolution. In the era of surveys covering billions of stars, an automated method of identifying these classes becomes necessary. Methods Many classes of stars are identified based on their emitted spectra. In this paper, we use a combination of multi-class multi-label methods from Machine Learning (ML) and the PySSED spectral-energy-distribution fitting algorithm to classify stars into nine different classes. The classifier is trained on subsets of the SIMBAD database. Particular challenges are the very high sparsity (large fraction of missing values) of the underlying data as well as the high class imbalance. We discuss the different variables available, such as photometric measurements on the one hand, and indirect predictors such as Galactic position on the other hand. Results We show the difference in performance when excluding certain variables, and discuss in which contexts which variables should be used. Finally, we show that increasing the number of samples of a particular type of star significantly increases the performance of the model for that particular type, while having little to no impact on other types. Conclusions While the current accuracy of the classifier is not high enough to be usable for non-experts, this work is an initial proof of feasibility for using ML to classify stars.

Attention-based early warning framework for abnormal operating conditions in fluid catalytic cracking units

Fluid Catalytic Cracking Unit (FCCU) is a critical processing technology in the oil refining industry, playing a vital role in energy efficiency and environmental protection. However, FCCU often encounters various abnormal operating conditions, leading to safety hazards, downtime, and reduced production efficiency. Early warning of these abnormal conditions is crucial but challenging due to high noise, strong hysteresis, and class imbalance problems. To tackle these challenges, a novel and universal attention-based framework called AEW-AOC (Attention-based Early Warning for Abnormal Operating Conditions) is specifically designed for FCCU applications. The proposed AEW-AOC framework incorporates three key components: (1) a Self-Correlation Denoiser (SCD) module is proposed to exploit spatiotemporal data correlation to effectively reduce noise; (2) a Convolutional Long Short-Term Memory (Conv-LSTM) module is employed to address the issue of strong hysteresis by capturing temporal variation features of process parameters; (3) an Anomaly Pattern Attention (APA) module is proposed to enhance the distinguishability of abnormal operating conditions based on clustering results from historical abnormal instances. Extensive experiments demonstrate the effectiveness and superiority of the proposed AEW-AOC framework, particularly in practical applications. Specifically, the AEW-AOC framework obtains an impressive fβ score of 91.00% on LIC201, 90.45% on LIC202, and 90.64% on LIC801. The proposed AEW-AOC framework shows great potential in enhancing safety, reducing downtime, optimizing efficiency, promoting sustainability, and expanding its applicability beyond FCCU. Its proactive and versatile nature makes it a valuable tool for improving industrial processes and driving advancements in the field of abnormal operating condition detection and prevention.

Research on Intrusion Detection Based on an Enhanced Random Forest Algorithm

To address the challenges posed by high data dimensionality and class imbalance during intrusion detection, which result in increased computational complexity, resource consumption, and reduced classification accuracy, this paper presents an intrusion-detection algorithm based on an improved Random Forest approach. The algorithm employs the Bald Eagle Search (BES) optimization technique to fine-tune the Kernel Principal Component Analysis (KPCA) algorithm, enabling optimized dimensionality reduction. The processed data are then fed into a cost-sensitive Random Forest classifier for training, with subsequent model validation conducted on the reduced-dimension data. Experimental results demonstrate that compared to traditional Random Forest algorithms, the proposed method reduces the training time by 11.32 s and achieves a 5.59% increase in classification accuracy, an 11.7% improvement in specificity, and a 0.0558 increase in the G-mean value. These findings underscore the promising application potential and performance of this approach in the field of network intrusion detection.

Data augmentation strategy for generating realistic samples on defect segmentation task

In the anomaly and defect detection tasks, the number of negative samples greatly exceeds the number of defective samples. As a result, a high-class imbalance exists among different classes in the detection task. In our work, we introduce a data-level solution for improving the generalization performance of the semantic segmentation of surface defects based on a data augmentation (DA) strategy. In particular, our DA approach comprised a generative stage to simulate synthetic defects and a validation stage to validate the synthetic image as close as possible to the real one. A Siamese network fully validates our synthetic samples to select only synthetic defects as close to the real ones. We demonstrated the effectiveness of our approach in a real-use case scenario to baseline DA approaches. Our DA approach allows balancing the minority classes while improving the overall generalization performance for semantic segmentation for defect detection.

Resource efficient aortic distensibility calculation by end to end spatiotemporal learning of aortic lumen from multicentre multivendor multidisease CMR images

Aortic distensibility (AD) is important for the prognosis of multiple cardiovascular diseases. We propose a novel resource-efficient deep learning (DL) model, inspired by the bi-directional ConvLSTM U-Net with densely connected convolutions, to perform end-to-end hierarchical learning of the aorta from cine cardiovascular MRI towards streamlining AD quantification. Unlike current DL aortic segmentation approaches, our pipeline: (i) performs simultaneous spatio-temporal learning of the video input, (ii) combines the feature maps from the encoder and decoder using non-linear functions, and (iii) takes into account the high class imbalance. By using multi-centre multi-vendor data from a highly heterogeneous patient cohort, we demonstrate that the proposed method outperforms the state-of-the-art method in terms of accuracy and at the same time it consumes ∼\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$\\sim$$\\end{document} 3.9 times less fuel and generates ∼\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$\\sim$$\\end{document} 2.8 less carbon emissions. Our model could provide a valuable tool for exploring genome-wide associations of the AD with the cognitive performance in large-scale biomedical databases. By making energy usage and carbon emissions explicit, the presented work aligns with efforts to keep DL’s energy requirements and carbon cost in check. The improved resource efficiency of our pipeline might open up the more systematic DL-powered evaluation of the MRI-derived aortic stiffness.