Model Training Research Articles

Determining the optimal level of automation for cash flow forecasts

PurposeThe aim of this paper is to develop an approach for identifying the optimal level of automation by maximizing the level of automation and accuracy while addressing problem areas of forecast quality.Design/methodology/approachWe use a unique set of forecasts planned by six subsidiaries of a multinational corporation to train and test various models. We compare the accuracy of three levels of automation and how they address prevalent forecasting process quality problem areas.FindingsThe findings indicate that accuracy alone is not a sufficient dimension to consider when selecting the optimal level of automation but that forecast process quality areas need to be assessed as well.Research limitations/implicationsThe limitations of this work are the inability to study the effects of our tool’s recommendations, the sample originating from a single company, the use of simple statistical methods and the limited number of dimensions to evaluate forecasts.Practical implicationsFirms should apply the structure offered in this paper to target individual components of the cash flow forecasting process when automating it and use it to structure their discussion, planning and implementation of automation.Originality/valueA novel approach for determining the optimal level of automation for cash flow forecasting combining the human information processing framework of Parasuraman et al. (2000) with the forecast quality problem areas by Fildes and Petropoulos (2015).

Machine learning-based design of double corrugated steel plate shear walls

PurposeIn this study, machine learning (ML) algorithms were employed to predict the shear capacity and behavior of DCSWs.Design/methodology/approachIn this study, ML algorithms were employed to predict the shear capacity and behavior of DCSWs. Various ML techniques, including linear regression (LR), support vector machine (SVM), decision tree (DT), random forest (RF), extreme gradient boosting (XGBoost) and artificial neural network (ANN), were utilized. The ML models were trained using a dataset of 462 numerical and experimental samples. Numerical models were generated and analyzed using the finite element (FE) software Abaqus. These models underwent push-over analysis, subjecting them to pure shear conditions by applying a target displacement solely to the top of the shear walls without interaction from a frame. The input data encompassed eight survey variables: geometric values and material types. The characterization of input FE data was randomly generated within a logical range for each variable. The training and testing phases employed 90 and 10% of the data, respectively. The trained models predicted two output targets: the shear capacity of DCSWs and the likelihood of buckling. Accurate predictions in these areas contribute to the efficient lateral enhancement of structures. An ensemble method was employed to enhance capacity prediction accuracy, incorporating select algorithms.FindingsThe proposed model achieved a remarkable 98% R-score for estimating shear strength and a corresponding 98% accuracy in predicting buckling occurrences. Among all the algorithms tested, XGBoost demonstrated the best performance.Originality/valueIn this study, for the first time, ML algorithms were employed to predict the shear capacity and behavior of DCSWs.

Model of Dynamic Mechanical Parameters and Vibration Analysis in Hot Rolling

As the core equipment of metal shaping processing, the vibration problem of rolling mill seriously affects the product quality and production efficiency. To address the problem of the limited prediction of mill vibration due to small sample data in non-steady states, we propose to predict dynamic mechanical parameters (rolling force and rolling torque) based on the TCN-LSTM step strategy transfer model. The prediction accuracies of the TCN-LSTM step strategy transfer model under 1000 sets of training data reach 95.8% and 93.2%, respectively, and at the same time, it saves the time of training and regulation of the deep learning model and can better meet the needs of online prediction. The horizontal-torsion hot rolling vibration model is further established to quantitatively analyze the relationship between process parameters and mill vibration. The experimental results show that with the vibration suppression measures of 10% reduction of the rolling speed and 10% reduction of the entrance thickness, the horizontal vibration displacement amplitude is reduced from 0.687 × 10−4 m to 0.229 × 10−4 m, and the rotational displacement amplitude is reduced from 0.0273 m to 0.0117 m, which effectively suppresses the vibration of the mill and improves the stability of the rolling process.

Exploring the limits of hierarchical world models in reinforcement learning.

Hierarchical model-based reinforcement learning (HMBRL) aims to combine the sample efficiency of model-based reinforcement learning with the abstraction capability of hierarchical reinforcement learning. While HMBRL has great potential, the structural and conceptual complexities of current approaches make it challenging to extract general principles, hindering understanding and adaptation to new use cases, and thereby impeding the overall progress of the field. In this work we describe a novel HMBRL framework and evaluate it thoroughly. We construct hierarchical world models that simulate the environment at various levels of temporal abstraction. These models are used to train a stack of agents that communicate top-down by proposing goals to their subordinate agents. A significant focus of this study is the exploration of a static and environment agnostic temporal abstraction, which allows concurrent training of models and agents throughout the hierarchy. Unlike most goal-conditioned H(MB)RL approaches, it also leads to comparatively low dimensional abstract actions. Although our HMBRL approach did not outperform traditional methods in terms of final episode returns, it successfully facilitated decision-making across two levels of abstraction. A central challenge in enhancing our method's performance, as uncovered through comprehensive experimentation, is model exploitation on the abstract level of our world model stack. We provide an in depth examination of this issue, discussing its implications and suggesting directions for future research to overcome this challenge. By sharing these findings, we aim to contribute to the broader discourse on refining HMBRL methodologies.

Automated acute skin toxicity scoring in a mouse model through deep learning.

This study presents a novel approach to skin toxicity assessment in preclinical radiotherapy trials through an advanced imaging setup and deep learning. Skin reactions, commonly associated with undesirable side effects in radiotherapy, were meticulously evaluated in 160 mice across four studies. A comprehensive dataset containing 7542 images was derived from proton/electron trials with matched manual scoring of the acute toxicity on the right hind leg, which was the target area irradiated in the trials. This dataset was the foundation for the subsequent model training. The two-step deep learning framework incorporated an object detection model for hind leg detection and a classification model for toxicity classification. An observer study involving five experts and the deep learning model, was conducted to analyze the retrospective capabilities and inter-observer variations. The results revealed that the hind leg object detection model exhibited a robust performance, achieving an accuracy of almost 99%. Subsequently, the classification model demonstrated an overall accuracy of about 85%, revealing nuanced challenges in specific toxicity grades. The observer study highlighted high inter-observer agreement and showcased the model's superiority in accuracy and misclassification distance. In conclusion, this study signifies an advancement in objective and reproducible skin toxicity assessment. The imaging and deep learning system not only allows for retrospective toxicity scoring, but also presents a potential for minimizing inter-observer variation and evaluation times, addressing critical gaps in manual scoring methodologies. Future recommendations include refining the system through an expanded training dataset, paving the way for its deployment in preclinical research and radiotherapy trials.

Establishing an AI model and application for automated capsule endoscopy recognition based on convolutional neural networks (with video).

Although capsule endoscopy (CE) is a crucial tool for diagnosing small bowel diseases, the need to process a vast number of images imposes a significant workload on physicians, leading to a high risk of missed diagnoses. This study aims to develop an artificial intelligence (AI) model and application based on convolutional neural networks that can automatically recognize various lesions in small bowel capsule endoscopy. Three small bowel capsule endoscopy datasets were used for AI model training, validation, and testing, encompassing 12 categories of images. The model's performance was evaluated using metrics such as AUC, sensitivity, specificity, precision, accuracy, and F1 score to select the best model. A human-machine comparison experiment was conducted using the best model and endoscopists with varying levels of experience. Model interpretability was analyzed using Grad-CAM and SHAP techniques. Finally, a clinical application was developed based on the best model using PyQt5 technology. A total of 34,303 images were included in this study. The best model, MobileNetv3-large, achieved a weighted average sensitivity of 87.17%, specificity of 98.77%, and an AUC of 0.9897 across all categories. The application developed based on this model performed exceptionally well in comparison with endoscopists, achieving an accuracy of 87.17% and a processing speed of 75.04 frames per second, surpassing endoscopists of varying experience levels. The AI model and application developed based on convolutional neural networks can quickly and accurately identify 12 types of small bowel lesions. With its high sensitivity, this system can effectively assist physicians in interpreting small bowel capsule endoscopy images.Future studies will validate the AI system for video evaluations and real-world clinical integration.

Decoding Radiomics: A Step-by-Step Guide to Machine Learning Workflow in Hand-Crafted and Deep Learning Radiomics Studies

Although radiomics research has experienced rapid growth in recent years, with numerous studies dedicated to the automated extraction of diagnostic and prognostic information from various imaging modalities, such as CT, PET, and MRI, only a small fraction of these findings has successfully transitioned into clinical practice. This gap is primarily due to the significant methodological challenges involved in radiomics research, which emphasize the need for a rigorous evaluation of study quality. While many technical aspects may lie outside the expertise of most radiologists, having a foundational knowledge is essential for evaluating the quality of radiomics workflows and contributing, together with data scientists, to the development of models with a real-world clinical impact. This review is designed for the new generation of radiologists, who may not have specialized training in machine learning or radiomics, but will inevitably play a role in this evolving field. The paper has two primary objectives: first, to provide a clear, systematic guide to radiomics study pipeline, including study design, image preprocessing, feature selection, model training and validation, and performance evaluation. Furthermore, given the critical importance of evaluating the robustness of radiomics studies, this review offers a step-by-step guide to the application of the METhodological RadiomICs Score (METRICS, 2024)—a newly proposed tool for assessing the quality of radiomics studies. This roadmap aims to support researchers and reviewers alike, regardless of their machine learning expertise, in utilizing this tool for effective study evaluation.

Joint Mixing Data Augmentation for Skeleton-based Action Recognition

Skeleton-based action recognition is beneficial for understanding human behavior in videos, and thus has received much attention in recent years as an important research area in action recognition. Current research focuses on designing more advanced algorithms to better extract spatio-temporal information from skeleton data. However, due to the small amount of data in the existing skeleton dataset and the lack of effective data augmentation methods, it is easy to lead to overfitting in model training. To address this challenge, we propose a mix-based data augmentation method, Joint Mixing Data Augmentation (JMDA), which can generally improve the effectiveness and robustness of various skeleton-based action recognition algorithms. In terms of spatial information, we introduce SpatialMix (SM), a method that projects the original 3D skeleton discrete information into a 2D space. Then, SM mixes the projected spatial information between two random samples during the training process to achieve the spatial-based mixing data augmentation. Concerning temporal information, we propose TemporalMix (TM). Leveraging the temporal continuity in skeleton data, we perform a temporal resize operation on the original skeleton data, and then merge two random samples during training to achieve the temporal-based mixed data augmentation. Additionally, we analyze the Feature Mismatch (FM) problem caused by introducing mix-based data augmentation into skeleton data. Then we propose a new data preprocessing method called Feature Alignment (FA) to effectively address this problem and improve model performance. Moreover, we propose a novel training pipeline, Joint Training Strategy (JTS), which combines multiple mix-based data augmentation methods for further improvement of model performance. Specifically, our proposed JMDA is plug-and-play and widely applicable to skeleton-based action recognition models. At the same time, the application of JMDA does not increase the model parameters and there is almost no additional training cost. We conduct extensive experiments on NTU RGB+D 60 and NTU RGB+D 120 datasets to demonstrate the effectiveness and robustness of the proposed JMDA on several mainstream skeleton-based action recognition algorithms.

Medical image recognition based on multilayer neural network

As a common accidental injury, accurate identification of burns and scalded injuries is of great significance to the development of treatment programs and prognosis assessment. Traditional identification methods are subjective and inaccurate, this paper proposes a burns and scalds grade identification technique based on multilayer neural network. The three-degree classification standard of burns and scalds is described, and the structure and principle of multilayer neural network are introduced, including the network structure composed of input layer, hidden layer and output layer, forward propagation, back propagation and training process. The implementation steps of burns and scald grade recognition technology based on multi-layer neural network are discussed in detail, such as data acquisition, image pre-processing, feature extraction, classifier design, model training and evaluation. Through experiments, the model is tested using a dataset containing 362 burns and scalds images, and the model achieves more than 90% accuracy on the test set with high accuracy and reliability. The multi-layer neural network model is used to classify the burn and scald data, which improves the diagnostic accuracy, shortens the delay time of the disease, reduces the pressure of professional doctors, and helps patients have a preliminary understanding of the degree of burn.

Federated learning with differential privacy via fast Fourier transform for tighter-efficient combining.

Spurred by the simultaneous need for data privacy protection and data sharing, federated learning (FL) has been proposed. However, it still poses a risk of privacy leakage in it. This paper, an improved Differential Privacy (DP) algorithm to protect the federated learning model. Additionally, the Fast Fourier Transform (FFT) is used in the computation of the privacy budget [Formula: see text], to minimize the impact of limited arithmetic resources and numerous users on the effectiveness of training model. Moreover, instead of direct analyses of the privacy budget ε through various methods, Privacy Loss Distribution (PLD) and privacy curves are adopted, while the number of artificial assignments hyperparameters is reduced, and the grid parameters delineated for FFT use are improved. The improved algorithm tightens parameter bounds and minimizes human factors' influence with minimal efficiency impact. It decreases the errors caused by truncation and discreteness of PLDs while expanding the discreteness interval to reduce the calculation workload. Furthermore, an improved activation function using a temper sigmoid with only one parameter [Formula: see text], smooths the accuracy curve and mitigates drastically fluctuating scenarios during model training. Finally, simulation results on real datasets show that our improved DP algorithm, which accounts for long trailing, facilitates a better balance between privacy and utility in federated learning models.

Machine Learning-Based Prediction for Incident Hypertension Based on Regular Health Checkup Data: Derivation and Validation in 2 Independent Nationwide Cohorts in South Korea and Japan.

Worldwide, cardiovascular diseases are the primary cause of death, with hypertension as a key contributor. In 2019, cardiovascular diseases led to 17.9 million deaths, predicted to reach 23 million by 2030. This study presents a new method to predict hypertension using demographic data, using 6 machine learning models for enhanced reliability and applicability. The goal is to harness artificial intelligence for early and accurate hypertension diagnosis across diverse populations. Data from 2 national cohort studies, National Health Insurance Service-National Sample Cohort (South Korea, n=244,814), conducted between 2002 and 2013 were used to train and test machine learning models designed to anticipate incident hypertension within 5 years of a health checkup involving those aged ≥20 years, and Japanese Medical Data Center cohort (Japan, n=1,296,649) were used for extra validation. An ensemble from 6 diverse machine learning models was used to identify the 5 most salient features contributing to hypertension by presenting a feature importance analysis to confirm the contribution of each future. The Adaptive Boosting and logistic regression ensemble showed superior balanced accuracy (0.812, sensitivity 0.806, specificity 0.818, and area under the receiver operating characteristic curve 0.901). The 5 key hypertension indicators were age, diastolic blood pressure, BMI, systolic blood pressure, and fasting blood glucose. The Japanese Medical Data Center cohort dataset (extra validation set) corroborated these findings (balanced accuracy 0.741 and area under the receiver operating characteristic curve 0.824). The ensemble model was integrated into a public web portal for predicting hypertension onset based on health checkup data. Comparative evaluation of our machine learning models against classical statistical models across 2 distinct studies emphasized the former's enhanced stability, generalizability, and reproducibility in predicting hypertension onset.

Focal mechanism determination by location-constrained deep learning: application to microseismic monitoring

Accurate and rapid determination of focal mechanism solutions is of great significance for real-time seismic monitoring. Previous deep learning approaches for focal mechanism determination typically use the waveform data, which is sensitive to the velocity model and inherently includes information about both the source location and the focal mechanism. We introduce a location-constrained deep learning algorithm for determining the focal mechanism for surface microseismic events. By using aligned P-wave data along with azimuth and take-off angle as input, we narrow the solution space for the focal mechanism problem and reduce the dependence on the velocity model. The model is trained using 40,000 theoretical samples generated with the geometry and velocity model of the field data. Validation tests, comparisons with a waveform-based network, velocity perturbation tests, and location error tests are performed to demonstrate the robustness and efficiency of our method. After applying the trained model to field data, the results demonstrate that our method is fast and achieves comparable accuracy to HASH results for high-quality events, making it promising for real-time microseismic monitoring.

Development and validation of a prognostic model for cachexia in postoperative gastric cancer patients with low nutritional risk: a dual-center retrospective cohort study.

Gastric cancer can lead to excessive catabolism in patients. After undergoing gastric surgery, patients may experience additional unintended weight loss, resulting in severe malnutrition and potentially cachexia. We selected and incorporated patients from two centers. Cohort 1 (n = 1393) served as the development cohort, while cohort 2 (n = 501) was designated as an external validation cohort. Within cohort 1, 70% of the patients were utilized for model training, with the remaining 30% reserved for internal validation. The training set initially underwent univariate logistic regression, followed by multivariate logistic regression. The factors ultimately incorporated were used to construct the model and create nomograms. The discriminative ability was assessed using ROC curves in all three datasets, calibration curves were used to evaluate consistency, and decision curves analysis was performed to assess the clinical application value. The model incorporated 12 factors, specifically: age (OR = 1.07), preoperative BMI (OR = 0.89), surgery type (Total Gastrectomy (TG), OR = 1.83), chemotherapy (yes, OR = 1.52), stage (III, OR = 1.40), anastomotic obstruction (yes, OR = 6.85), Postsurgical Gastroparesis Syndrome (PGS) (yes, OR = 2.27), albumin (OR = 0.95), hemoglobin (OR = 0.98), triglycerides (OR = 0.36), CRP (OR = 1.07), and Neutrophil to Lymphocyte Ratio (NLR) (OR = 1.05). The model demonstrated robust performance in ROC with AUC values of 0.805 in the training set, 0.767 in the validation set, and 0.795 in Cohort 2. Calibration curves in all three datasets exhibited a high degree of concordance between actual and predicted probabilities. Decision curve analysis (DCA) indicated that the model holds substantial clinical utility across all three datasets. We have developed a predictive model for cachexia in patients undergoing gastric cancer surgery. This model enables healthcare professionals to accurately estimate the risk of cachexia in postoperative patients with nutritional deficits, allowing for timely nutritional interventions to enhance patient quality of life and prognosis.

Exploring the Nexus Between Fertility Rates and Geopolitical Risk with Intelligence Methods: A Multifaceted Analysis

Background/Objectives: This paper presents an analysis of birth rate statistics, specifically focusing on recorded births in Scotland. The main research objective focuses on investigating the influence of geopolitical concerns on birth rate forecasts. Specifically, we examine whether individuals may choose to postpone or abstain from having children during times of conflict or political turmoil due to concerns about personal safety, the welfare of their children, or uncertainty about the future caused by geopolitical risks. Additionally, this study examines how disruptions to healthcare services, such as limited access to prenatal care and maternal health facilities, can affect birth outcomes and lead to changes in birth rates. Methods: To approach the research objective both machine learning algorithms and classical statistical procedures. Also, as part of the current analysis, the Geopolitical Risk Index has been applied as an extra factor to predict the birth rate. Results: The results of our study demonstrate the effectiveness of machine learning in producing precise predictions in this field, while emphasizing the significant influence of geopolitical risk on comprehending the dynamics of birth rates in Scotland. Conclusions: This study examines the effectiveness of several machine learning regression models in accurately predicting the number of births in Scotland using data that is not included in the model training process. Findings show promising outcomes in predicting births, while geopolitical instability has been indicated as a substantial influence on birth rates and fertility rates.

Using CT images to assist the segmentation of MR images via generalization: Segmentation of the renal parenchyma of renal carcinoma patients.

Developing deep learning models for segmenting medical images in multiple modalities with less data and annotation is an attractive and challenging task, which was previously discussed as being accomplished by complex external frameworks for bridging the gap between different modalities. Exploring the generalization ability of networks in medical images in different modalities could provide more simple and accessible methods, yet comprehensive testing could still be needed. To explore the feasibility and robustness of using computed tomography (CT) images to assist the segmentation of magnetic resonance (MR) images via the generalization, in the segmentation of renal parenchyma of renal cell carcinoma (RCC) patients. Nephrographic CT images and fat-suppressed T2-weighted (fs-T2W) images were retrospectively collected. The pure CT dataset included 116 CT images. Additionally, 240 MR images were randomly divided into subsets A and B. From subset A, three training datasets were constructed, each containing 40, 80, and 120 images, respectively. Similarly, three datasets were constructed from subset B. Subsequently, datasets with mixed modality were created by combining these pure MR datasets with the 116 CT images. The 3D-UNET models for segmenting the renal parenchyma in two steps were trained using these 13 datasets: segmenting kidneys and then the renal parenchyma. These models were evaluated in internal MR (n=120), CT (n=65) validation datasets, and an external validation dataset of CT (n=79), using the mean of the dice similarity coefficient (DSC). To demonstrate the robustness of generalization ability over different proportions of modalities, we compared the models trained with mixed modality in three different proportions and pure MR, using repeated measures analysis of variance (RM-ANOVA). We developed a renal parenchyma volume quantification tool by the trained models. The mean differences and Pearson correlation coefficients between the model segmentation volume and the ground truth segmentation volume were calculated for its evaluation. The mean DSCs of models trained with 116 data in CT in the validation of MR were 0.826, 0.842, and 0.953, respectively, for the predictions of kidney segmentation model on whole image, renal parenchymal segmentation model on kidneys with RCC and without RCC. For all models trained with mixed modality, the means of DSC were above 0.9, in all validations of CT and MR. According to the results of the comparison between models trained with mixed modality and pure MR, the means of DSC of the former were significantly greater or equal to the latter, at all three different proportions of modalities. The differences of volumes were all significantly lower than one-third of the volumetric quantification error of a previous method, and the Pearson correlation coefficients of volumes were all above 0.96 on kidneys with and without RCC of three validations. CT images could be used to assist the segmentation of MR images via the generalization, with or without the supervision of MR data. This ability showed acceptable robustness. A tool for accurately measuring renal parenchymal volume on CT and MR images was established.

Feature Transfer Learning for Fatigue Life Prediction of Additive Manufactured Metals With Small Samples

ABSTRACTA feature transfer learning (FTL)‐based model is proposed to address small‐sample problems in fatigue life prediction of additively manufactured (AM) metals. Transfer component analysis (TCA) is studied for data alignment before model training. Correspondingly, two TCA improvement strategies are further considered to aggregate training data from distinct AM processing conditions. An experimental database consisting of 103 fatigue data is built for model evaluation. The results demonstrate that the proposed model outperforms conventional machine learning models and other transfer learning‐based models in terms of accuracy and data demand, showing good applicability for AM fatigue life assessment.

Balanced Training Sets Improve Deep Learning-Based Prediction of CRISPR sgRNA Activity.

CRISPR-Cas systems have transformed the field of synthetic biology by providing a versatile method for genome editing. The efficiency of CRISPR systems is largely dependent on the sequence of the constituent sgRNA, necessitating the development of computational methods for designing active sgRNAs. While deep learning-based models have shown promise in predicting sgRNA activity, the accuracy of prediction is primarily governed by the data set used in model training. Here, we trained a convolutional neural network (CNN) model and a large language model (LLM) on balanced and imbalanced data sets generated from CRISPR-Cas12a screening data for the yeast Yarrowia lipolytica and evaluated their ability to predict high- and low-activity sgRNAs. We further tested whether prediction performance can be improved by training on imbalanced data sets augmented with synthetic sgRNAs. Lastly, we demonstrated that adding synthetic sgRNAs to inherently imbalanced CRISPR-Cas9 data sets from Y. lipolytica and Komagataella phaffii leads to improved performance in predicting sgRNA activity, thus underscoring the importance of employing balanced training sets for accurate sgRNA activity prediction.

Impact of Adverse Weather and Image Distortions on Vision-Based UAV Detection: A Performance Evaluation of Deep Learning Models

Unmanned aerial vehicle (UAV) detection in real-time is a challenging task despite the advances in computer vision and deep learning techniques. The increasing use of UAVs in numerous applications has generated worries about possible risks and misuse. Although vision-based UAV detection methods have been proposed in recent years, a standing open challenge and overlooked issue is that of adverse weather. This work is the first, to the best of our knowledge, to investigate the impact of adverse weather conditions and image distortions on vision-based UAV detection methods. To achieve this, a custom training dataset was curated with images containing a variety of UAVs in diverse complex backgrounds. In addition, this work develops a first-of-its-kind dataset, to the best of our knowledge, with UAV-containing images affected by adverse conditions. Based on the proposed datasets, a comprehensive benchmarking study is conducted to evaluate the impact of adverse weather and image distortions on the performance of popular object detection methods such as YOLOv5, YOLOv8, Faster-RCNN, RetinaNet, and YOLO-NAS. The experimental results reveal the weaknesses of the studied models and the performance degradation due to adverse weather, highlighting avenues for future improvement. The results show that even the best UAV detection model’s performance degrades in mean average precision (mAP) by −50.62 points in torrential rain conditions, by −52.40 points in high noise conditions, and by −77.0 points in high motion blur conditions. To increase the selected models’ resilience, we propose and evaluate a strategy to enhance the training of the selected models by introducing weather effects in the training images. For example, the YOLOv5 model with the proposed enhancement strategy gained +35.4, +39.3, and +44.9 points higher mAP in severe rain, noise, and motion blur conditions respectively. The findings presented in this work highlight the advantages of considering adverse weather conditions during model training and underscore the significance of data enrichment for improving model generalization. The work also accentuates the need for further research into advanced techniques and architectures to ensure more reliable UAV detection under extreme weather conditions and image distortions.

ATGAN: attention-based temporal GAN for EEG data augmentation in personal identification

With the development of brain–computer interfaces in recent years, deep learning models have been widely used in EEG-based personal identification. The training of complex models requires a large amount of data, but the acquisition of EEG signals is very time-consuming and energy-consuming. EEG data augmentation can reduce the overhead of EEG data acquisition and provide the data volume required by various complex models, which is conducive to applying EEG identification in practice. The generative adversarial network (GAN) has been applied to augment EEG data features. But the augmentation of temporal EEG signals is less studied. Therefore, this paper proposes a method of temporal EEG data augmentation based on an attention mechanism time series generative adversarial network (ATGAN), which can expand the original EEG dataset to facilitate the training of downstream deep learning classifiers. The experiment is conducted on the BCI Competition IV 2a dataset. The ATGAN model is trained with the temporal EEG signals after preprocessing, and the generated augmentation data by ATGAN is used to expand the original dataset. The accuracy of the EEG-based identification task on the augmented dataset is 7.78% higher than that on the original dataset, which proves the effectiveness of the proposed method.

Artificial Neural Network Model to Predict the Exportation of Traditional Products of Colombia

This article develops the design, training, and validation of a computational model to predict the exportation of traditional Colombian products using artificial neural networks. This work aims to obtain a model using a single multilayer neural network. The number of historical input data (delays), the number of layers, and the number of neurons were considered for the neural network design. In this way, an experimental design of 64 configurations of the neural network was performed. The main arduousness addressed in this work is the significant difference (in tons) in the values of the considered products. The results show the effect that occurs due to the different range values, and one of the proposals made allows this limitation to be handled appropriately. In summary, this work seeks to provide essential information for formulating a model for efficient and practical application.