Limited Training Data Research Articles

Improving ranking-based question answering with weak supervision for low-resource Qur’anic texts

AbstractThis work tackles the challenge of ranking-based machine reading comprehension (MRC), where a question answering (QA) system generates a ranked list of relevant answers for each question instead of simply extracting a single answer. We highlight the limitations of traditional learning methods in this setting, particularly under limited training data. To address these issues, we propose a novel ranking-inspired learning method that focuses on ranking multiple answer spans instead of single answer extraction. This method leverages lexical overlap as weak supervision to guide the ranking process. We evaluate our approach on the Qur’an Reading Comprehension Dataset (QRCD), a low-resource Arabic dataset over the Holy Qur’an. We employ transfer learning with external resources to fine-tune various transformer-based models, mitigating the low-resource challenge. Experimental results demonstrate that our proposed method significantly outperforms standard mechanisms across different models. Furthermore, we show its better alignment with the ranking-based MRC task and the effectiveness of external resources for this low-resource dataset. Our best performing model achieves a state-of-the-art partial Reciprocal Rank (pRR) score of 63.82%, surpassing the previous best-known score of 58.60%. To foster further research, we release code [GitHub repository:github.com/mohammed-elkomy/weakly-supervised-mrc], trained models, and detailed experiments to the community.

An Enhanced Feature Learning CNN Using Auto-encoder for Partially Occluded Target Recognition in SAR Imagery

Deep learning algorithms have found extensive utility in Automatic Target Recognition (ATR) using Synthetic Aperture Radar (SAR). Despite this, conventional deep learning-based ATR models often lack in their performance when subjected to Extended Operating Conditions (EOCs). Significant variations in depression, noise interference, limited training data, and partial occlusion are primary factors in EOCs. In this paper, we proposed a deep neural network-based technique to recognize SAR targets for Standard Operating Conditions (SOC) and Extended Operating Conditions (EOC). The proposed approach comprises an encoder and decoder network. The encoder network integrates multiple dilated convolutions to address noise-induced recognition challenges for extracting multi-scale features. Furthermore, a feature refinement module is integrated within the multi-scale channels to enhance the extraction of discriminative and resilient features. Feature refinement selectively emphasizes informative features while suppressing irrelevant ones. Additionally, a salient attention block and a spatial feature learning module are introduced for feature selection. Selected features are then utilized in the top-level network layer of the encoder network to preserve spatial relationships among diverse features. This spatial feature learning module aids in recognition performance, particularly when training data is scarce. The decoder network consists of stacked transposed convolution layers, facilitating the encoder network's acquisition of discriminative image features. Experimental evaluations were conducted on the Moving and Stationary Target Acquisition and Recognition (MSTAR) dataset to validate the efficacy of the proposed methodology.

Variable resolution machine learning optimization of antennas using global sensitivity analysis.

The significance of rigorous optimization techniques in antenna engineering has grown significantly in recent years. For many design tasks, parameter tuning must be conducted globally, presenting a challenge due to associated computational costs. The popular bio-inspired routines often necessitate thousands of merit function calls to converge, generating prohibitive expenses whenever the design process relies on electromagnetic (EM) simulation models. Surrogate-assisted methods offer acceleration, yet constructing reliable metamodels is hindered in higher-dimensional spaces and systems with highly nonlinear characteristics. This work suggests an innovative technique for global antenna optimization embedded within a machine-learning framework. It involves iteratively refined kriging surrogates and particle swarm optimization for generating infill points. The search process operates within a reduced-dimensionality region established through fast global sensitivity analysis. Domain confinement enables the creation of accurate behavioral models using limited training data, resulting in low CPU costs for optimization. Additional savings are realized by employing variable-resolution EM simulations, where low-fidelity models are utilized during the global search stage (including sensitivity analysis), and high-fidelity ones are reserved for final (gradient-based) tuning of antenna parameters. Comprehensive verification demonstrates the consistent performance of the proposed procedure, its superiority over benchmark techniques, and the relevance of the mechanisms embedded into the algorithm for enhancing search process reliability, design quality, and computational efficiency.

Proto-DS: A Self-Supervised Learning-Based Nondestructive Testing Approach for Food Adulteration with Imbalanced Hyperspectral Data

Conventional food fraud detection using hyperspectral imaging (HSI) relies on the discriminative power of machine learning. However, these approaches often assume a balanced class distribution in an ideal laboratory environment, which is impractical in real-world scenarios with diverse label distributions. This results in suboptimal performance when less frequent classes are overshadowed by the majority class during training. Thus, the critical research challenge emerges of how to develop an effective classifier on a small-scale imbalanced dataset without significant bias from the dominant class. In this paper, we propose a novel nondestructive detection approach, which we call the Dice Loss Improved Self-Supervised Learning-Based Prototypical Network (Proto-DS), designed to address this imbalanced learning challenge. The proposed amalgamation mitigates the label bias on the most frequent class, further improving robustness. We validate our proposed method on three collected hyperspectral food image datasets with varying degrees of data imbalance: Citri Reticulatae Pericarpium (Chenpi), Chinese herbs, and coffee beans. Comparisons with state-of-the-art imbalanced learning techniques, including the Synthetic Minority Oversampling Technique (SMOTE) and class-importance reweighting, reveal our method’s superiority. Notably, our experiments demonstrate that Proto-DS consistently outperforms conventional approaches, achieving the best average balanced accuracy of 88.18% across various training sample sizes, whereas the Logistic Model Tree (LMT), Multi-Layer Perceptron (MLP), and Convolutional Neural Network (CNN) approaches attain only 59.42%, 60.38%, and 66.34%, respectively. Overall, self-supervised learning is key to improving imbalanced learning performance and outperforms related approaches, while both prototypical networks and the Dice loss can further enhance classification performance. Intriguingly, self-supervised learning can provide complementary information to existing imbalanced learning approaches. Combining these approaches may serve as a potential solution for building effective models with limited training data.

PCNet: a human pose compensation network based on incremental learning for sports actions estimation

Human pose estimation has a wide range of applications. Existing methods perform well in conventional domains, but there are certain defects when they are applied to sports activities. The first is lack of estimation of the extremity posture, making it impossible to comprehensively evaluate the movement posture; the second is insufficient occlusion handling. Therefore, we propose a human pose compensation network based on incremental learning, which obtains shared weights to extract detailed features under the premise of limited extremity training data. We propose a higher-order feature compensator (HOF-compensator) to embed the attributes of the extremity into the torso and limbs topology structure, building a complete higher-order feature. In addition, to improve the occlusion handling performance, we propose an occlusion feature enhancement attention mechanism (OFE-attention) that can identify occluded keypoints and enhance attention to occlusion areas. We design comparative experiments on three public datasets and a self-built sports dataset, achieving the highest mean accuracy among all comparative methods. In addition, we design a series of ablation analysis and visualization displays to verify that our method performs best in sports pose estimation.

Modeling and analysis of nonlinear dynamics of axisymmetric vector nozzle based on deep neural network

Abstract The axisymmetric nozzle mechanism is the core part for thrust vectoring of aero engine, which contains complex rigid-flexible coupled multibody system with joints clearance and significantly reduces the efficiency in modeling and calculation, therefore the kinematics and dynamics analysis of axisymmetric vectoring nozzle mechanism based on deep neural network is proposed. The deep neural network model of the axisymmetric vector nozzle is established according to the limited training data from the physical dynamic model and then used to predict the kinematics and dynamics response of the axisymmetric vector nozzle. This study analyses the effects of joint clearance on the kinematics and dynamics of the axisymmetric vector nozzle mechanism by a data-driven model. It is found that the angular acceleration of the expanding blade and the driving force are mostly affected by joint clearance followed by the angle, angular velocity and position of the expanding blade. Larger joint clearance results in more pronounced fluctuations of the dynamic response of the mechanism, which is due to the greater relative velocity and contact force between the bushing and the pin. Since axisymmetric vector nozzles are highly complex nonlinear systems, traditional numerical methods of dynamics are extremely time-consuming. Our work indicates that the data-driven approach greatly reduces the computational cost while maintaining accuracy, and can be used for rapid evaluation and iterative computation of complex multibody dynamics of engine nozzle mechanism.

Exploring Cutout and Mixup for Robust Human Activity Recognition on Sensor and Skeleton Data

Human Activity Recognition (HAR) is an essential area of research in Artificial Intelligence and Machine Learning, with numerous applications in healthcare, sports science, and smart environments. While several advancements in the field, such as attention-based models and Graph Neural Networks, have made great strides, this work focuses on data augmentation methods that tackle issues like data scarcity and task variability in HAR. In this work, we investigate and expand the use of mixup and cutout data augmentation methods to sensor-based and skeleton-based HAR datasets. These methods were first widely used in Computer Vision and Natural Language Processing. We use both augmentation techniques, customized for time-series and skeletal data, to improve the robustness and performance of HAR models by diversifying the data and overcoming the drawbacks of having limited training data. Specifically, we customize mixup data augmentation for sensor-based datasets and cutout data augmentation for skeleton-based datasets with the goal of improving model accuracy without adding more data. Our results show that using mixup and cutout techniques improves the accuracy and generalization of activity recognition models on both sensor-based and skeleton-based human activity datasets. This work showcases the potential of data augmentation techniques on transformers and Graph Neural Networks by offering a novel method for enhancing time series and skeletal HAR tasks.

Application of convolutional neural networks in image classification and applications of improved convolutional neural networks

Abstract. This paper reviews the application and improvement of convolutional neural networks (CNNs) in image classification. Firstly, a shallow CNN for interstitial lung disease image classification is presented. This model suppresses overfitting through a unique network architecture and optimisation algorithm. Next, the improved VGG16 architecture and MIDNet18 model are discussed and their superior performance in brain tumour image classification is demonstrated. Subsequently, a CNN-CapsNet model for cervical cancer image classification and its improvement are presented and the customised model is compared with the conventional VGG-16 CNN architecture in the paper. Next, the application of sparse convolutional kernels and hybrid sparse convolutional kernels (HDCs) in solving the problem of computational resource consumption is presented. Subsequently, methods for solving the problem of limited training data through transfer learning and network data augmentation techniques are discussed, as well as GAN-generated datasets for solving the overfitting problem. Finally, the effect of degraded images on the classification effectiveness of CNNs is explored. The results show that the improved CNN architecture and algorithms have significant effects in solving the problems of overfitting and computational resource consumption, and can significantly improve the accuracy and efficiency of image classification. And degraded images do adversely affect the accuracy of CNN for image classification.

Application of Prompt Learning Models in Identifying the Collaborative Problem Solving Skills in an Online Task

Collaborative problem solving (CPS) competence is considered one of the essential 21st-century skills. To facilitate the assessment and learning of CPS competence, researchers have proposed a series of frameworks to conceptualize CPS and explored ways to make sense of the complex processes involved in collaborative problem solving. However, encoding explicit behaviors into subskills within the frameworks of CPS skills is still a challenging task. Traditional studies have relied on manual coding to decipher behavioral data for CPS, but such coding methods can be very time-consuming and cannot support real-time analyses. Scholars have begun to explore approaches for constructing automatic coding models. Nevertheless, the existing models built using machine learning or deep learning techniques depend on a large amount of training data and have relatively low accuracy. To address these problems, this paper proposes a prompt-based learning pre-trained model. The model can achieve high performance even with limited training data. In this study, three experiments were conducted, and the results showed that our model not only produced the highest accuracy, macro F1 score, and kappa values on large training sets, but also performed the best on small training sets of the CPS behavioral data. The application of the proposed prompt-based learning pre-trained model contributes to the CPS skills coding task and can also be used for other CSCW coding tasks to replace manual coding.

Towards Integrating Federated Learning with Split Learning via Spatio-temporal Graph Framework for Brain Disease Prediction.

Functional Magnetic Resonance Imaging (fMRI) is used for extracting blood oxygen signals from brain regions to map brain functional connectivity for brain disease prediction. Despite its effectiveness, fMRI has not been widely used: on the one hand, collecting and labeling the data is time-consuming and costly, which limits the amount of valid data collected at a single healthcare site; on the other hand, integrating data from multiple sites is challenging due to data privacy restrictions. To address these issues, we propose a novel, integrated Federated learning and Split learning Spatio-temporal Graph framework (FS2G). Specifically, we introduce federated learning and split learning techniques to split a spatio-temporal model into a client temporal model and a server spatial model. In the client temporal model, we propose a time-aware mechanism to focus on changes in brain functional states and use an InceptionTime model to extract information about changes in the brain states of each subject. In the server spatial model, we propose a united graph convolutional network to integrate multiple graph convolutional networks. Integrating federated learning and split learning, FS2G can utilize multi-site fMRI data without violating data privacy protection and reduce the risk of overfitting as it is capable of learning from limited training data sets. Moreover, it boosts the extraction of spatio-temporal features of fMRI using spatio-temporal graph networks. Experiments on ABIDE and ADHD200 datasets demonstrate that our proposed method outperforms state-of-the-art methods. In addition, we explore biomarkers associated with brain disease prediction using community discovery algorithms using intermediate results of FS2G. The source code is available at https://github.com/yutian0315/FS2G.

Contrastive learning-enabled digital twin framework for fault diagnosis of rolling bearing

Abstract Rolling bearings are essential components in various industrial machines, and their failures can lead to significant downtime and maintenance costs. Traditional data-driven fault diagnosis methods often require extensive fault datasets for training, which may not always be available in critical industrial scenarios, limiting their practicality. Digital twins, virtual representations of physical entities reflecting their operational conditions, offer a promising solution for the fault diagnosis of rolling bearings with limited fault data. In this paper, we propose a novel digital twin-driven framework to address the challenge of limited training data in rolling bearing fault diagnosis. Firstly, a virtual bearing simulation model is used to generate the simulated data. Subsequently, a transformer-based network is introduced to learn the discrepancy features from the raw data. Then, a Maximum Mean Discrepancy loss and a supervised contrastive learning loss for raw and augmentation data are established to achieve global domain alignment and instance-based domain alignment. Finally, an unsupervised contrastive learning loss for the augmentation data of the target domain is established to further improve the diagnostic performance. In five cross-domain fault diagnosis tasks representing real industrial scenarios set in this study, the proposed method achieved an average diagnostic accuracy of 84.40%, exceeding two existing advanced domain adaptation methods by more than 10%. The experimental results demonstrate that the proposed method achieves high diagnostic performance in real industrial scenarios where labeled data is lacking. This shows its significant benefits for monitoring the condition of critical bearings.

S2Match: Self-paced sampling for data-limited semi-supervised learning

Data-limited semi-supervised learning tends to be severely degraded by miscalibration (i.e., misalignment between confidence and correctness of predicted pseudo labels) and stuck at poor local minima while learning from the same set of over-confident yet incorrect pseudo labels repeatedly. We design a simple and effective self-paced sampling technique that can greatly alleviate the impact of miscalibration and learn more accurate semi-supervised models from limited training data. Instead of employing static or dynamic confidence thresholds which is sensitive to miscalibration, the proposed self-paced sampling follows a simple linear policy to select pseudo labels which eases repeated learning from the same set of falsely predicted pseudo labels at the early training stage and lowers the chance of being stuck at local minima effectively. Despite its simplicity, extensive evaluations over multiple data-limited semi-supervised tasks show the proposed self-paced sampling outperforms the state-of-the-art consistently by large margins.

A temporal–spectral fusion transformer with subject-specific adapter for enhancing RSVP-BCI decoding

The Rapid Serial Visual Presentation (RSVP)-based Brain–Computer Interface (BCI) is an efficient technology for target retrieval using electroencephalography (EEG) signals. The performance improvement of traditional decoding methods relies on a substantial amount of training data from new test subjects, which increases preparation time for BCI systems. Several studies introduce data from existing subjects to reduce the dependence of performance improvement on data from new subjects, but their optimization strategy based on adversarial learning with extensive data increases training time during the preparation procedure. Moreover, most previous methods only focus on the single-view information of EEG signals, but ignore the information from other views which may further improve performance. To enhance decoding performance while reducing preparation time, we propose a Temporal-Spectral fusion transformer with Subject-specific Adapter (TSformer-SA). Specifically, a cross-view interaction module is proposed to facilitate information transfer and extract common representations across two-view features extracted from EEG temporal signals and spectrogram images. Then, an attention-based fusion module fuses the features of two views to obtain comprehensive discriminative features for classification. Furthermore, a multi-view consistency loss is proposed to maximize the feature similarity between two views of the same EEG signal. Finally, we propose a subject-specific adapter to rapidly transfer the knowledge of the model trained on data from existing subjects to decode data from new subjects. Experimental results show that TSformer-SA significantly outperforms comparison methods and achieves outstanding performance with limited training data from new subjects. This facilitates efficient decoding and rapid deployment of BCI systems in practical use.

A Multi-Tree Genetic Programming-Based Ensemble Approach to Image Classification With Limited Training Data [Research Frontier

A Multi-Tree Genetic Programming-Based Ensemble Approach to Image Classification With Limited Training Data [Research Frontier

A novel method for subgrade cumulative deformation prediction of high-speed railways based on empiricism-constrained neural network and SHapley Additive exPlanations analysis

Understanding the long-term deformation of high-speed railway subgrade is essential for solving deformation issues and managing operations. Machine learning methods are commonly used to predict subgrade cumulative deformation (SCD). However, traditional machine learning models for SCD prediction have poor generalization to new data and lack visualization. Hence, this study proposes a novel method using an empiricism-constrained neural network (ECNN) and SHapley Additive exPlanations (SHAP) analysis for predicting SCD of high-speed railways. Firstly, the SCD prediction dataset is constructed and divided into training and test sets. Then, neural network models are developed using the training set, and the optimal model is determined based on the comprehensive scoring results on the test set. The optimal model couples empirical information into the neural network with loss function modification, to create the ECNN model. Finally, the interpretability of the ECNN model is analyzed using the SHAP method. The results indicate that the Bi-directional Gated Recurrent Unit (Bi-GRU) model is the optimal model with the highest CSI value of 23. The ECNN model outperforms the Bi-GRU in generalization to new data, especially in long-term SCD prediction with limited training data. Contribution analysis shows that the top two features influencing the prediction are St-1 (54.4%) and St-2 (30.4%), consistent with the findings of the ablation analysis. The research results can provide a new reference for predicting the SCD of high-speed railways.

Redefining the accumulated temperature index for accurate prediction of rice flowering time in diverse environments.

Accurate prediction of flowering time across diverse environments is crucial for effective crop management and breeding. While the accumulated temperature index (ATI) is widely used as an indicator for estimating flowering time, its traditional definition lacks systematic evaluation and genetic basis understanding. Here, using data from 422 rice hybrids across 47 locations, we identified the optimal ATI calculation window as 1 day after sowing to 26 days before flowering. Based on this redefined ATI, we developed a single-parameter model that outperforms the state-of-the-art reaction norm index model in both accuracy and stability, especially with limited training data. We identified 10 loci significantly associated with ATI variation, including two near known flowering time genes and four linked to ecotype differentiation. To enhance practical utility, we developed an efficient flowering time prediction kit using 28 functionally relevant markers, complemented by a user-friendly online tool (http://xielab.hzau.edu.cn/ATI). Our approach can be easily applied to other crops, as ATI is commonly used across various agricultural systems.

Applications of ChatGPT in Heart Failure Prevention, Diagnosis, Management, and Research: A Narrative Review.

Heart failure (HF) is a leading cause of mortality, morbidity, and financial burden worldwide. The emergence of advanced artificial intelligence (AI) technologies, particularly Generative Pre-trained Transformer (GPT) systems, presents new opportunities to enhance HF management. In this review, we identified and examined existing studies on the use of ChatGPT in HF care by searching multiple medical databases (PubMed, Google Scholar, Medline, and Scopus). We assessed the role of ChatGPT in HF prevention, diagnosis, and management, focusing on its influence on clinical decision-making and patient education. However, ChatGPT faces limited training data, inherent biases, and ethical issues that hinder its widespread clinical adoption. We review these limitations and highlight the need for improved training approaches, greater model transparency, and robust regulatory compliance. Additionally, we explore the effectiveness of ChatGPT in managing HF, particularly in reducing hospital readmissions and improving patient outcomes with customized treatment plans while addressing social determinants of health (SDoH). In this review, we aim to provide healthcare professionals and policymakers with an in-depth understanding of ChatGPT's potential and constraints within the realm of HF care.

Impact of metadata in multimodal classification of bone tumours

The accurate classification of bone tumours is crucial for guiding clinical decisions regarding treatment and follow-up. However, differentiating between various tumour types is challenging due to the rarity of certain entities, high intra-class variability, and limited training data in clinical practice. This study proposes a multimodal deep learning model that integrates clinical metadata and X-ray imaging to improve the classification of primary bone tumours. The dataset comprises 1,785 radiographs from 804 patients collected between 2000 and 2020, including metadata such as age, affected bone site, tumour position, and gender. Ten tumour types were selected, with histopathology or tumour board decisions serving as the reference standard.MethodsOur model is based on the NesT image classification model and a multilayer perceptron with a joint fusion architecture. Descriptive statistics included incidence and percentage ratios for discrete parameters, and mean, standard deviation, median, and interquartile range for continuous parameters.ResultsThe mean age of the patients was 33.62 ± 18.60 years, with 54.73% being male. Our multimodal deep learning model achieved 69.7% accuracy in classifying primary bone tumours, outperforming the Vision Transformer model by five percentage points. SHAP values indicated that age had the most substantial influence among the considered metadata.ConclusionThe joint fusion approach developed in this study, integrating clinical metadata and imaging data, outperformed state-of-the-art models in classifying primary bone tumours. The use of SHAP values provided insights into the impact of different metadata on the model’s performance, highlighting the significant role of age. This approach has potential implications for improving diagnostic accuracy and understanding the influence of clinical factors in tumour classification.

Geometric neural network based on phase space for BCI-EEG decoding.

\textbf{Objective:} 
The integration of Deep Learning (DL) algorithms on brain signal analysis is still in its nascent stages compared to their success in fields like Computer Vision. This is particularly true for BCI, where the brain activity is decoded to control external devices without requiring muscle control.
Electroencephalography (EEG) is a widely adopted choice for designing BCI systems due to its non-invasive and cost-effective nature and excellent temporal resolution. Still, it comes at the expense of limited training data, poor signal-to-noise, and a large variability across and within-subject recordings. 
Finally, setting up a BCI system with many electrodes takes a long time, hindering the widespread adoption of reliable DL architectures in BCIs outside research laboratories. To improve adoption, we need to improve user comfort using, for instance, reliable algorithms that operate with few electrodes. 
\textbf{Approach:} Our research aims to develop a DL algorithm that delivers effective results with a limited number of electrodes. Taking advantage of the Augmented Covariance Method and the framework of SPDNet, we propose the \method{} architecture and analyze its performance and the interpretability of the results. The evaluation is conducted on 5-fold cross-validation, using only three electrodes positioned above the Motor Cortex. The methodology was tested on nearly 100 subjects from several open-source datasets using the Mother Of All BCI Benchmark (MOABB) framework. 
\textbf{Main results:} The results of our \method{} demonstrate that the augmented approach combined with the SPDNet significantly outperforms all the current state-of-the-art DL architecture in MI decoding. 
\textbf{Significance:} This new architecture is explainable and with a low number of trainable parameters.

Camera-Based Respiratory Imaging System for Monitoring Infant Thoracoabdominal Patterns of Respiration.

Existing respiratory monitoring techniques primarily focus on respiratory rate measurement, neglecting the potential of using thoracoabdominal patterns of respiration for infant lung health assessment. To bridge this gap, we exploit the unique advantage of spatial redundancy of a camera sensor to analyze the infant thoracoabdominal respiratory motion. Specifically, we propose a camera-based respiratory imaging (CRI) system that utilizes optical flow to construct a spatio-temporal respiratory imager for comparing the infant chest and abdominal respiratory motion, and employs deep learning algorithms to identify infant abdominal, thoracoabdominal synchronous, and thoracoabdominal asynchronous patterns of respiration. To alleviate the challenges posed by limited clinical training data and subject variability, we introduce a novel multiple-expert contrastive learning (MECL) strategy to CRI. It enriches training samples by reversing and pairing different-class data, and promotes the representation consistency of same-class data through multi-expert collaborative optimization. Clinical validation involving 44 infants shows that MECL achieves 70% in sensitivity and 80.21% in specificity, which validates the feasibility of CRI for respiratory pattern recognition. This work investigates a novel video-based approach for assessing the infant thoracoabdominal patterns of respiration, revealing a new value stream of video health monitoring in neonatal care.