Training Dataset Research Articles

Scientific machine learning based reduced-order models for plasma turbulence simulations

This paper investigates non-intrusive Scientific Machine Learning (SciML) Reduced-Order Models (ROMs) for plasma turbulence simulations. In particular, we focus on Operator Inference (OpInf) to build low-cost physics-based ROMs from data for such simulations. As a representative example, we consider the (classical) Hasegawa–Wakatani (HW) equations used for modeling two-dimensional electrostatic drift-wave turbulence. For a comprehensive perspective of the potential of OpInf to construct predictive ROMs, we consider three setups for the HW equations by varying a key parameter, namely, the adiabaticity coefficient. These setups lead to the formation of complex and nonlinear dynamics, which makes the construction of predictive ROMs of any kind challenging. We generate the training datasets by performing direct numerical simulations of the HW equations and recording the computed state data and outputs over a time horizon of 100 time units in the turbulent phase. We then use these datasets to construct OpInf ROMs for predictions over 400 additional time units, that is, 400% more than the training horizon. Our results show that the OpInf ROMs capture important statistical features of the turbulent dynamics and generalize beyond the training time horizon while reducing the computational effort of the high-fidelity simulation by up to five orders of magnitude. In the broader context of fusion research, this shows that non-intrusive SciML ROMs have the potential to drastically accelerate numerical studies, which can ultimately enable tasks such as the design of optimized fusion devices.

From Policy to Prediction: Assessing Forecasting Accuracy in an Integrated Framework with Machine Learning and Disease Models.

To improve the forecasting accuracy of the spread of infectious diseases, a hybrid model was recently introduced where the commonly assumed constant disease transmission rate was actively estimated from enforced mitigating policy data by a machine learning (ML) model and then fed to an extended susceptible-infected-recovered model to forecast the number of infected cases. Testing only one ML model, that is, gradient boosting model (GBM), the work left open whether other ML models would perform better. Here, we compared GBMs, linear regressions, k-nearest neighbors, and Bayesian networks (BNs) in forecasting the number of COVID-19-infected cases in the United States and Canadian provinces based on policy indices of future 35 days. There was no significant difference in the mean absolute percentage errors of these ML models over the combined dataset []. In two provinces, a significant difference was observed [], yet posthoc tests revealed no significant difference in pairwise comparisons. Nevertheless, BNs significantly outperformed the other models in most of the training datasets. The results put forward that the ML models have equal forecasting power overall, and BNs are best for data-fitting applications.

Precise Image-level Localization of Intracranial Hemorrhage on Head CT Scans with Deep Learning Models Trained on Study-level Labels.

Purpose To develop a highly generalizable weakly supervised model to automatically detect and localize image-level intracranial hemorrhage (ICH) by using study-level labels. Materials and Methods In this retrospective study, the proposed model was pretrained on the image-level Radiological Society of North America dataset and fine-tuned on a local dataset by using attention-based bidirectional long short-term memory networks. This local training dataset included 10 699 noncontrast head CT scans in 7469 patients, with ICH study-level labels extracted from radiology reports. Model performance was compared with that of two senior neuroradiologists on 100 random test scans using the McNemar test, and its generalizability was evaluated on an external independent dataset. Results The model achieved a positive predictive value (PPV) of 85.7% (95% CI: 84.0, 87.4) and an area under the receiver operating characteristic curve of 0.96 (95% CI: 0.96, 0.97) on the held-out local test set (n = 7243, 3721 female) and 89.3% (95% CI: 87.8, 90.7) and 0.96 (95% CI: 0.96, 0.97), respectively, on the external test set (n = 491, 178 female). For 100 randomly selected samples, the model achieved performance on par with two neuroradiologists, but with a significantly faster (P < .05) diagnostic time of 5.04 seconds per scan (vs 86 seconds and 22.2 seconds for the two neuroradiologists, respectively). The model's attention weights and heatmaps visually aligned with neuroradiologists' interpretations. Conclusion The proposed model demonstrated high generalizability and high PPVs, offering a valuable tool for expedited ICH detection and prioritization while reducing false-positive interruptions in radiologists' workflows. Keywords: Computer-Aided Diagnosis (CAD), Brain/Brain Stem, Hemorrhage, Convolutional Neural Network (CNN), Transfer Learning Supplemental material is available for this article. © RSNA, 2024 See also the commentary by Akinci D'Antonoli and Rudie in this issue.

Leveraging deep learning for identification and segmentation of “CAF-1/p60-positive” nuclei in oral squamous cell carcinoma tissue samples

In the current study, we introduced a unique method for identifying and segmenting oral squamous cell carcinoma (OSCC) nuclei, concentrating on those predicted to have significant CAF-1/p60 protein expression. Our suggested model uses the StarDist architecture, a deep-learning framework designed for biomedical image segmentation tasks. The training dataset comprises painstakingly annotated masks created from tissue sections previously stained with hematoxylin and eosin (H&E) and then restained with immunohistochemistry (IHC) for p60 protein. Our algorithm uses subtle morphological and colorimetric H&E cellular characteristics to predict CAF-1/p60 IHC expression in OSCC nuclei. The StarDist-based architecture performs exceptionally well in localizing and segmenting H&E nuclei, previously identified by IHC-based ground truth. In summary, our innovative approach harnesses deep learning and multimodal information to advance the automated analysis of OSCC nuclei exhibiting specific protein expression patterns. This methodology holds promise for expediting accurate pathological assessment and gaining deeper insights into the role of CAF-1/p60 protein within the context of oral cancer progression.

A CSI-Based Data-Driven Localization Framework Using Small-Scale Training Datasets in Single-Site MIMO Systems

A CSI-Based Data-Driven Localization Framework Using Small-Scale Training Datasets in Single-Site MIMO Systems

Regressive Machine Learning for Real-Time Monitoring of Bed-Based Patients

This study introduces an ensemble model designed for real-time monitoring of bedridden patients. The model was developed using a unique dataset, specifically acquired for this study, that captures six typical movements. The dataset was balanced using the Synthetic Minority Over-sampling Technique, resulting in a diverse distribution of movement types. Three models were evaluated: a Decision Tree Regressor, a Gradient Boosting Regressor, and a Bagging Regressor. The Decision Tree Regressor achieved an accuracy of 0.892 and an R2 score of 1.0 on the training dataset, and 0.939 on the test dataset. The Boosting Regressor achieved an accuracy of 0.908 and an R2 score of 0.99 on the training dataset, and 0.943 on the test dataset. The Bagging Regressor was selected due to its superior performance and trade-offs such as computational cost and scalability. It achieved an accuracy of 0.950, an R2 score of 0.996 for the training data, and an R2 score of 0.959 for the test data. This study also employs K-Fold cross-validation and learning curves to validate the robustness of the Bagging Regressor model. The proposed system addresses practical implementation challenges in real-time monitoring, such as data latency and false positives/negatives, and is designed for seamless integration with hospital IT infrastructure. This research demonstrates the potential of machine learning to enhance patient safety in healthcare settings.

Plant Disease Detection Using Machine Learning

In this project, we have developed a plant disease detection model using a machine learning approach on a dataset named “Plant Disease Expert” which is available on Kaggle. The dataset contains 199,611 images across 58 subdirectories. The splitting of the dataset in training, validation and testing is done in ratio 90:5:5. The data preprocessing included resizing the images, normalizing pixel values and applying augmentation techniques. We have used different libraries throughout the project such as NumPy, Pandas, Operating system, Time, Matplotlib, OpenCV, Shutil, TensorFlow, Keras, Seaborn, etc. EfficientNetB3 is used as the base model. EfficientNetB3 is a convolutional neural network architecture developed by Google Brain researchers [1]. It is known for its efficiency and effectiveness in image classification tasks. Additional custom layers such as max pooling, batch normalization, dense layers, and activation functions are used to optimize the performance. The adamax optimizer and categorical cross-entropy were used during training. The model was trained for 10 epochs with a batch size of 20 and a learning rate of 0.01. For evaluation, classification report, accuracy, precision, recall, F1-score and support were used. The model achieved an accuracy of 98.76% [1].

Automated Lunar Surface Image Classification Using Deep Convolutional Neural Networks for Geological Feature Detection

Robotic rovers have vastly expanded our understanding of the lunar surface, providing detailed imagery crucial for scientific research and future exploration. However, manually classifying this imagery is time-consuming and prone to errors, necessitating automated solutions. Automated classification of lunar surface imagery is vital for efficient data analysis, site selection for future missions, and advancing lunar exploration. Developing accurate and efficient image classification systems tailored for lunar terrain is thus imperative. The objective of this study is to develop and assess an image classification system utilizing Deep Convolutional Neural Networks (DCNNs) specifically for lunar surface images. The aim is to achieve high accuracy and efficiency in identifying geological features such as craters and dunes, as observed by robotic rovers. A curated dataset of lunar surface images was partitioned into training, testing, and validation subsets. DCNNs models were trained on the training dataset and evaluated using testing and validation datasets. Hyperparameter tuning and optimization techniques were employed to enhance model performance. The classification system based on DCNNs showed promising outcomes. Model B and F achieved an accuracy of 91.1%, while Model A and D achieved 87.5%. Model C attained an accuracy of 89.3%, and Model E reached 83.9%. Visualizations of training and validation metrics revealed distinct performance patterns across models, highlighting the potential for further advancements in lunar exploration research.

Analysis of the Effectiveness of Multi-Armed Bandit Algorithms in Crop Pricing

This study explores the application of multi-armed bandit (MAB) algorithms in dynamic pricing of crops, with a focus on evaluating the adaptive upper confidence bound (asUCB) and Thompson Sampling (TS) algorithms. Through simulation experiments on historical data, the study analyzed the performance of these algorithms in fitting actual market price trends and responding to future price fluctuations. The results indicate that the asUCB algorithm performed particularly well in both the training dataset and simulation tests, demonstrating low mean squared error (MSE) and minimal cumulative regret, reflecting its rapid convergence and stable pricing capabilities. In contrast, although the TS algorithm initially performed slightly less effectively, it demonstrated unique advantages in dealing with market volatility due to its strong adaptability. This study demonstrates the potential application of MAB algorithms in dynamic pricing, providing valuable insights for pricing strategies in the agricultural product market.

Self-learning based joint multi image super-resolution and sub-pixel registration

Multi Image Super-resolution (MISR) refers to the task of enhancing the spatial resolution of a stack of low-resolution (LR) images representing the same scene. Although many deep learning-based single image super-resolution (SISR) technologies have recently been developed, deep learning has not been widely exploited for MISR, even though it can achieve higher reconstruction accuracy because more information can be extracted from the stack of LR images. One of the primary obstacles encountered by deep networks when addressing the MISR problem is the variability in the number of LR images that act as input to the network. This impedes the feasibility of adopting an end-to-end learning approach, because the varying number of input images makes it difficult to construct a training dataset for the network. Another challenge arises from the requirement to align the LR input images to generate high-resolution (HR) image of high quality, which requires complex and sophisticated methods.In this paper, we propose a self-learning based method that can simultaneously perform super-resolution and sub-pixel registration of multiple LR images. The proposed method trains a neural network with only the LR images as input and without any true target HR images; i.e., the proposed method requires no extra training dataset. Therefore, it is easy to use the proposed method to deal with different numbers of input images. To our knowledge this is the first time that a neural network is trained using only LR images to perform a joint MISR and sub-pixel registration. Experimental results confirmed that the HR images generated by the proposed method achieved better results in both quantitative and qualitative evaluations than those generated by other deep learning-based methods.

Devising a method for integrated dataset formation and selecting a model for recognizing the technical condition of unmanned aerial vehicle

The object of this study is the process of forming a training dataset for diagnosing the technical condition of unmanned aerial vehicles (UAVs) using machine-learning algorithms. UAV flights are extremely important for various aspects of troop deployment. Combat UAV flights are performed under the influence of negative factors that cause flight special cases (FSC), which hinder the execution of combat missions, lead to mission failures, and result in the aircraft damage or loss. The available capabilities of autopilots are not enough for control under complex conditions, and in certain situations, the human operator cannot timely recognize a flight special case, including evaluation of the destructive impact of enemy’s electronic warfare systems on communication channels and operation of UAV. Therefore, the urgent issue is the intellectualization of onboard control systems, particularly towards recognizing the current technical state of UAV using artificial intelligence methods. To design such systems, labeled datasets are required. The procedure for forming datasets that consider the specificity of UAV construction and their combat use under adversarial conditions is not defined, necessitating the development of an appropriate method. Based on the well-known CRISP-DM methodology, a method for dataset formation has been proposed for subsequent use in artificial intelligence systems that use various machine-learning methods. This method differs from existing ones by considering the specificity of combat mission execution under adversarial conditions, which allowed for an 8.0 % increase in the accuracy of recognizing special cases in UAV flights by the onboard system. It also enabled timely detection of electronic warfare impacts on UAV and evaluation of the effectiveness of radio signal receivers jamming

Multi-Step Ageing Prediction of NMC Lithium-Ion Batteries Based on Temperature Characteristics

The performance of lithium-ion batteries depends strongly on their ageing state; therefore, the monitoring and the prediction of the battery state of health (SoH) is necessary for an optimized and secured functioning of battery systems. This paper evaluates and compares three artificial neural network architectures for multi-step ageing prediction of lithium-ion cells: Recurrent Neural Network (RNN), Gated Recurrent Unit (GRU) and Long short-term memory (LSTM). These models use the features extracted from the cell’s temperature to predict the cell’s capacity. The features are extracted from experimental measurements of the cell’s surface temperature and selected based on Spearman correlation analysis. The prediction results were evaluated and compared considering three different percentages of the training dataset: 60%, 70%, and 80%. Training and testing data were generated experimentally based on accelerated ageing cycling tests. During these experiments, four Nickel Manganese Cobalt/Graphite (NMC) cells were cycled under a controlled temperature environment based on a dynamic current profile extracted from the Worldwide Harmonized Light Vehicles Test Cycles.

The impact of contextual understanding on neural machine translation accuracy: A case study of Indonesian cultural idioms in English translation

The translation of Indonesian idiomatic expressions into English using Neural Machine Translation (NMT) systems presents significant challenges due to the intricate nature of idiomatic language. Idioms are culturally embedded constructs conveying meanings different from their literal interpretations. This study explores the effectiveness of NMT systems in capturing and translating Indonesian idiomatic expressions into English. Using a qualitative approach, 150 diverse Indonesian idioms were evaluated through Google Translate and DeepL. Each idiom was assessed for semantic accuracy, syntactic coherence, and contextual fidelity. Qualitative analysis provided a comprehensive evaluation of translation quality. Findings reveal significant challenges for NMT systems in translating idiomatic expressions. Both Google Translate and DeepL show strengths in some areas, but also have limitations. While they generally capture literal meanings, they often miss metaphorical nuances and cultural connotations. Syntactic errors such as incorrect word order and tense inconsistencies are common, especially in complex idioms. In addition, contextual fidelity analysis shows that NMT systems struggle with contextual appropriateness and pragmatic usage, resulting in translations lacking cultural sensitivity and relevance. These challenges highlight the need for improved algorithms to better interpret and translate idiomatic language across cultural contexts. Accurate idiom translation is crucial for language learners to understand cultural nuances and idiomatic usage. Enhancing the capabilities of NMT systems requires refining algorithms, developing context-aware models, and expanding training datasets with diverse idiomatic expressions. Overcoming these challenges will advance machine translation capabilities and improve cross-cultural communication and language learning experiences.

Advanced Algorithms for Offline Signature Recognition and Enhanced Forgery Detection Mechanisms

In contemporary society, signatures hold significant importance in various critical documents such as bank cheques, passports, and driver's licenses. Unfortunately, they can be counterfeited through various means, giving rise to issues like fraudulent identifications, identity theft, and cyber-attacks. To mitigate this problem, our project is centered around the development of a system that discerns the authenticity of a signature, distinguishing between genuine and forged signatures within a dataset. We have employed Convolutional Neural Networks (CNN) and deep learning for this purpose. This choice is driven by the understanding that signatures evolve over time due to a range of behavioral factors like age, mental state, and physical well-being. Consequently, our system is designed to adapt and learn from diverse training datasets, enhancing its accuracy in detection. While online and offline signature verification methods exist, our project primarily focuses on the latter, specifically targeting the detectionof forged offline signatures.

A simplified methodology to determine the latent heat of macroencapsulated phase change materials for building envelope applications

Incorporating macroencapsulated phase change materials (PCMs) in building envelopes is a strategy to reduce energy consumption. Current characterization methods focus on small quantities of PCM alone, even if it was demonstrated that they are not representative of the thermal behavior of macroencapsulated PCMs, and rely on highly specific equipment. This work proposes a novel methodology to determine the latent heat of whole macroencapsulated PCMs for building envelope applications using only temperature measurements. This method takes into account the greater quantity of PCM and the capsule material, unlike previous works, while it simplifies the necessary equipment. To develop the predictive model, a wide set of PCM panels was subjected to controlled temperature variations of 0.5 °C/h and 1 °C/h employing an insulated hotbox connected to a climatic module. Transient surface temperature and heat flux were measured. Experimental results demonstrated that not only heat flux measurements, but also temperature can be used to identify the beginning and end of the phase change process. The experimental results were then divided into a training and validation dataset. The training dataset was used to build the predictive model that correlates the area under the heat flux curve to that enclosed by the temperature curve during phase transitions. The model was validated with the remaining experimental data and optimized to assess the influence of temperature variation rate and phase change processes (melting and solidification). We found that the quotient of the heat flux area over the temperature area is 20.5 W/m2·K. Based on this, we identified that the latent heat can be accurately estimated using only temperature data, with an average error of less than 6 %. This methodology is simple but efficient requiring only thermocouples, and no other sophisticated instrumentation. The proposed model predicts the thermal response of macroencapsulated PCM with robustness, versatility, and accuracy (R2 value above 0.95). The trained empirical model applies to other macroencapsulated PCMs, contributing to the effective deployment of PCM-based thermal storage in buildings. Numerical models of envelopes with macroencapsulated PCMs will benefit from the results presented in this work, since the presented model determines the latent heat of whole PCM macrocapsules, unlike previous studies where only small quantities of PCM alone were analyzed.

Data-Driven Optimised XGBoost for Predicting the Performance of Axial Load Bearing Capacity of Fully Cementitious Grouted Rock Bolting Systems

This article investigates the application of eXtreme gradient boosting (XGBoost) and hybrid metaheuristics optimisation techniques to predict the axial load bearing capacity of fully grouted rock bolting systems. For this purpose, a comprehensive dataset of 72 pull-out tests was built, considering various influential parameters such as three water-to-grout (W/G) ratios, five ranges of curing time (CT), three different grout admixtures with two different fly ash (FA) contents, and two different diameter confinements (DCs). Additionally, to find out the effect of the mechanical behaviour of grouts on the performance of fully grouted rock bolting systems, seventy-two uniaxial compression strength (UCS) samples were cast and tested simultaneously with pull-out samples. The UCS samples were prepared with the same details as the pull-out samples to avoid any inconsistency. The results highlight that peak load values generally increase with longer curing times, lower W/G, and higher UCS and DC values. The main novelty of this paper lies in its data-driven approach, using various XGBoost models. This method offers a time-, cost-, and labour-efficient alternative to traditional experimental methods for predicting rock bolt performance. For this purpose, after building the dataset and dividing it randomly into two training and testing datasets, five different XGBoost models were developed: a standalone XGBoost model and four hybrid models incorporating Harris hawk optimisation (HHO), the jellyfish search optimiser (JSO), the dragonfly algorithm (DA), and the firefly algorithm (FA). These models were subsequently evaluated for their ability to predict peak load values. The results demonstrate that all models effectively predicted peak load values, but the XGBoost-JSO hybrid model demonstrated superior performance, achieving the highest R-squared coefficients of 0.987 and 0.988 for the training and testing datasets, respectively. Sensitivity analysis revealed that UCS values were the most influential parameter, while FA content had the least impact on the maximum peak load values of fully cementitious grouted rock bolts.

Deep learning-based segmentation of kidneys and renal cysts on T2-weighted MRI from patients with autosomal dominant polycystic kidney disease

BackgroundOur aim was to train and test a deep learning-based algorithm for automatically segmenting kidneys and renal cysts in patients with autosomal dominant polycystic kidney disease (ADPKD).MethodsWe retrospectively selected all ADPKD patients who underwent renal MRI with coronal T2-weighted imaging at our institution from 2008 to 2022. The 20 most recent examinations constituted the test dataset, to mimic pseudoprospective enrolment. The remaining ones constituted the training dataset to which eight normal renal MRIs were added. Kidneys and cysts ground truth segmentations were performed on coronal T2-weighted images by a junior radiologist supervised by an experienced radiologist. Kidneys and cysts of the 20 test MRIs were segmented by the algorithm and three independent human raters. Segmentations were compared using overlap metrics. The total kidney volume (TKV), total cystic volume (TCV), and cystic index (TCV divided by TKV) were compared using Bland–Altman analysis.ResultsWe included 164 ADPKD patients. Dice similarity coefficients ranged from 85.9% to 87.4% between the algorithms and the raters’ segmentations and from 84.2% to 86.2% across raters’ segmentations. For TCV assessment, the biases ± standard deviations (SD) were 3–19 ± 137–151 mL between the algorithm and the raters, and 22–45 ± 49–57 mL across raters. The algorithm underestimated TKV and TCV in two outliers with TCV > 2800 mL. For cystic index assessment, the biases ± SD were 2.5–6.9% ± 6.7–8.3% between the algorithm and the raters, and 2.1–9.4 ± 7.4–11.6% across raters.ConclusionThe algorithm’s performance fell within the range of inter-rater variability, but large TKV and TCV were underestimated.Relevance statementAccurate automated segmentation of the renal cysts will enable the large-scale evaluation of the prognostic value of TCV and cystic index in ADPKD patients. If these biomarkers are prognostic, then automated segmentation will facilitate their use in daily routine.Key PointsCystic volume is an emerging biomarker in ADPKD.The algorithm’s performance in segmenting kidneys and cysts fell within interrater variability.The segmentation of very large cysts, under-represented in the training dataset, needs improvement.Graphical

Reliable uncertainty estimation in emotion recognition in conversation using conformal prediction framework

Abstract Emotion recognition in conversation (ERC) faces two major challenges: biased predictions and poor calibration. Classifiers often disproportionately favor certain emotion categories, such as neutral, due to the structural complexity of classifiers, the subjective nature of emotions, and imbalances in training datasets. This bias results in poorly calibrated predictions where the model’s predicted probabilities do not align with the true likelihood of outcomes. To tackle these problems, we introduce the application of conformal prediction (CP) into ERC tasks. CP is a distribution-free method that generates set-valued predictions to ensure marginal coverage in classification, thus improving the calibration of models. However, inherent biases in emotion recognition models prevent baseline CP from achieving a uniform conditional coverage across all classes. We propose a novel CP variant, class spectrum conformation, which significantly reduces coverage bias in CP methods. The methodologies introduced in this study enhance the reliability of prediction calibration and mitigate bias in complex natural language processing tasks.

Deep learning approaches for protecting IoT devices in smart homes from MitM attacks

The primary objective of this paper is to enhance the security of IoT devices in Software-Defined Networking (SDN) environments against Man-in-the-Middle (MitM) attacks in smart homes using Artificial Intelligence (AI) methods as part of an Intrusion Detection and Prevention System (IDPS) framework. This framework aims to authenticate communication parties, ensure overall system and network security within SDN environments, and foster trust among users and stakeholders. The experimental analysis focuses on machine learning (ML) and deep learning (DL) algorithms, particularly those employed in Intrusion Detection Systems (IDS), such as Naive Bayes (NB), k-Nearest Neighbors (kNN), Random Forest (RF), and Convolutional Neural Networks (CNN). The CNN algorithm demonstrates exceptional performance on the training dataset, achieving 99.96% accuracy with minimal training time. It also shows favorable results in terms of detection speed, requiring only 1 s, and maintains a low False Alarm Rate (FAR) of 0.02%. Subsequently, the proposed framework was deployed in a testbed SDN environment to evaluate its detection capabilities across diverse network topologies, showcasing its efficiency compared to existing approaches.

An improved deep convolutional generative adversarial network for quantification of catechins in fermented black tea

The rapid and non-destructive quantification of catechins in fermented black tea is crucial for evaluating the quality of black tea. The combination of hyperspectral imaging and chemometrics has been applied for quantitative detection, but its performance is usually constrained by the limited dataset size. Targeted at the challenge of insufficient samples in regression analysis of catechins, this study proposes an improved deep convolutional generative adversarial network with labeling module, named as DCGAN-L for hyperspectral data augmentation. The DCGAN-L consists of the spectral and label generating modules. First the synthetic spectra were generated, and an indicator was proposed to evaluate their quality. Then, the corresponding label values were generated, including epicatechin gallate (ECG), epicatechin (EGC), catechin (C), and total catechin (CC). For label generating, the Euclidean distances between the synthetic spectrum and all true spectra were measured, followed by allocating weights for calculating the label values based on these distances. Subsequently, the training dataset was augmented with the generated synthetic data. The effect of data augmentation was finally evaluated based on two regression models of random forest (RF) and broad learning system (BLS) for the quantification of catechins. Compared with the results before data augmentation, the average R2 of RF and BLS models increased by 0.044 and 0.164, respectively. The proposed DCGAN-L model allows for the rapid, non-destructive quantification of catechins in black tea in the case of limited sample size.