Sample Imbalance Research Articles

FCAN : Speech emotion recognition network based on focused contrastive learning

Traditional speech emotion recognition (SER) models predominantly utilize cross-entropy loss for supervised optimization. Yet, given the limited size of SER datasets, relying solely on cross-entropy loss does not adequately capture the intricacies within the data. Moreover, conventional contrastive learning techniques, though improving performance, do not completely harness the inherent data structure across samples. To surmount these limitations, we introduce an innovative emotion recognition network founded on focused contrastive learning. Our model employs bidirectional gated recurrent unit (GRU) networks to seize critical contextual dependencies. By integrating an attention mechanism, we adeptly merge the contextual features of dialogues with individual utterance attributes, forming a comprehensive multi-stage feature fusion framework. This strategy extracts more elaborate emotional features from the constrained dataset. Additionally, by amalgamating the focal loss function into our training protocol, we mitigate imbalanced sample issues to a degree. We further refine our contrastive loss function by enacting a weighted strategy for intra-dialogue category assignments, where higher penalty weights are ascribed to harder-to-identify categories, thus incentivizing the model to concentrate on distinguishing subtler category differences. This refinement significantly bolsters the model’s proficiency in learning various emotional feature distributions, thereby enhancing its emotion recognition capabilities. Experimental validation on IEMOCAP and MELD benchmark datasets underpins our model’s superiority.

FIQ: A Fastener Inspection and Quantization Method Based on Mask FRCN

Rail-fastening components are essential for ensuring the safety of urban rail systems by securing rails to sleepers. Traditional inspection methods rely heavily on manual labor and are inefficient. This paper introduces a novel approach to address these inefficiencies and the challenges faced by computer vision-based inspections, such as missed detections due to imbalanced samples and limitations in conventional image segmentation techniques. Our approach transitions the industry’s focus from qualitative to a more precise quantitative analysis of rail-fastening components. We propose Mask-FRCN, an advanced image segmentation network that incorporates three key technological enhancements: the fully refined convolutional network module (FRCN),which refines the segmentation boundaries for SFC-type fasteners; the Channel-WiseKnowledge Distillation (CWD) algorithm, which boosts the model’s inference efficiency; and the FCRM methodology, which enhances the extraction capabilities for features specific to SFC-type fasteners. Furthermore, we introduce a fastener system inspection and quantization method based on the Mask FRCN method (FIQ), a novel technique for quantifying the condition of components by using image features, template matching with random forests, and a clustering calculation method derived from segmentation results. Experimental results validate that our method significantly surpasses existing techniques in accuracy, thereby offering a more efficient solution for inspecting rail-fastening components. The enhanced Mask-FRCN achieves a segmentation accuracy of 96.01% and a reduced network size of 36.1 M. Additionally, the FIQ method improves fault detection accuracy for SFC-type fasteners to 95.13%, demonstrating the efficacy and efficiency of our innovative approach.

An oversampling method based on Gaussian Mixture Model for multi-bolt looseness monitoring using Lamb waves`

The imbalance in the number of healthy and faulty samples poses a significant hindrance to the successful implementation of multi-bolt looseness monitoring and fault classification in engineering applications. The Synthetic Minority Oversampling Technique (SMOTE) is widely used in addressing data imbalance issues. However, traditional SMOTE methods and their improvements have not taken into account the problem of sensor network signals with multiple paths. In this paper, an improved oversampling technique, namely Gaussian Mixture Model (GMM)-SMOTE, is introduced. The proposed SMOTE method is based on GMM and aims to solve the sample imbalance problem under different bolt states and different sensor paths in sensor networks. The GMM is utilized to cluster the minority classes in multi-bolt looseness data. The assignment of weights to different clusters is based on the clustering density function, which is subsequently followed by oversampling. The KL distance is utilized to screen the SMOTE sample data, which improves the quality of the data by achieving inter-class balance and intra-class balance. To verify the effectiveness of the method, a multi-bolt looseness monitoring procedure is implemented. The experimental results demonstrate that the proposed method can improve the accuracy of diagnosing early-stage bolt looseness.

Semi-supervised soft sensor development based on dynamic dimensionality reduction-assisted large-scale pseudo label optimization and sample-weighted quality-relevant deep learning

Data-driven soft sensors have become popular tools for estimating critical quality variables in the process industry. However, in practical applications, it is very common that the unlabeled data are abundant but the labeled data are scarce, which poses a great challenge for developing high-performance data-based soft sensors. Thus, a dynamic dimensionality reduction-assisted large-scale pseudo label optimization method (DDR-LSPLO) is proposed for achieving sample expansion. This method repeatedly converts the LSPLO issue into a reduced-dimension pseudo label optimization problem with the low-confidence pseudo labels as new decision variables during the evolutionary optimization process. Meanwhile, to tackle the sample imbalance problem resulting from the inclusion of large-scale pseudo-labeled samples, a sample expansion and weighting-based quality-relevant autoencoder (SEWQAE) is developed for semi-supervised soft sensor modeling. The effectiveness and superiority of the proposed DDR-LSPLO and SEWQAE methods are verified through an industrial chlortetracycline (CTC) fermentation process and a simulated Tennessee Eastman (TE) chemical process.

DAJLENet: A neural network based on dual attention and joint learning for explainable heart failure adverse event prediction

Heart Failure (HF) is a common, high-risk, complex cardiovascular disease that is often accompanied by other diseases and has a high mortality rate. The use of electronic health records (EHRs) to predict the risk of adverse events, such as death, intensive care units (ICU) transfer, and readmission, in patients with HF is an important topic in the field of personalized treatment and prognosis. However, the heterogeneity and diversity of patient EHRs make it difficult for traditional machine learning models to adequately mine the hidden features and well train patient representations. In addition, significant sample imbalance and lack of labeled data can limit model performance. To address the above issues, this study proposes a neural network called DAJLENet for predicting adverse events in heart failure with explainability. The model obtains feature-level and visit-level patient representations through feature mining, feature embedding, and dual attention mechanisms. Based on the patient representation, a joint training method of classification task and contrastive learning task is used to learn shared representations to alleviate the sample imbalance problem and improve the model performance. Experimental results on four heart failure prediction tasks show that DAJLENet has an improvement of 1%–3% on various metrics compared to the most advanced models, and an improvement of more than 5% on unbalanced datasets.

Fault diagnosis of power equipment based on variational autoencoder and semi‐supervised learning

SummaryThe issue of fault diagnosis in power equipment is receiving increasing attention from scholars. Due to the important role played by bearings in power equipment, bearing faults have become the main factor causing the shutdown of wind turbines units. Therefore, this paper takes bearing equipment as an example for research. In order to solve the problem of insufficient and unbalanced fault sample data of wind turbines bearings, a fault diagnosis (FD) method based on variational autoencoder and semi‐supervised learning is proposed in this paper. Firstly, based on Label Propagation‐random forests (LP‐RFs) and a small number of labeled fault samples, a semi‐supervised learning algorithm is proposed to label the original data samples. Secondly, a small number of training samples are preprocessed by the variational autoencoder to reduce the imbalance of the fault samples. Then, the RFs‐based method is adopted to train the processed fault samples to obtain a mature FD classifier. Finally, the proposed method is applied to FD for bearings, and the results show that the proposed method can realize bearings fault diagnosis (BFD). And meanwhile, the proposed method can also be applied for fault diagnosis in power transmission and transformation systems.

Lithology identification based on ramified structure model using generative adversarial network for imbalanced data

Lithology identification plays a crucial role in the process of geological exploration and oil and gas development. Currently, the primary challenges in lithology identification include the inadequate differences of adjacent lithologies and the imbalance of lithology samples. This paper focuses on addressing these issues from three aspects. Firstly, a ramified lithology identification model based on neural network (CNN) and bi-directional long short-term memory (Bi-LSTM) is established to solve the problem of poor identification effect of adjacent lithology of single model. The task of lithology identification is split through a group network, and several identification models are designed to learn the subtle differences of adjacent lithologies. Secondly, to deal with the issue of ineffective feature learning due to limited samples, a data augmentation strategy is proposed. This strategy combines K-means SMOTE algorithm and Wasserstein conditional generative adversarial network (WCGAN). Finally, the k-nearest neighbor algorithm (KNN) is employed as a data screening method. Compared with other machine learning methods, the ramified lithology identification model demonstrates superior performance, particularly in identifying adjacent lithologies, on the well logging dataset from Kansas. Furthermore, the data augmentation strategy effectively enhances the identification accuracy of classes with fewer samples, thereby enhancing the overall accuracy of lithology identification.

An Accurate Recognition Method for Landslides Based on a Semi-Supervised Generative Adversarial Network: A Case Study in Lanzhou City

With the development of computer technology, landslide recognition based on machine learning methods has been widely applied in geological disaster management and research. However, in landslide identification, the problems of an insufficient number of samples and an imbalance of samples are often ignored; that is, landslide samples are much smaller than non-landslide samples. In order to solve this problem, taking the main urban area of Lanzhou City as an example, this paper proposes to construct a semi-supervised generated countermeasure network (SSGAN) model, which aims to achieve high performance with a limited number of labeled samples for precise landslide identification, and to help prevent and reduce the harm caused by disasters. In order to express the environmental characteristics of landslide development and the optical texture features of landslide occurrence, the study constructs three sets of samples to represent landslide features, including a landslide influencing factor sample set, a Sentinel-2A optical remote sensing sample set, a joint influencing factor and Sentinel-2A sample set. The three kinds of sample sets are transferred to SSGAN for training to form a comparative study. The results show that the joint sample set has excellent feature results in discriminator and generator. Through the experimental comparison, the model proposed in this paper is compared with the model without semi-supervised generated confrontation training. The experimental results show that the proposed method is better than the unsupervised adversarial learning model in terms of accuracy, F1 score, Kappa coefficient, and MIoU. A total of 160 landslides have been identified in the study area, with a total area of 10.328 km2, with an accuracy rate of 83%. Therefore, the generated results are accurate and reliable, and show that SSGAN can better distinguish landslides from non-landslides in an image, under the condition of obtaining a large number of unmarked environmental features; enhance the effect of landslide classification in complex geographical environment; and then put forward effective suggestions for the prevention and control of landslides and geological disasters in the main urban area of Lanzhou.

Multifactorial Tomato Leaf Disease Detection Based on Improved YOLOV5

Target detection algorithms can greatly improve the efficiency of tomato leaf disease detection and play an important technical role in intelligent tomato cultivation. However, there are some challenges in the detection process, such as the diversity of complex backgrounds and the loss of leaf symmetry due to leaf shadowing, and existing disease detection methods have some disadvantages in terms of deteriorating generalization ability and insufficient accuracy. Aiming at the above issues, a target detection model for tomato leaf disease based on deep learning with a global attention mechanism, TDGA, is proposed in this paper. The main idea of TDGA includes three aspects. Firstly, TDGA adds a global attention mechanism (GAM) after up-sampling and down-sampling, as well as in the SPPF module, to improve the feature extraction ability of the target object, effectively reducing the interference of invalid targets. Secondly, TDGA uses a switchable atrous convolution (SAConv) in the C3 module to improve the model’s ability to detect. Thirdly, TDGA adopts the efficient IoU loss (EIoU) instead of complete IoU loss (CIoU) to solve the ambiguous definition of aspect ratio and sample imbalance. In addition, the influences of different environmental factors such as single leaf, multiple leaves, and shadows on the performance of tomato disease detection are extensively experimented with and analyzed in this paper, which also verified the robustness of TDGA. The experimental results show that the average accuracy of TDGA reaches 91.40%, which is 2.93% higher than that of the original YOLOv5 network, which is higher than YOLOv5, YOLOv7, YOLOHC, YOLOv8, SSD, Faster R-CNN, RetinaNet and other target detection networks, so that TDGA can be utilized for the detection of tomato leaf disease more efficiently and accurately, even in complex environments.

An end-to-end method for predicting compound-protein interactions based on simplified homogeneous graph convolutional network and pre-trained language model

Identification of interactions between chemical compounds and proteins is crucial for various applications, including drug discovery, target identification, network pharmacology, and elucidation of protein functions. Deep neural network-based approaches are becoming increasingly popular in efficiently identifying compound-protein interactions with high-throughput capabilities, narrowing down the scope of candidates for traditional labor-intensive, time-consuming and expensive experimental techniques. In this study, we proposed an end-to-end approach termed SPVec-SGCN-CPI, which utilized simplified graph convolutional network (SGCN) model with low-dimensional and continuous features generated from our previously developed model SPVec and graph topology information to predict compound-protein interactions. The SGCN technique, dividing the local neighborhood aggregation and nonlinearity layer-wise propagation steps, effectively aggregates K-order neighbor information while avoiding neighbor explosion and expediting training. The performance of the SPVec-SGCN-CPI method was assessed across three datasets and compared against four machine learning- and deep learning-based methods, as well as six state-of-the-art methods. Experimental results revealed that SPVec-SGCN-CPI outperformed all these competing methods, particularly excelling in unbalanced data scenarios. By propagating node features and topological information to the feature space, SPVec-SGCN-CPI effectively incorporates interactions between compounds and proteins, enabling the fusion of heterogeneity. Furthermore, our method scored all unlabeled data in ChEMBL, confirming the top five ranked compound-protein interactions through molecular docking and existing evidence. These findings suggest that our model can reliably uncover compound-protein interactions within unlabeled compound-protein pairs, carrying substantial implications for drug re-profiling and discovery. In summary, SPVec-SGCN demonstrates its efficacy in accurately predicting compound-protein interactions, showcasing potential to enhance target identification and streamline drug discovery processes.Scientific contributionsThe methodology presented in this work not only enables the comparatively accurate prediction of compound-protein interactions but also, for the first time, take sample imbalance which is very common in real world and computation efficiency into consideration simultaneously, accelerating the target identification and drug discovery process.

SCV Filter: A Hybrid Deep Learning Model for SARS-CoV-2 Variants Classification

Background: The high mutability of severe acute respiratory syndrome coronavirus 2(SARS-CoV-2) makes it easy for mutations to occur during transmission. As the epidemic continues to develop, several mutated strains have been produced. Researchers worldwide are working on the effective identification of SARS-CoV-2. Objective: In this paper, we propose a new deep learning method that can effectively identify SARSCoV-2 Variant sequences, called SCVfilter, which is a deep hybrid model with embedding, attention residual network, and long short-term memory as components. Methods: Deep learning is effective in extracting rich features from sequence data, which has significant implications for the study of Coronavirus Disease 2019 (COVID-19), which has become prevalent in recent years. In this paper, we propose a new deep learning method that can effectively identify SARS-CoV-2 Variant sequences, called SCVfilter, which is a deep hybrid model with embedding, attention residual network, and long short-term memory as components. Results: The accuracy of the SCVfilter is 93.833% on Dataset-I consisting of different variant strains; 90.367% on Dataset-II consisting of data collected from China, Taiwan, and Hong Kong; and 79.701% on Dataset-III consisting of data collected from six continents (Africa, Asia, Europe, North America, Oceania, and South America). Conclusion: When using the SCV filter to process lengthy and high-homology SARS-CoV-2 data, it can automatically select features and accurately detect different variant strains of SARS-CoV-2. In addition, the SCV filter is sufficiently robust to handle the problems caused by sample imbalance and sequence incompleteness. Other: The SCVfilter is an open-source method available at https://github.com/deconvolutionw/SCVfilter.

Tri-Flow-YOLO: Counter helps to improve cross-domain object detection

The excellence of intelligent detection models has been widely recognized, but in terms of cross-domain scenes, they still face performance degradation and low accuracy. A multi-supervised Tri-Flow-YOLO model is proposed to improve the accuracy of objects with various scales under cross-domain conditions. Based on the full-supervised traditional detection branch of YOLOv5, another two mutually supporting task branches are designed intently. In brief, we add unsupervised adversarial classification training flow to the backend, to realize the feature alignment requirements and improve the cross-domain performance stability of the model. Meanwhile, a weakly-supervised object counting flow is proposed to improve the model's attention to all the objects and the detection ability is efficiently enforced. In addition, I-Mosaic and iCIOU are designed especially for small hard objects, enriching the positive samples during the training process. With the auxiliary of both improved strategies, the imbalance of positive and negative samples in the anchor-based model is relieved accordingly. The experimental results show that the improved Tri-Flow-YOLO model achieves 56.0 mAP in the Cityscapes→Foggy-Cityscapes task, and 49.8 mAP in the VOC→Clipart task.

Road marking defect detection based on CFG_SI_YOLO network

Existing road marking detection primarily focuses on the direction, position, and color of road markings. However, clear and accurate road markings directly impact issues such as directional guidance, lane selection, speed limits, and parking positions. Therefore, we introduce the CFG_SI_YOLO model for road marking defect detection. This model introduces a multi-channel CoordConv module, which enhances the network's focus on fine details based on the distribution characteristics of road markings defect. This helps prevent the loss of road marking information caused by model compression and pooling operations; Moreover, the model introduces the Focal-EIoU loss function to address the issue of imbalanced samples between easy and difficult cases. Additionally, the GELU activation function is incorporated to prevent gradient explosions, enhance the network's non-linear expressiveness, and improve the detection accuracy of the model. Finally, we add a similarity attention module to enhance the network's focus on the target, reduce interference from other objects, and mitigate false detection defects caused by inter-class similarity. Experiments conducted on a self-made dataset containing various types of road markings have shown that our approach achieved Precision, Recall, F1, IoU, and mAP of 85.7%, 85.8%, 85.7%, 75.1% and 82.8%, respectively. These results are significantly better than other methods, confirming the effectiveness and feasibility of our approach. Our code and results can be found on https://github.com/ly6660/Road_marking_line_code_data.

A hybrid demultiplexing strategy that improves performance and robustness of cell hashing.

Cell hashing, a nucleotide barcode-based method that allows users to pool multiple samples and demultiplex in downstream analysis, has gained widespread popularity in single-cell sequencing due to its compatibility, simplicity, and cost-effectiveness. Despite these advantages, the performance of this method remains unsatisfactory under certain circumstances, especially in experiments that have imbalanced sample sizes or use many hashtag antibodies. Here, we introduce a hybrid demultiplexing strategy that increases accuracy and cell recovery in multi-sample single-cell experiments. This approach correlates the results of cell hashing and genetic variant clustering, enabling precise and efficient cell identity determination without additional experimental costs or efforts. In addition, we developed HTOreader, a demultiplexing tool for cell hashing that improves the accuracy of cut-off calling by avoiding the dominance of negative signals in experiments with many hashtags or imbalanced sample sizes. When compared to existing methods using real-world datasets, this hybrid approach and HTOreader consistently generate reliable results with increased accuracy and cell recovery.

Not seeing the wood for the trees: Influences on random forest accuracy

Machine learning classifiers are increasingly widely used. This research note explores how a particular widely used classifier, the Random Forest, performs when faced with samples which are imbalanced and noisy data. Both are known to affect accuracy, but if their effects are independent or not has not been explored. Based on an experiment using synthetic data generated for the study we find that the effects of noise and sample balance interact with each other; classification accuracy is worse when faced with both noisy data and sample imbalance. This has implications for the use of RF in market research, but also for how methods to address either sample imbalance or noise are assessed.

MediDRNet: Tackling category imbalance in diabetic retinopathy classification with dual-branch learning and prototypical contrastive learning

Background and Objective:The classification of diabetic retinopathy (DR) aims to utilize the implicit information in images for early diagnosis, to prevent and mitigate the further worsening of the condition. However, existing methods are often limited by the need to operate within large, annotated datasets to show significant advantages. Additionally, the number of samples for different categories within the dataset needs to be evenly distributed, because the characteristic of sample imbalance distribution can lead to an excessive focus on high-frequency disease categories, while neglecting the less common but equally important disease categories. Therefore, there is an urgent need to develop a new classification method that can effectively alleviate the issue of sample distribution imbalance, thereby enhancing the accuracy of diabetic retinopathy classification. Methods:In this work, we propose MediDRNet, a dual-branch network model based on prototypical contrastive learning. This model adopts prototype contrastive learning, creating prototypes for different levels of lesions, ensuring they represent the core features of each lesion level. It classifies by comparing the similarity between data points and their category prototypes. Our dual-branch network structure effectively resolves the issue of category imbalance and improves classification accuracy by emphasizing subtle differences in retinal lesions. Moreover, our approach combines a dual-branch network with specific lesion-level prototypes for core feature representation and incorporates the convolutional block attention module for enhanced lesion feature identification. Results:Our experiments using both the Kaggle and UWF classification datasets have demonstrated that MediDRNet exhibits exceptional performance compared to other advanced models in the industry, especially on the UWF DR classification dataset where it achieved state-of-the-art performance across all metrics. On the Kaggle DR classification dataset, it achieved the highest average classification accuracy (0.6327) and Macro-F1 score (0.6361). Particularly in the classification tasks for minority categories of diabetic retinopathy on the Kaggle dataset (Grades 1, 2, 3, and 4), the model reached high classification accuracies of 58.08%, 55.32%, 69.73%, and 90.21%, respectively. In the ablation study, the MediDRNet model proved to be more effective in feature extraction from diabetic retinal fundus images compared to other feature extraction methods. Conclusions:This study employed prototype contrastive learning and bidirectional branch learning strategies, successfully constructing a grading system for diabetic retinopathy lesions within imbalanced diabetic retinopathy datasets. Through a dual-branch network, the feature learning branch effectively facilitated a smooth transition of features from the grading network to the classification learning branch, accurately identifying minority sample categories. This method not only effectively resolved the issue of sample imbalance but also provided strong support for the precise grading and early diagnosis of diabetic retinopathy in clinical applications, showcasing exceptional performance in handling complex diabetic retinopathy datasets. Moreover, this research significantly improved the efficiency of prevention and management of disease progression in diabetic retinopathy patients within medical practice. We encourage the use and modification of our code, which is publicly accessible on GitHub: https://github.com/ReinforceLove/MediDRNet.

An Improved Face Mask Detection Simulation Algorithm Based on YOLOv5 Model

This paper proposes an advanced approach for detecting faces with mask occlusion based on YOLOv5 to address various challenges encountered in face detection, including illumination blur and occlusion. The proposed methodology involves the integration of a convolutional block attention module into the backbone network and different network levels in the neck of the YOLOv5s. This approach enables the suppression of irrelevant features and emphasizes the identification of masked facial features. Replacing the conventional loss function with the Focal Loss function addresses the problem of sample imbalance. The enhanced YOLOv5s network was applied to detect mask-occluded faces. Empirical evaluations were conducted on the WIDER Face and AIZOO datasets. The simulation comparison results demonstrate that the proposed method achieves superior real-time detection performance, fulfilling the objective of developing a lightweight detection model.

MSGC-YOLO: An Improved Lightweight Traffic Sign Detection Model under Snow Conditions

Traffic sign recognition plays a crucial role in enhancing the safety and efficiency of traffic systems. However, in snowy conditions, traffic signs are often obscured by particles, leading to a severe decrease in detection accuracy. To address this challenge, we propose an improved YOLOv8-based model for traffic sign recognition. Initially, we introduce a Multi-Scale Group Convolution (MSGC) module to replace the C2f module in the YOLOv8 backbone. Data indicate that MSGC enhances detection accuracy while maintaining model lightweightness. Subsequently, to improve the recognition ability for small targets, we introduce an enhanced small target detection layer, which enhances small target detection accuracy while reducing parameters. In addition, we replaced the original BCE loss with the improved EfficientSlide loss to improve the sample imbalance problem. Finally, we integrate Deformable Attention into the model to improve the detection efficiency and performance of complex targets. The resulting fused model, named MSGC-YOLOv8, is evaluated on an enhanced dataset of snow-covered traffic signs. Experimental results show that the MSGC-YOLOv8 model is used for snow road traffic sign recognition. Compared with the YOLOv8n model mAP@0.5:0.95, mAP@0.5:0.95 is increased by 17.7% and 18.1%, respectively, greatly improving the detection accuracy. Compared with the YOLOv8s model, while the parameters are reduced by 59.6%, mAP@0.5 only loses 1.5%. Considering all aspects of the data, our proposed model shows high detection efficiency and accuracy under snowy conditions.

A fault diagnosis framework based on heterogeneous ensemble learning for air conditioning chiller with unbalanced samples

Big data-based air conditioning fault diagnosis research has developed rapidly in recent years, but in actual engineering, the fault sample size of air conditioning systems is much smaller than the normal sample size, and the resulting sample imbalance problem makes conventional data-driven diagnostic methods based on low accuracy and poor stability. In order to solve the problem of unbalanced fault diagnosis of air-conditioning chillers, this paper proposes an integrated learning-based diagnostic model, which achieves diagnosis by combining multiple base models and by majority voting. The method uses four classification models, namely, random forest model, decision tree model, k nearest neighbor model, and isomorphic integration model, as base classifiers, and synthesizes the four base classifiers into a heterogeneous integration algorithmic model (IMV) through integrated learning, and performs diagnostic detection of seven types of typical faults of chiller units using the majority voting method of integrated learning. The effectiveness of the proposed algorithm is verified on the RP-1043 dataset, and the experimental results show that the accuracy of the heterogeneous integrated algorithm model (IMV) can reach 96.87%, which is a significant improvement compared with the accuracy of the other four base classifier models (81.04%–96.25%). Therefore, the integrated learning model has some application prospects in fault diagnosis when targeting unbalanced datasets.

Sample-imbalanced wafer map defects classification based on auxiliary classifier denoising diffusion probability model

This paper research on the classification task of wafer map defects under imbalanced sample. To improve the number of imbalanced wafer maps, a denoising diffusion probabilistic model with auxiliary classifier is proposed to generate new wafer maps. The developed model has an U-net structure with multiple inputs and outputs, taking wafer images, noise degree and category features as inputs, and outputs predicted noise and prediction category. When performing data augmentation, the trained model gradually removes prediction noise from the initial random noise and obtains new wafers. Finally, the balanced wafer defect maps are classified using a residual neural network. The proposed auxiliary classifier denoising diffusion probabilistic model and residual neural networks (ACDDPM-ResNet) method has validated on the MIR-WM811K dataset and MixedWM38 dataset, and the defect classification results of the imbalanced wafer map are remarkable improved after data augmentation. In addition, the paper also discusses and analyzes the influence of the maximum noise addition steps and the data augmentation size on the accuracy of wafer classification. It is further validated that the proposed wafer map classification method based on data augmentation can solve the classification problem caused by sample imbalance.