Class Probability Distributions Research Articles

A Vision-Language Model-Based Traffic Sign Detection Method for High-Resolution Drone Images: A Case Study in Guyuan, China.

As a fundamental element of the transportation system, traffic signs are widely used to guide traffic behaviors. In recent years, drones have emerged as an important tool for monitoring the conditions of traffic signs. However, the existing image processing technique is heavily reliant on image annotations. It is time consuming to build a high-quality dataset with diverse training images and human annotations. In this paper, we introduce the utilization of Vision-language Models (VLMs) in the traffic sign detection task. Without the need for discrete image labels, the rapid deployment is fulfilled by the multi-modal learning and large-scale pretrained networks. First, we compile a keyword dictionary to explain traffic signs. The Chinese national standard is used to suggest the shape and color information. Our program conducts Bootstrapping Language-image Pretraining v2 (BLIPv2) to translate representative images into text descriptions. Second, a Contrastive Language-image Pretraining (CLIP) framework is applied to characterize not only drone images but also text descriptions. Our method utilizes the pretrained encoder network to create visual features and word embeddings. Third, the category of each traffic sign is predicted according to the similarity between drone images and keywords. Cosine distance and softmax function are performed to calculate the class probability distribution. To evaluate the performance, we apply the proposed method in a practical application. The drone images captured from Guyuan, China, are employed to record the conditions of traffic signs. Further experiments include two widely used public datasets. The calculation results indicate that our vision-language model-based method has an acceptable prediction accuracy and low training cost.

GANFAT: Robust federated adversarial learning with label distribution skew

As privacy concerns and regulatory constraints on data protection continue to grow, the distribution of collected data has become more dispersed, resembling a ”data silo” style. To harness these data effectively without exchanging raw data, federated learning has emerged as a prominent solution. However, distributions of user-generated data often exhibit imbalances between devices and labels, which adversely affect model performance, especially in the presence of adversarial attacks, making models more susceptible. To address the challenge of balancing natural accuracy and robustness in federated training, especially under skewed label distribution scenarios, we propose a novel approach based on Generative Adversarial Networks for Federated Adversarial Training (GANFAT). GANFAT leverages GAN to enhance the authenticity and effectiveness of adversarial samples and addresses label distribution skew issues by incorporating class probability distribution information. Through a balanced interplay of natural accuracy loss and adversarial loss, GANFAT demonstrates significantly superior performance across multiple datasets under various settings compared to other frameworks. Particularly on the SVHN dataset, GANFAT achieves a remarkable 9.30% enhancement in robustness against FGSM attacks compared to the best baseline method (FedRBN). On the CIFAR-100 dataset, GANFAT showcases a noteworthy 6.68% improvement in natural accuracy compared to the best baseline method (CalFAT). GANFAT provides a powerful solution for confronting diverse attacks, yielding models comparable to those produced by centralized training. Experimental results underscore GANFAT’s outstanding performance, offering a robust solution for scenarios characterized by uneven data distribution and adversarial attacks.

An Active Learning Approach for Detecting Customer Induced Damages in Motherboards with Deep Neural Networks

Identifying Customer Induced Damage (CID) is a key part in warranty programs of electronics manufacturers. CID is defined as any damage in the unit performed by an unauthorized person including the customer in a Printed Circuit Board (PCB). In such cases, damaged units are not covered by warranty. The inspection of CIDs is usually performed by humans which may be costly and error prone. Modern computer vision techniques for object detection using deep neural networks can automatically and accurately detect CIDs on PCBs. The training of such networks requires a large labeled dataset of image examples of CIDs. Daily, hardware factories and repair centers generate hundreds of unlabeled images. Labeling them manually is laborious and time-consuming. Therefore, it is crucial to label the minimum amount of images such that the trained neural network can achieve comparable accuracy as if it were trained with the whole dataset. To this end, we propose an active learning approach that selects the most informative images for the object detector. For that, our approach is based on the uncertainty of the object detector, i.e., it selects new images based on class probability distribution given by the object detector. Also, we tackle some challenges that are intrinsic to this problem: i) it is a multiclass object detection problem since there are many types of defects; ii) there is a class accuracy imbalance; iii) there is a focus on recall, e.g. false positives are less harmful than false negatives, and iv) there are many images with no object which should not be selected for labeling. We evaluate this approach by using it to iteratively sample data, train and evaluate a model, and compare it with randomly sampled data. The results show that our method consistently outperforms random sampling by an average margin of 21.6%, proving to be a viable alternative for reducing the labeling cost and increasing detection accuracy in this domain.

Label-Aware Calibration and Relation-Preserving in Visual Intention Understanding.

Visual intention understanding is a challenging task that explores the hidden intention behind the images of publishers in social media. Visual intention represents implicit semantics, whose ambiguous definition inevitably leads to label shifting and label blemish. The former indicates that the same image delivers intention discrepancies under different data augmentations, while the latter represents that the label of intention data is susceptible to errors or omissions during the annotation process. This paper proposes a novel method, called Label-aware Calibration and Relation-preserving (LabCR) to alleviate the above two problems from both intra-sample and inter-sample views. First, we disentangle the multiple intentions into a single intention for explicit distribution calibration in terms of the overall and the individual. Calibrating the class probability distributions in augmented instance pairs provides consistent inferred intention to address label shifting. Second, we utilize the intention similarity to establish correlations among samples, which offers additional supervision signals to form correlation alignments in instance pairs. This strategy alleviates the effect of label blemish. Extensive experiments have validated the superiority of the proposed method LabCR in visual intention understanding and pedestrian attribute recognition. Code is available at https://github.com/ShiQingHongYa/LabCR.

PVCsNet : A Specialized Artificial Intelligence-Based Model to Classify Premature Ventricular Contractions from ECG Images.

Premature ventricular complexes (PVCs) are irregularities in heart rhythm where the ventricles contract earlier than expected, disrupting the normal cardiac cycle. Identifying the origin of PVCs before surgery is crucial as it can reduce operation duration, lower radiation exposure, and potentially enhance ablation success rates. Current detection methods face limitations in accuracy and data processing, often requiring large datasets and complex interpretations. This study presents PVCsNet, a deep-learning network specifically designed for classifying premature ventricular complexes (PVCs) in ECG images. It incorporates residual structures and attention mechanisms to enhance classification performance. PVCsNet consists of four 3×3 convolutional layers as feature extractors, followed by residual connections and attention blocks. This design enables the network to map image features to class probability distributions, enhancing performance even with limited data. Our experimental results demonstrate that using the SE Block with MaxPool and a ratio of 4, PVCsNet achieves an overall accuracy of 94.49%, with high precision in critical categories and a moderate parameter size. We successfully categorize the data into six distinct classes based on their origin locations in the heart: right ventricular outflow tract (RVOT), left ventricular outflow tract (LVOT), papillary muscle (PM), valvular annulus (VA), summit, and His-Purkinje system (HPS). Among these, RVOT is the most common and crucial origin of PVCs. PM and HPS are also significant origins due to their clinical implications. This study demonstrates the potential of PVCsNet in clinical diagnostics, providing promising results in classifying ECG images and contributing to future medical research and diagnosis.

Deep Gated Multi-modal Fusion for Image Privacy Prediction

With the rapid development of technologies in mobile devices, people can post their daily lives on social networking sites such as Facebook, Flickr, and Instagram. This leads to new privacy concerns due to people’s lack of understanding that private information can be leaked and used to their detriment. Image privacy prediction models are developed to predict whether images contain sensitive information (private images) or are safe to be shared online (public images). Despite significant progress on this task, there are still some crucial problems that remain to be solved. Firstly, images’ content and tags are found to be useful modalities to automatically predict images’ privacy. To date, most image privacy prediction models use single modalities (image-only or tag-only), which limits their performance. Secondly, we observe that current image privacy prediction models are surprisingly vulnerable to even small perturbations in the input data. Attackers can add small perturbations to input data and easily damage a well-trained image privacy prediction model. To address these challenges, in this article, we propose a new decision-level Gated multi-modal fusion (GMMF) approach that fuses object, scene, and image tags modalities to predict privacy for online images. In particular, the proposed approach identifies fusion weights of class probability distributions generated by single-modal classifiers according to their reliability of the privacy prediction for each target image in a sample-by-sample manner and performs a weighted decision-level fusion, so that modalities with high reliability are assigned with higher fusion weights while ones with low reliability are restrained with lower fusion weights. The results of our experiments show that the gated multi-modal fusion network effectively fuses single modalities and outperforms state-of-the-art models for image privacy prediction. Moreover, we perform adversarial training on our proposed GMMF model using multiple types of noise on input data (i.e., images and/or tags). When some modalities are failed by input data with noise attacks, our approach effectively utilizes clean modalities and minimizes negative influences brought by degraded ones using fusion weights, achieving significantly stronger robustness over traditional fusion methods for image privacy prediction. The robustness of our GMMF model against data noise can even be generalized to more severe noise levels. To the best of our knowledge, we are the first to investigate the robustness of image privacy prediction models against noise attacks. Moreover, as the performance of decision-level multi-modal fusion depends highly on the quality of single-modal networks, we investigate self-distillation on single-modal privacy classifiers and observe that transferring knowledge from a trained teacher model to a student model is beneficial in our proposed approach.

MCIBI++: Soft Mining Contextual Information Beyond Image for Semantic Segmentation.

Co-occurrent visual pattern makes context aggregation become an essential paradigm for semantic segmentation. The existing studies focus on modeling the contexts within image while neglecting the valuable semantics of the corresponding category beyond image. To this end, we propose a novel soft mining contextual information beyond image paradigm named MCIBI++ to further boost the pixel-level representations. Specifically, we first set up a dynamically updated memory module to store the dataset-level distribution information of various categories and then leverage the information to yield the dataset-level category representations during network forward. After that, we generate a class probability distribution for each pixel representation and conduct the dataset-level context aggregation with the class probability distribution as weights. Finally, the original pixel representations are augmented with the aggregated dataset-level and the conventional image-level contextual information. Moreover, in the inference phase, we additionally design a coarse-to-fine iterative inference strategy to further boost the segmentation results. MCIBI++ can be effortlessly incorporated into the existing segmentation frameworks and bring consistent performance improvements. Also, MCIBI++ can be extended into the video semantic segmentation framework with considerable improvements over the baseline. Equipped with MCIBI++, we achieved the state-of-the-art performance on seven challenging image or video semantic segmentation benchmarks.

Efficient Data-Driven Machine Learning Models for Cardiovascular Diseases Risk Prediction.

Cardiovascular diseases (CVDs) are now the leading cause of death, as the quality of life and human habits have changed significantly. CVDs are accompanied by various complications, including all pathological changes involving the heart and/or blood vessels. The list of pathological changes includes hypertension, coronary heart disease, heart failure, angina, myocardial infarction and stroke. Hence, prevention and early diagnosis could limit the onset or progression of the disease. Nowadays, machine learning (ML) techniques have gained a significant role in disease prediction and are an essential tool in medicine. In this study, a supervised ML-based methodology is presented through which we aim to design efficient prediction models for CVD manifestation, highlighting the SMOTE technique's superiority. Detailed analysis and understanding of risk factors are shown to explore their importance and contribution to CVD prediction. These factors are fed as input features to a plethora of ML models, which are trained and tested to identify the most appropriate for our objective under a binary classification problem with a uniform class probability distribution. Various ML models were evaluated after the use or non-use of Synthetic Minority Oversampling Technique (SMOTE), and comparing them in terms of Accuracy, Recall, Precision and an Area Under the Curve (AUC). The experiment results showed that the Stacking ensemble model after SMOTE with 10-fold cross-validation prevailed over the other ones achieving an Accuracy of 87.8%, Recall of 88.3%, Precision of 88% and an AUC equal to 98.2%.

Class probability distribution based maximum entropy model for classification of datasets with sparse instances

Due to the digital revolution, the amount of data to be processed is growing every day. One of the more common functions used to process these data is classification. However, the results obtained by most existing classifiers are not satisfactory, as they often depend on the number and type of attributes within the datasets. In this paper, a maximum entropy model based on class probability distribution is proposed for classifying data in sparse datasets with fewer attributes and instances. Moreover, a new idea of using Lagrange multipliers is suggested for estimating class probabilities in the process of class label prediction. Experimental analysis indicates that the proposed model has an average accuracy of 89.9% and 86.93% with 17 and 36 datasets. Besides, statistical analysis of the results indicates that the proposed model offers greater classification accuracy for over 50% of datasets with fewer attributes and instances than other competitors.

Classification of hyperspectral images with copulas

In the last decade, supervised learning methods for the classification of remotely sensed images (RSI) have grown significantly, especially for hyper-spectral (HS) images. Recently, deep learning-based approaches have produced encouraging results for the land cover classification of HS images. In particular, the Convolutional Neural Networks (CNN) and Recurrent Neural Networks (RNN) have shown good performance. However, these methods suffer for the problem of the hyperparameter optimization or tuning that requires a high computational cost; moreover, they are sensitive to the number of observations in the learning phase. In this work we propose a novel supervised learning algorithm based on the use of copula functions for the classification of hyperspectral images called CopSCHI (Copula Supervised Classification of Hyperspectral Images). In particular, we start with a dimensionality reduction technique based on Singular Value Decomposition (SVD) in order to extract a small number of relevant features that best preserve the characteristics of the original image. Afterward, we learn the classifier through a dynamic choice of copulas that allows us to identify the distribution of the different classes within the dataset. The use of copulas proves to be a good choice due to their ability to recognize the probability distribution of classes and hence an accurate final classification with low computational cost can be conducted. The proposed approach was tested on two benchmark datasets widely used in literature. The experimental results confirm that CopSCHI outperforms the state-of-the-art methods considered in this paper as competitors.

ABOUT MEASURING THE PARAMETERS OF THE SUBJECT-METHODICAL COMPETENCE OF BIOLOGY TEACHERS IN GENERAL SECONDARY EDUCATION INSTITUTIONS

Taking into account the professional standard of the teacher in the general secondary education institution, the concept of measuring indicators of the level of the subject component of the subject-methodical competence of the biology teacher in general secondary education institutions was developed. The corresponding concept applies the idea of a probabilistic sampling approach, which has already been tested by the authors, for evaluating parameters of the educators professional competence. This idea consists in selecting a hypothetical set of tasks corresponding to a teachers potential professional activity with the appropriate probability distribution of classes of tasks and construction of a representative sample of tasks. The indicator of the corresponding level of competence is considered as the probability to successfully solve any task from the general population, therefore it is theoretically calculated according to the formula of full probability, and the indicator estimate is the average weighted relative frequency of the removed classes of tasks. When forming the profile of the general population of tasks to assess subject competence, the authors relied on the classification of tasks types of the Ukrainian Centre for Education Quality Assessment, which reflects the structure of the school course in Biology and the Principles of Ecology. When approbating the results, 231 respondents were involved, the sample of teachers is representative, reflecting the age and qualification of biology teachers in the region. The research was carried out at professional development courses for teaching staff. Measurements showed a sufficiently high level of the indicator of the subject competence of biology teachers. According to the obtained estimates, the average value is 81.1%; the distribution of points is abnormal with a clearly expressed right asymmetry. The hypothesis about the independence of the results from teaching experience was not confirmed, the deviation of this hypothesis is at the 5% significance level. A more detailed analysis shows higher results for teachers with above average experience. The corresponding analysis provides grounds for differentiation in the learning process at teacher training courses depending on the pedagogical experience.

Investigating the relationships between class probabilities and users’ appropriate trust in computer vision classifications of ambiguous images

The large-scale adoption of systems that automate classifications using Machine Learning (ML) algorithms raises pressing challenges as they support or make decisions with profound consequences for human beings. It is important to understand how users’ trust is affected by ML models’ suggestions, even when those models are wrong. Many research efforts have focused on the user’s ability to interpret what a model has learned. In this paper, we seek to understand another aspect of ML interpretability: whether and how the presence of classification probabilities and their different distributions are related to users’ trust in model outcomes, especially in ambiguous instances. To this end, we conducted two online surveys in which we asked participants to evaluate their agreement with image classifications of pictures of animals made by an ML model. In the first, we analyze their trust before and after presenting them the model classification probabilities. In the second, we investigate the relationships between class probability distributions and users’ trust in the model. We found that, in some cases, the additional information is correlated with undue trust in the model’s classifications. However, in others, they are associated with inappropriate skepticism. • Additional information about model decisions may be detrimental to users’ trust. • Disclosure of additional information about model decisions may lead to overtrust. • Distributions of class probabilities may have an impact on users’ appropriate trust.

Automated bone marrow cytology using deep learning to generate a histogram of cell types

BackgroundBone marrow cytology is required to make a hematological diagnosis, influencing critical clinical decision points in hematology. However, bone marrow cytology is tedious, limited to experienced reference centers and associated with inter-observer variability. This may lead to a delayed or incorrect diagnosis, leaving an unmet need for innovative supporting technologies.MethodsWe develop an end-to-end deep learning-based system for automated bone marrow cytology. Starting with a bone marrow aspirate digital whole slide image, our system rapidly and automatically detects suitable regions for cytology, and subsequently identifies and classifies all bone marrow cells in each region. This collective cytomorphological information is captured in a representation called Histogram of Cell Types (HCT) quantifying bone marrow cell class probability distribution and acting as a cytological patient fingerprint.ResultsOur system achieves high accuracy in region detection (0.97 accuracy and 0.99 ROC AUC), and cell detection and cell classification (0.75 mean average precision, 0.78 average F1-score, Log-average miss rate of 0.31).ConclusionsHCT has potential to eventually support more efficient and accurate diagnosis in hematology, supporting AI-enabled computational pathology.

Improving Post-Processing of Audio Event Detectors Using Reinforcement Learning

We apply post-processing to the class probability distribution outputs of audio event classification models and employ reinforcement learning to jointly discover the optimal parameters for various stages of a post-processing stack, such as the classification thresholds and the kernel sizes of median filtering algorithms used to smooth out model predictions. To achieve this we define a reinforcement learning environment where: 1) a state is the class probability distribution provided by the model for a given audio sample, 2) an action is the choice of a candidate optimal value for each parameter of the post-processing stack, 3) the reward is based on the classification accuracy metric we aim to optimize, which is the audio event-based macro F1-score in our case. We apply our post-processing to the class probability distribution outputs of two audio event classification models submitted to the DCASE Task4 2020 challenge. We find that by using reinforcement learning to discover the optimal per-class parameters for the post-processing stack that is applied to the outputs of audio event classification models, we can improve the audio event-based macro F1-score (the main metric used in the DCASE challenge to compare audio event classification accuracy) by 4-5% compared to using the same post-processing stack with manually tuned parameters.

Industrial fault diagnosis based on active learning and semi-supervised learning using small training set

Industrial fault diagnosis has been investigated for many years, and many approaches have been proposed to identify industrial faults. However, the size of the actual training set is usually small, which severely degrades the performance of existing fault diagnostic models. To solve this problem, a new fault diagnosis method was proposed based on active and semi-supervised learning. First, uncertain unlabelled samples were selected by estimating the first two values in the class probability distribution of the samples. They were labelled by experts to update the performance of the models learned from a small training set. Second, heterogeneous classifiers were adopted to increase the diversity of the base classifiers, and noise samples were deleted using a sample pruning operation. The weights of the base classifier were designed for ensemble learning based on the test error rates. An evaluation using the Case Western Reserve University and Intelligent Maintenance Systems data showed that the performance of the proposed method was better than those of the other methods in the experiment. The experimental results showed that this study provided a promising and useful methodology for fault diagnosis under a small training set.

Hybrid feature selection model for classification of lung disorders

Feature selection in computer aided diagnosis is now becoming a challenging part in the classification of lung diseases. This is because, it needs to deliver results with improved accuracy and it also requires a greater number of features for analysis. The major demerit of widely utilized single-objective feature selection (FS) algorithm is that it proffers only a single optimum solution for a feature set. Here, a hybridized multi-objective particle swarm optimization with a local Tabu search (MOPSO-TS) algorithm is proposed to overcome the above demerit of the traditional single objective algorithm by producing a bag of optimum solutions which trade disparate objectives amongst themselves. The work is validated against a feature set which consists of GLCM features, shape features and GLRLM (gray-level run length matrix) extracted from lung chest tomography (CT) images. Classification is done using k-nearest neighbor with class probability and normal distribution (ND). The proposed FS method’s performance is analyzed against widely used bio-inspired FS methods such as Firefly, Particle Swarm Optimization along with Bee Colony Optimization algorithms. The numerical analysis of this model indicates that the proposed hybrid FS algorithm achieves improved performance compared to a single objective optimization algorithm in respect of specificity, accuracy, F-score, precision, sensitivity and error rate. The proposed algorithm obtains the result of 90.588% in both and specificity accuracy rate, (77.143) precision, 87.667 sensitivity rate and error rate of 0.1 which are higher on considering the other prevailing methodologies

Hyperspectral Open Set Classification towards Deep Networks Based on Boxplot

Recently, hyperspectral imaging (HSI) supervised classification has achieved an astonishing performance by using deep learning. However, most of them take the ideal assumption of ‘closed set’, where all testing classes have been known during training. In fact, in the real world, new classes unseen in training may appear during testing. Obviously, traditional supervised classification methods cannot operate correctly in the real world, which requires classifiers not only to classify known classes, but reject the unknown in order to avoid false positives. This challenge is called ‘open set classification’(OSC). Considering the increased applications of deep learning in the real world, rejecting unknown classes during classification is of vital importance. To tackle it, we present a simple but effective HSI OSC method toward deep networks. In this method, we tighten the decision boundaries of SoftMax function of the last layer of the deep networks by using boxplots to analysis the statistical characteristics of the probability distribution of known classes and generate a proper rejection threshold for each known class. To test the performance of the proposed HSI OSC method, experiments are conducted on three HSI datasets. The results show that the proposed method outperforms existing state-of-the-art HSI OSC methods.

EasiEdge: A Novel Global Deep Neural Networks Pruning Method for Efficient Edge Computing

Deep neural networks (DNNs) have shown tremendous success in many areas, such as signal processing, computer vision, and artificial intelligence. However, the DNNs require intensive computation resources, hindering their practical applications on the edge devices with limited storage and computation resources. Filter pruning has been recognized as a useful technique to compress and accelerate the DNNs, but most existing works tend to prune filters in a layerwise manner, facing some significant drawbacks. First, the layerwise pruning methods require prohibitive computation for per-layer sensitivity analysis. Second, layerwise pruning suffers from the accumulation of pruning errors, leading to performance degradation of pruned networks. To address these challenges, we propose a novel global pruning method, namely, EasiEdge, to compress and accelerate the DNNs for efficient edge computing. More specifically, we introduce an alternating direction method of multipliers (ADMMs) to formulate the pruning problem as a performance improving subproblem and a global pruning subproblem. In the global pruning subproblem, we propose to use information gain (IG) to quantify the impact of filters removal on the class probability distributions of network output. Besides, we propose a Taylor-based approximate algorithm (TBAA) to efficiently calculate the IG of filters. Extensive experiments on three data sets and two edge computing platforms verify that our proposed EasiEdge can efficiently accelerate DNNs on edge computing platforms with nearly negligible accuracy loss. For example, when EasiEdge prunes 80% filters in VGG-16, the accuracy drops by 0.22%, but inference latency on CPU of Jetson TX2 decreases from 76.85 to 8.01 ms.

Information Bottleneck Classification in Extremely Distributed Systems.

We present a new decentralized classification system based on a distributed architecture. This system consists of distributed nodes, each possessing their own datasets and computing modules, along with a centralized server, which provides probes to classification and aggregates the responses of nodes for a final decision. Each node, with access to its own training dataset of a given class, is trained based on an auto-encoder system consisting of a fixed data-independent encoder, a pre-trained quantizer and a class-dependent decoder. Hence, these auto-encoders are highly dependent on the class probability distribution for which the reconstruction distortion is minimized. Alternatively, when an encoding–quantizing–decoding node observes data from different distributions, unseen at training, there is a mismatch, and such a decoding is not optimal, leading to a significant increase of the reconstruction distortion. The final classification is performed at the centralized classifier that votes for the class with the minimum reconstruction distortion. In addition to the system applicability for applications facing big-data communication problems and or requiring private classification, the above distributed scheme creates a theoretical bridge to the information bottleneck principle. The proposed system demonstrates a very promising performance on basic datasets such as MNIST and FasionMNIST.

Improving data and model quality in crowdsourcing using cross-entropy-based noise correction

Crowdsourcing services provide a fast, efficient, and cost-effective approach to obtaining labeled data, particularly for human-like tasks. In a crowdsourcing scenario, after ground truth inference methods have been employed to obtain integrated instance labels, label noise remains present in the integrated labels. Label noise handling techniques can then be implemented to mitigate the effects of this noise. In this study, we propose a Cross-Entropy-based Noise Correction (CENC) method for crowdsourcing. CENC uses the entropies of the label distributions generated from multiple noisy label sets to filter noisy instances. It then exploits the cross-entropies between each possible true class probability distribution and each predicted class probability distribution to rectify the noisy instances. Using both simulated benchmark data and real-world crowdsourced data, we show that CENC outperforms all other existing state-of-the-art noise correction methods.