CoDC: Accurate Learning with Noisy Labels via Disagreement and Consistency

Inspired by the biological nervous system, deep neural networks (DNNs) are able to achieve remarkable performance in various tasks. However, they struggle to handle label noise, which can poison the memorization effects of DNNs. Co-teaching-based methods are popular in learning with noisy labels. These methods cross-train two DNNs based on the small-loss criterion and employ a strategy using either "disagreement" or "consistency" to obtain the divergence of the two networks. However, these methods are sample-inefficient for generalization in noisy scenarios. In this paper, we propose CoDC, a novel Co-teaching-basedmethod for accurate learning with label noise via both Disagreement and Consistency strategies. Specifically, CoDC maintains disagreement at the feature level and consistency at the prediction level using a balanced loss function. Additionally, a weighted cross-entropy loss is proposed based on information derived from the historical training process. Moreover, the valuable knowledge involved in "large-loss" samples is further developed and utilized by assigning pseudo-labels. Comprehensive experiments were conducted on both synthetic and real-world noise and under various noise types. CoDC achieved 72.81% accuracy on the Clothing1M dataset and 76.96% (Top1) accuracy on the WebVision1.0 dataset. These superior results demonstrate the effectiveness and robustness of learning with noisy labels.

Leveraging Multi-Annotator Label Uncertainties as Privileged Information for Acute Respiratory Distress Syndrome Detection in Chest X-ray Images

Acute Respiratory Distress Syndrome (ARDS) is a life-threatening lung injury for which early diagnosis and evidence-based treatment can improve patient outcomes. Chest X-rays (CXRs) play a crucial role in the identification of ARDS; however, their interpretation can be difficult due to non-specific radiological features, uncertainty in disease staging, and inter-rater variability among clinical experts, thus leading to prominent label noise issues. To address these challenges, this study proposes a novel approach that leverages label uncertainty from multiple annotators to enhance ARDS detection in CXR images. Label uncertainty information is encoded and supplied to the model as privileged information, a form of information exclusively available during the training stage and not during inference. By incorporating the Transfer and Marginalized (TRAM) network and effective knowledge transfer mechanisms, the detection model achieved a mean testing AUROC of 0.850, an AUPRC of 0.868, and an F1 score of 0.797. After removing equivocal testing cases, the model attained an AUROC of 0.973, an AUPRC of 0.971, and an F1 score of 0.921. As a new approach to addressing label noise in medical image analysis, the proposed model has shown superiority compared to the original TRAM, Confusion Estimation, and mean-aggregated label training. The overall findings highlight the effectiveness of the proposed methods in addressing label noise in CXRs for ARDS detection, with potential for use in other medical imaging domains that encounter similar challenges.

Robust Learning with Noisy Ship Trajectories by Adaptive Noise Rate Estimation

Ship trajectory classification is of great significance for shipping analysis and marine security governance. However, in order to cover up their illegal fishing or espionage activities, some illicit ships will forge the ship type information in the Automatic Identification System (AIS), and this label noise will significantly impact the algorithm's classification accuracy. Sample selection is a common and effective approach in the field of learning from noisy labels. However, most of the existing methods based on sample selection need to determine the noise rate of the data through prior means. To address these issues, we propose a noise rate adaptive learning mechanism that operates without prior conditions. This mechanism is integrated with the robust training paradigm JoCoR (joint training with co-regularization), giving rise to a noise rate adaptive learning robust training paradigm called A-JoCoR. Experimental results on real-world trajectories provided by the Danish Maritime Authority verified the effectiveness of A-JoCoR. It not only realizes the adaptive learning of the data noise rate during the training process, but also significantly improves the classification performance compared with the original method.

Impact of Label Noise on the Learning Based Models for a Binary Classification of Physiological Signal

Label noise is omnipresent in the annotations process and has an impact on supervised learning algorithms. This work focuses on the impact of label noise on the performance of learning models by examining the effect of random and class-dependent label noise on a binary classification task: quality assessment for photoplethysmography (PPG). PPG signal is used to detect physiological changes and its quality can have a significant impact on the subsequent tasks, which makes PPG quality assessment a particularly good target for examining the impact of label noise in the field of biomedicine. Random and class-dependent label noise was introduced separately into the training set to emulate the errors associated with fatigue and bias in labeling data samples. We also tested different representations of the PPG, including features defined by domain experts, 1D raw signal and 2D image. Three different classifiers are tested on the noisy training data, including support vector machine (SVM), XGBoost, 1D Resnet and 2D Resnet, which handle three representations, respectively. The results showed that the two deep learning models were more robust than the two traditional machine learning models for both the random and class-dependent label noise. From the representation perspective, the 2D image shows better robustness compared to the 1D raw signal. The logits from three classifiers are also analyzed, the predicted probabilities intend to be more dispersed when more label noise is introduced. From this work, we investigated various factors related to label noise, including representations, label noise type, and data imbalance, which can be a good guidebook for designing more robust methods for label noise in future work.

Label noise and self-learning label correction in cardiac abnormalities classification

OBJECTIVE

Learning to classify cardiac abnormalities requires large and high-quality labeled datasets, which is a challenge in medical applications. Small datasets from various sources are often aggregated to meet this requirement, resulting in a final dataset prone to label noise owing to inter- and intra-observer variability, and different expertise. It is well known that label noise can affect the performance and generalizability of the trained models. In this work, we explore the impact of label noise and self-learning label correction on the classification of cardiac abnormalities on large heterogeneous datasets of electrocardiogram (ECG) signals.

APPROACH

A state-of-the-art self-learning multi-class label correction method for image classification is adapted to learn a multi-label classifier for electrocardiogram signals. We evaluated our performance using 5-fold cross-validation on the publicly available PhysioNet/Computing in Cardiology (CinC) 2021 Challenge data, with full and reduced sets of leads. Due to the unknown label noise in the testing set, we tested our approach on the MNIST dataset. We investigated the performance under different levels of structured label noise for both datasets.

MAIN RESULTS

Under high levels of noise, the cross-validation results of self-learning label correction showed an improvement of approximately 3% in the Challenge score for the PhysioNet/CinC 2021 Challenge dataset and, an improvement in accuracy of 5$\%$ and reduction of the expected calibration error of 0.03 for the MNIST dataset. We demonstrate that self-learning label correction can be used to effectively deal with the presence of unknown label noise, also when using a reduced number of ECG leads.

Deep Learning Using CT Images to Grade Clear Cell Renal Cell Carcinoma: Development and Validation of a Prediction Model

Simple Summary

Clear cell renal cell carcinoma (ccRCC) pathologic grade identification is essential to both monitoring patients’ conditions and constructing individualized subsequent treatment strategies. However, biopsies are typically used to obtain the pathological grade, entailing tremendous physical and mental suffering as well as heavy economic burden, not to mention the increased risk of complications. Our study explores a new way to provide grade assessment of ccRCC on the basis of the individual’s appearance on CT images. A deep learning (DL) method that includes self-supervised learning is constructed to identify patients with high grade for ccRCC. We confirmed that our grading network can accurately differentiate between different grades of CT scans of ccRCC patients using a cohort of 706 patients from West China Hospital. The promising diagnostic performance indicates that our DL framework is an effective, non-invasive and labor-saving method for decoding CT images, offering a valuable means for ccRCC grade stratification and individualized patient treatment.

Abstract

This retrospective study aimed to develop and validate deep-learning-based models for grading clear cell renal cell carcinoma (ccRCC) patients. A cohort enrolling 706 patients (n = 706) with pathologically verified ccRCC was used in this study. A temporal split was applied to verify our models: the first 83.9% of the cases (years 2010–2017) for development and the last 16.1% (year 2018–2019) for validation (development cohort: n = 592; validation cohort: n = 114). Here, we demonstrated a deep learning(DL) framework initialized by a self-supervised pre-training method, developed with the addition of mixed loss strategy and sample reweighting to identify patients with high grade for ccRCC. Four types of DL networks were developed separately and further combined with different weights for better prediction. The single DL model achieved up to an area under curve (AUC) of 0.864 in the validation cohort, while the ensembled model yielded the best predictive performance with an AUC of 0.882. These findings confirms that our DL approach performs either favorably or comparably in terms of grade assessment of ccRCC with biopsies whilst enjoying the non-invasive and labor-saving property.

Performance of Regression Models as a Function of Experiment Noise

Background:

A challenge in developing machine learning regression models is that it is difficult to know whether maximal performance has been reached on the test dataset, or whether further model improvement is possible. In biology, this problem is particularly pronounced as sample labels (response variables) are typically obtained through experiments and therefore have experiment noise associated with them. Such label noise puts a fundamental limit to the metrics of performance attainable by regression models on the test dataset.

Results:

We address this challenge by deriving an expected upper bound for the coefficient of determination (R²) for regression models when tested on the holdout dataset. This upper bound depends only on the noise associated with the response variable in a dataset as well as its variance. The upper bound estimate was validated via Monte Carlo simulations and then used as a tool to bootstrap performance of regression models trained on biological datasets, including protein sequence data, transcriptomic data, and genomic data.

Conclusions:

The new method for estimating upper bounds for model performance on test data should aid researchers in developing ML regression models that reach their maximum potential. Although we study biological datasets in this work, the new upper bound estimates will hold true for regression models from any research field or application area where response variables have associated noise.

On Robustness of Neural Architecture Search Under Label Noise

Neural architecture search (NAS), which aims at automatically seeking proper neural architectures given a specific task, has attracted extensive attention recently in supervised learning applications. In most real-world situations, the class labels provided in the training data would be noisy due to many reasons, such as subjective judgments, inadequate information, and random human errors. Existing work has demonstrated the adverse effects of label noise on the learning of weights of neural networks. These effects could become more critical in NAS since the architectures are not only trained with noisy labels but are also compared based on their performances on noisy validation sets. In this paper, we systematically explore the robustness of NAS under label noise. We show that label noise in the training and/or validation data can lead to various degrees of performance variations. Through empirical experiments, using robust loss functions can mitigate the performance degradation under symmetric label noise as well as under a simple model of class conditional label noise. We also provide a theoretical justification for this. Both empirical and theoretical results provide a strong argument in favor of employing the robust loss function in NAS under high-level noise.

Label Noise Cleaning with an Adaptive Ensemble Method Based on Noise Detection Metric

Real-world datasets are often contaminated with label noise; labeling is not a clear-cut process and reliable methods tend to be expensive or time-consuming. Depending on the learning technique used, such label noise is potentially harmful, requiring an increased size of the training set, making the trained model more complex and more prone to overfitting and yielding less accurate prediction. This work proposes a cleaning technique called the ensemble method based on the noise detection metric (ENDM). From the corrupted training set, an ensemble classifier is first learned and used to derive four metrics assessing the likelihood for a sample to be mislabeled. For each metric, three thresholds are set to maximize the classifying performance on a corrupted validation dataset when using three different ensemble classifiers, namely Bagging, AdaBoost and k-nearest neighbor (k-NN). These thresholds are used to identify and then either remove or correct the corrupted samples. The effectiveness of the ENDM is demonstrated in performing the classification of 15 public datasets. A comparative analysis is conducted concerning the homogeneous-ensembles-based majority vote method and consensus vote method, two popular ensemble-based label noise filters.

Effects of Label Noise on Deep Learning-Based Skin Cancer Classification

Recent studies have shown that deep learning is capable of classifying dermatoscopic images at least as well as dermatologists. However, many studies in skin cancer classification utilize non-biopsy-verified training images. This imperfect ground truth introduces a systematic error, but the effects on classifier performance are currently unknown. Here, we systematically examine the effects of label noise by training and evaluating convolutional neural networks (CNN) with 804 images of melanoma and nevi labeled either by dermatologists or by biopsy. The CNNs are evaluated on a test set of 384 images by means of 4-fold cross validation comparing the outputs with either the corresponding dermatological or the biopsy-verified diagnosis. With identical ground truths of training and test labels, high accuracies with 75.03% (95% CI: 74.39–75.66%) for dermatological and 73.80% (95% CI: 73.10–74.51%) for biopsy-verified labels can be achieved. However, if the CNN is trained and tested with different ground truths, accuracy drops significantly to 64.53% (95% CI: 63.12–65.94%, p < 0.01) on a non-biopsy-verified and to 64.24% (95% CI: 62.66–65.83%, p < 0.01) on a biopsy-verified test set. In conclusion, deep learning methods for skin cancer classification are highly sensitive to label noise and future work should use biopsy-verified training images to mitigate this problem.

Automatic Annotation for Human Activity Recognition in Free Living Using a Smartphone

Data annotation is a time-consuming process posing major limitations to the development of Human Activity Recognition (HAR) systems. The availability of a large amount of labeled data is required for supervised Machine Learning (ML) approaches, especially in the case of online and personalized approaches requiring user specific datasets to be labeled. The availability of such datasets has the potential to help address common problems of smartphone-based HAR, such as inter-person variability. In this work, we present (i) an automatic labeling method facilitating the collection of labeled datasets in free-living conditions using the smartphone, and (ii) we investigate the robustness of common supervised classification approaches under instances of noisy data. We evaluated the results with a dataset consisting of 38 days of manually labeled data collected in free living. The comparison between the manually and the automatically labeled ground truth demonstrated that it was possible to obtain labels automatically with an 80–85% average precision rate. Results obtained also show how a supervised approach trained using automatically generated labels achieved an 84% f-score (using Neural Networks and Random Forests); however, results also demonstrated how the presence of label noise could lower the f-score up to 64–74% depending on the classification approach (Nearest Centroid and Multi-Class Support Vector Machine).