Abstract:Noise transition matrix estimation is a promising approach for learning with label noise. It can infer clean posterior probabilities, known as Label Distribution (LD), based on noisy ones and reduce the impact of noisy labels. However, this estimation is challenging, since the ground truth labels are not always available. Most existing methods estimate a global noise transition matrix using either correctly labeled samples (anchor points) or detected reliable samples (pseudo anchor points). These methods heavily rely on the existence of anchor points or the quality of pseudo ones, and the global noise transition matrix can hardly provide accurate label transition information for each sample, since the label noise in real applications is mostly instance-dependent. To address these challenges, we propose an Instance-dependent Label Distribution Estimation (ILDE) method to learn from noisy labels for image classification. The method's workflow has three major steps. First, we estimate each sample's noisy posterior probability, supervised by noisy labels. Second, since mislabeling probability closely correlates with inter-class correlation, we compute the inter-class correlation matrix to estimate the noise transition matrix, bypassing the need for (pseudo) anchor points. Moreover, for a precise approximation of the instance-dependent noise transition matrix, we calculate the inter-class correlation matrix using only mini-batch samples rather than the entire training dataset. Third, we transform the noisy posterior probability into instance-dependent LD by multiplying it with the estimated noise transition matrix, using the resulting LD for enhanced supervision to prevent DCNNs from memorizing noisy labels. The proposed ILDE method has been evaluated against several state-of-the-art methods on two synthetic and three real-world noisy datasets. Our results indicate that the proposed ILDE method outperforms all competing methods, no matter whether the noise is synthetic or real noise.

Alleviating Noisy-label Effects in Image Classification Via Probability Transition Matrix.

Noise is the Fatal Poison: A Noise-aware Network for Noisy Dataset Classification

Instance-dependent Label Distribution Estimation for Learning with Label Noise

Multi-Label Noise Transition Matrix Estimation with Label Correlations: Theory and Algorithm

Co-Correcting: Noise-Tolerant Medical Image Classification via Mutual Label Correction

DECODE: Deep Confidence Network for Robust Image Classification

TBC-MI : Suppressing Noise Labels by Maximizing Cleaning Samples for Robust Image Classification

Learning Confident Classifiers in the Presence of Label Noise

Typicality- and Instance-Dependent Label Noise-Combating: a Novel Framework for Simulating and Combating Real-World Noisy Labels for Endoscopic Polyp Classification

A Holistic View of Label Noise Transition Matrix in Deep Learning and Beyond

Class-Independent Regularization for Learning with Noisy Labels.

Multi-Objective Interpolation Training for Robustness to Label Noise

Improving Medical Images Classification with Label Noise Using Dual-Uncertainty Estimation

A Simple Probabilistic Method for Deep Classification under Input-Dependent Label Noise

Estimating Instance-dependent Label-noise Transition Matrix using DNNs

Synergistic Network Learning and Label Correction for Noise-robust Image Classification

Robustness of Classification Algorithm in the Face of Label Noise

Multi-category classification with label noise by robust binary loss

SigCo: Eliminate the inter-class competition via sigmoid for learning with noisy labels

Meta Transition Adaptation for Robust Deep Learning with Noisy Labels

Boosting Co-teaching with Compression Regularization for Label Noise