Abstract:Although deep neural networks have been proved effective in many applications, they are data hungry, and training deep models often requires laboriously labeled data. However, when labeled data contain erroneous labels, they often lead to model performance degradation. A common solution is to assign each sample with a dynamic weight during optimization, and the weight is adjusted in accordance with the loss. However, those weights are usually unreliable since they are measured by the losses of corrupted labels. Thus, this scheme might impede the discriminative ability of neural networks trained on noisy data. To address this issue, we propose a novel reweighting method, dubbed self-reweighting from class centroids (SRCC), by assigning sample weights based on the similarities between the samples and our online learned class centroids. Since we exploit statistical class centers in the image feature space to reweight data samples in learning, our method is robust to noise caused by corrupted labels. In addition, even after reweighting the noisy data, the decision boundaries might still suffer distortions. Thus, we leverage mixed inputs that are generated by linearly interpolating two random images and their labels to further regularize the boundaries. We employ the learned class centroids to evaluate the confidence of our generated mixed data via measuring feature similarities. During the network optimization, the class centroids are updated as more discriminative feature representations of original images are learned. In doing so, SRCC will generate more robust weighting coefficients for noisy and mixed data and facilitates our feature representation learning in return. Extensive experiments on both the synthetic and real image recognition tasks demonstrate that our method SRCC outperforms the state of the art on learning with noisy data.

Joint Negative-Positive-learning Based Sample Reweighting for Hyperspectral Image Classification with Label Noise

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

Agreement and Disagreement-Based Co-Learning with Dual Network for Hyperspectral Image Classification with Noisy Labels.

Making Deep Neural Networks Robust to Label Noise: A Reweighting Loss and Data Filtration

Noise Modeling and Learning-Based Hyperspectral Image Denoising Used for Hyperspectral Unmixing

Learning with Noisy Labels Using Hyperspherical Margin Weighting

A Hybrid Neural Network for Hyperspectral Image Classification

Ensemble Network-Based Distillation for Hyperspectral Image Classification in the Presence of Label Noise

A Superpixel Guided Sample Selection Neural Network for Handling Noisy Labels in Hyperspectral Image Classification

When Broad Learning System Meets Label Noise Learning: A Reweighting Learning Framework.

Dual-Channel Residual Network for Hyperspectral Image Classification with Noisy Labels

Hyperspectral Images Weakly Supervised Classification with Noisy Labels

Negative Samples Mining Matters: Reconsidering Hyperspectral Image Classification with Contrastive Learning

Robust Image Classification with Noisy Labels by Negative Learning and Feature Space Renormalization

Learning with Noisy Labels Via Self-Reweighting from Class Centroids

Hyperspectral Image Classification in the Presence of Noisy Labels

CMW-Net: Learning a Class-Aware Sample Weighting Mapping for Robust Deep Learning

Hyperspectral Image Classification Via A Joint Weighted K-Nearest Neighbour Approach

Hybrid Noise-Oriented Multilabel Learning.

Dynamic Loss Reweighting Method Based on Cumulative Classification Scores for Long-Tailed Remote Sensing Image Classification

Noise-Robust Bidirectional Learning with Dynamic Sample Reweighting