Abstract:Accurate predictions from deep neural networks are crucial for distinguishing clean data and correcting noisy labels in current label noise learning methods. However, the conventional Softmax classifier used in most relevant works is highly sensitive to label noise due to its inherent competition-prompting mechanism, i.e., similar categories are encouraged to compete for limited confidence scores during class activation, especially between the noisy classes and the ground-truth, which can inevitably lead to suboptimal predictions and eventually hamper model performance. To address this inter-class competition problem, we propose a novel Sigmoid-based Sample Selection and Correction method named SigCo for learning with noisy labels. Different from previous works, we develop a Sigmoid-based network in which each Sigmoid classifier independently predicts its respective class, improving the reliability of the selection and correction process through more accurate predictions. Besides, in order to mitigate the negative impact of noisy labels, we design a noise-adaptive learning strategy by imposing stringent class masking constraints on clean samples to enhance the learning of discriminative features, while adopting a loose masking strategy for noisy data to improve the robustness to label noise. Additionally, we introduce a co-training strategy between our Sigmoid-based network and the conventional Softmax-based network to implicitly boost the generalization capability of the model. Extensive experiments on synthetic and real-world benchmarks show that SigCo consistently outperforms state-of-the-art methods. Especially on CIFAR-100N with 80% and 90% symmetric noise ratios, it improves test accuracy by 5.10% and 18.44%, respectively.

Co-sampling: Training Robust Networks for Extremely Noisy Supervision.

Cooperative Learning for Noisy Supervision.

Co-teaching: Robust Training of Deep Neural Networks with Extremely Noisy Labels

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

Co-learning: Learning from Noisy Labels with Self-supervision

Learning with Noisy Labels Via Self-supervised Adversarial Noisy Masking

Mitigating Noisy Supervision Using Synthetic Samples with Soft Labels

Robust Noisy Label Learning via Two-Stream Sample Distillation

Mutual calibration training: Training deep neural networks with noisy labels using dual-models

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

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

Enhancing Robustness in Learning with Noisy Labels: an Asymmetric Co-Training Approach

Coresets for Robust Training of Neural Networks against Noisy Labels

Robust Training for Multi-View Stereo Networks with Noisy Labels

An Ensemble Noise-Robust K-fold Cross-Validation Selection Method for Noisy Labels

Asymmetric Co-teaching with Multi-view Consensus for Noisy Label Learning

Cross-head Supervision for Crowd Counting with Noisy Annotations.

Coordinated Sparse Recovery of Label Noise

Combating Label Noise With A General Surrogate Model For Sample Selection

CJC-net: A cyclical training method with joint loss and co-teaching strategy net for deep learning under noisy labels

Learning from Noisy Labels with Coarse-to-Fine Sample Credibility Modeling