We do not consider the uncertainty-based models (e.g., Monte Carlo (MC) dropout [ Figure 5: Blurred and restored images using different image denoising methods. All images are normalized and augmented by random horizontal flipping. Five networks with ResNet-18 are trained from scratch using PyTorch 1.9.0. Default PyTorch initialization is used on all layers. The model warm-up can help better separate noisy data and clean data.

We do not consider the uncertainty-based models (e.g., Monte Carlo (MC) dropout [ Figure 5: Blurred and restored images using different image denoising methods. All images are normalized and augmented by random horizontal flipping. Five networks with ResNet-18 are trained from scratch using PyTorch 1.9.0. Default PyTorch initialization is used on all layers. The model warm-up can help better separate noisy data and clean data.

Deep neural networks (DNNs) have achieved remarkable success in a variety of computer vision tasks, where massive labeled images are routinely required for model optimization. Yet, the data collected from the open world are unavoidably polluted by noise, which may significantly undermine the efficacy of the learned models. Various attempts have been made to reliably train DNNs under data noise, but they separately account for either the noise existing in the labels or that existing in the images. A naive combination of the two lines of works would suffer from the limitations in both sides, and miss the opportunities to handle the two kinds of noise in parallel. This work provides a first, unified framework for reliable learning under the joint (image, label)-noise. Technically, we develop a confidence-based sample filter to progressively filter out noisy data without the need of pre-specifying noise ratio. Then, we penalize the model uncertainty of the detected noisy data instead of letting the model continue over-fitting the misleading information in them. Experimental results on various challenging synthetic and real-world noisy datasets verify that the proposed method can outperform competing baselines in the aspect of classification performance.