SplitNet: Learnable Clean-Noisy Label Splitting for Learning with Noisy Labels
This work improves deep learning robustness by enabling more effective handling of label noise, which is crucial for real-world applications where data annotation is often imperfect.
The paper tackles the problem of learning with noisy labels by introducing SplitNet, a learnable module for clean-noisy label splitting, which addresses confirmation bias from handcrafted methods and achieves state-of-the-art performance, particularly in high noise ratio settings on various benchmarks.
Annotating the dataset with high-quality labels is crucial for performance of deep network, but in real world scenarios, the labels are often contaminated by noise. To address this, some methods were proposed to automatically split clean and noisy labels, and learn a semi-supervised learner in a Learning with Noisy Labels (LNL) framework. However, they leverage a handcrafted module for clean-noisy label splitting, which induces a confirmation bias in the semi-supervised learning phase and limits the performance. In this paper, we for the first time present a learnable module for clean-noisy label splitting, dubbed SplitNet, and a novel LNL framework which complementarily trains the SplitNet and main network for the LNL task. We propose to use a dynamic threshold based on a split confidence by SplitNet to better optimize semi-supervised learner. To enhance SplitNet training, we also present a risk hedging method. Our proposed method performs at a state-of-the-art level especially in high noise ratio settings on various LNL benchmarks.