Rethinking Noisy Label Learning in Real-world Annotation Scenarios from the Noise-type Perspective
This work addresses the problem of noisy label learning for machine learning practitioners, offering an incremental improvement by refining existing methods with a novel noise-type perspective.
The paper tackles learning with noisy labels in real-world scenarios by categorizing noise into factual and ambiguity types, and proposes Proto-semi, a sample selection approach that uses prototypes and semi-supervised learning to correct labels, achieving robustness and effectiveness in empirical evaluations on a real-world dataset.
In this paper, we investigate the problem of learning with noisy labels in real-world annotation scenarios, where noise can be categorized into two types: factual noise and ambiguity noise. To better distinguish these noise types and utilize their semantics, we propose a novel sample selection-based approach for noisy label learning, called Proto-semi. Proto-semi initially divides all samples into the confident and unconfident datasets via warm-up. By leveraging the confident dataset, prototype vectors are constructed to capture class characteristics. Subsequently, the distances between the unconfident samples and the prototype vectors are calculated to facilitate noise classification. Based on these distances, the labels are either corrected or retained, resulting in the refinement of the confident and unconfident datasets. Finally, we introduce a semi-supervised learning method to enhance training. Empirical evaluations on a real-world annotated dataset substantiate the robustness of Proto-semi in handling the problem of learning from noisy labels. Meanwhile, the prototype-based repartitioning strategy is shown to be effective in mitigating the adverse impact of label noise. Our code and data are available at https://github.com/fuxiAIlab/ProtoSemi.