Which is Better for Learning with Noisy Labels: The Semi-supervised Method or Modeling Label Noise?

In real life, accurately annotating large-scale datasets is sometimes difficult. Datasets used for training deep learning models are likely to contain label noise. To make use of the dataset containing label noise, two typical methods have been proposed. One is to employ the semi-supervised method by exploiting labeled \textit{confident examples} and unlabeled \textit{non-confident examples}. The other one is to \textit{model label noise} and design \textit{statistically consistent} classifiers. A natural question remains unsolved: which one should be used for a specific real-world application? In this paper, we answer the question from the perspective of \textit{causal data generative process}. Specifically, the semi-supervised method depends heavily on the data generation process while the modeling label noise method is independent of the generation process. For example, for a given dataset, if it has a causal generative structure that the features cause the label, the semi-supervised method would not be helpful. When the causal structure is unknown, we provide an intuitive method to discover the causal structure for a given dataset containing label noise.