Partial Label Learning with Self-Guided Retraining
This work addresses a key challenge in weakly supervised learning for scenarios with ambiguous labels, offering a more efficient and effective solution for tasks like image classification or text tagging.
The paper tackles the problem of partial label learning, where each training instance has multiple candidate labels but only one is correct, by introducing a self-training approach with a novel regularization method that automatically identifies the true label, resulting in significant performance improvements over state-of-the-art methods on various datasets.
Partial label learning deals with the problem where each training instance is assigned a set of candidate labels, only one of which is correct. This paper provides the first attempt to leverage the idea of self-training for dealing with partially labeled examples. Specifically, we propose a unified formulation with proper constraints to train the desired model and perform pseudo-labeling jointly. For pseudo-labeling, unlike traditional self-training that manually differentiates the ground-truth label with enough high confidence, we introduce the maximum infinity norm regularization on the modeling outputs to automatically achieve this consideratum, which results in a convex-concave optimization problem. We show that optimizing this convex-concave problem is equivalent to solving a set of quadratic programming (QP) problems. By proposing an upper-bound surrogate objective function, we turn to solving only one QP problem for improving the optimization efficiency. Extensive experiments on synthesized and real-world datasets demonstrate that the proposed approach significantly outperforms the state-of-the-art partial label learning approaches.