Appeal: Allow Mislabeled Samples the Chance to be Rectified in Partial Label Learning
This work addresses a specific bottleneck in partial label learning by enhancing existing methods to handle mislabeled samples, which is an incremental advancement for the field.
The paper tackles the problem of mislabeled samples in partial label learning by proposing an appeal-based framework that introduces a partner classifier and a collaborative term to improve identification and rectification of mislabeled samples, resulting in significant improvements in appeal and disambiguation abilities for several established PLL methods.
In partial label learning (PLL), each instance is associated with a set of candidate labels among which only one is ground-truth. The majority of the existing works focuses on constructing robust classifiers to estimate the labeling confidence of candidate labels in order to identify the correct one. However, these methods usually struggle to identify and rectify mislabeled samples. To help these mislabeled samples "appeal" for themselves and help existing PLL methods identify and rectify mislabeled samples, in this paper, we propose the first appeal-based PLL framework. Specifically, we introduce a novel partner classifier and instantiate it predicated on the implicit fact that non-candidate labels of a sample should not be assigned to it, which is inherently accurate and has not been fully investigated in PLL. Furthermore, a novel collaborative term is formulated to link the base classifier and the partner one. During each stage of mutual supervision, both classifiers will blur each other's predictions through a blurring mechanism to prevent overconfidence in a specific label. Extensive experiments demonstrate that the appeal and disambiguation ability of several well-established stand-alone and deep-learning based PLL approaches can be significantly improved by coupling with this learning paradigm.