Diffusion Disambiguation Models for Partial Label Learning

TL;DR

This paper introduces a diffusion-based approach for partial label learning, using generative models to disambiguate labels by iteratively refining initial guesses and leveraging instance-label relationships.

Contribution

It proposes the diffusion disambiguation model (DDMP) that constructs pseudo-clean labels and dynamically updates label estimates during training for improved partial label learning.

Findings

DDMP outperforms existing methods in partial label learning tasks.

The model effectively refines ambiguous labels through diffusion processes.

Experimental results demonstrate the approach's robustness and accuracy.

Abstract

Learning from ambiguous labels is a long-standing problem in practical machine learning applications. The purpose of \emph{partial label learning} (PLL) is to identify the ground-truth label from a set of candidate labels associated with a given instance. Inspired by the remarkable performance of diffusion models in various generation tasks, this paper explores their potential to denoise ambiguous labels through the reverse denoising process. Therefore, this paper reformulates the label disambiguation problem from the perspective of generative models, where labels are generated by iteratively refining initial random guesses. This perspective enables the diffusion model to learn how label information is generated stochastically. By modeling the generation uncertainty, we can use the maximum likelihood estimate of the label for classification inference. However, such ambiguous labels lead to a mismatch between instance and label, which reduces the quality of generated data. To address this issue, this paper proposes a \emph{diffusion disambiguation model for PLL} (DDMP), which first uses the potential complementary information between instances and labels to construct pseudo-clean labels for initial diffusion training. Furthermore, a transition-aware matrix is introduced to estimate the potential ground-truth labels, which are dynamically updated during the diffusion generation. During training, the ground-truth label is progressively refined, improving the classifier. Experiments show the advantage of the DDMP and its suitability for PLL.