Mitigating Instance Entanglement in Instance-Dependent Partial Label Learning

2026-03-05

AI Summary

This work addresses the class confusion problem in instance-dependent partial label learning (ID-PLL) caused by instance entanglement. To this end, the authors propose a Class-specific Augmentation and Decoupling (CAD) framework that enhances intra-class consistency through within-class feature augmentation and cross-instance alignment of the same class, while simultaneously improving inter-class separability via a label-ambiguity-aware weighted penalty loss. By jointly optimizing these components, CAD learns decoupled representations that mitigate ambiguity in label assignments. As the first mechanism specifically designed to tackle instance entanglement in ID-PLL, CAD demonstrates significant performance gains across multiple benchmark datasets, effectively alleviating class confusion and validating its efficacy in sharpening classification boundaries.

Technology Category

• Machine Learning: Multi-class/Multi-label Learning & Extreme Classification
• Constraint Satisfaction and Optimization: Constraint Learning and Acquisition
• Computer Vision: Representation Learning for Vision

Application Category

• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking

Abstract

Partial label learning is a prominent weakly supervised classification task, where each training instance is ambiguously labeled with a set of candidate labels. In real-world scenarios, candidate labels are often influenced by instance features, leading to the emergence of instance-dependent PLL (ID-PLL), a setting that more accurately reflects this relationship. A significant challenge in ID-PLL is instance entanglement, where instances from similar classes share overlapping features and candidate labels, resulting in increased class confusion. To address this issue, we propose a novel Class-specific Augmentation based Disentanglement (CAD) framework, which tackles instance entanglement by both intra- and inter-class regulations. For intra-class regulation, CAD amplifies class-specific features to generate class-wise augmentations and aligns same-class augmentations across instances. For inter-class regulation, CAD introduces a weighted penalty loss function that applies stronger penalties to more ambiguous labels, encouraging larger inter-class distances. By jointly applying intra- and inter-class regulations, CAD improves the clarity of class boundaries and reduces class confusion caused by entanglement. Extensive experimental results demonstrate the effectiveness of CAD in mitigating the entanglement problem and enhancing ID-PLL performance. The code is available at https://github.com/RyanZhaoIc/CAD.git.

Problem

• instance entanglement
• partial label learning
• instance-dependent
• class confusion
• weakly supervised classification

Innovation

• instance-dependent partial label learning
• instance entanglement
• class-specific augmentation
• disentanglement
• weighted penalty loss