Learning from Similarity-Confidence and Confidence-Difference

2025-08-07

AI Summary

Weakly supervised learning suffers from severe label scarcity and high noise levels in training labels. Method: This paper proposes a novel classification framework that jointly leverages two complementary weak supervision signals—similarity confidence and confidence discrepancy—under multiple relational perspectives. Its core innovation is the first interactive unbiased risk estimator, which models the coupling among weak labels via convex combination and incorporates a risk correction mechanism to mitigate overfitting induced by negative empirical risk. The method requires no additional annotations and significantly enhances generalization under few-shot settings. Contribution/Results: Extensive experiments demonstrate consistent superiority over state-of-the-art baselines across diverse weak supervision paradigms, including positive-unlabeled learning and complementary-label learning. Moreover, the approach exhibits strong robustness against both class-prior estimation errors and label noise.

Technology Category

• Machine Learning: Semi-Supervised Learning
• Reasoning under Uncertainty: Relational Probabilistic Models
• Computer Vision: Learning & Optimization for CV

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Web Mining and Content Analysis: Robustness and generalizability of Web mining methods
• Graph Algorithms and Modeling for the Web: Algorithms and analysis for incomplete, noisy, or partially observed Web-related graphs

Abstract

In practical machine learning applications, it is often challenging to assign accurate labels to data, and increasing the number of labeled instances is often limited. In such cases, Weakly Supervised Learning (WSL), which enables training with incomplete or imprecise supervision, provides a practical and effective solution. However, most existing WSL methods focus on leveraging a single type of weak supervision. In this paper, we propose a novel WSL framework that leverages complementary weak supervision signals from multiple relational perspectives, which can be especially valuable when labeled data is limited. Specifically, we introduce SconfConfDiff Classification, a method that integrates two distinct forms of weaklabels: similarity-confidence and confidence-difference, which are assigned to unlabeled data pairs. To implement this method, we derive two types of unbiased risk estimators for classification: one based on a convex combination of existing estimators, and another newly designed by modeling the interaction between two weak labels. We prove that both estimators achieve optimal convergence rates with respect to estimation error bounds. Furthermore, we introduce a risk correction approach to mitigate overfitting caused by negative empirical risk, and provide theoretical analysis on the robustness of the proposed method against inaccurate class prior probability and label noise. Experimental results demonstrate that the proposed method consistently outperforms existing baselines across a variety of settings.

Problem

• Leveraging multiple weak supervision signals for learning
• Integrating similarity-confidence and confidence-difference weak labels
• Mitigating overfitting and improving robustness in weak supervision

Innovation

• Leverages similarity-confidence and confidence-difference weaklabels
• Uses convex combination and interaction-based unbiased risk estimators
• Introduces risk correction to mitigate overfitting