EReLiFM: Evidential Reliability-Aware Residual Flow Meta-Learning for Open-Set Domain Generalization under Noisy Labels

2025-10-14

AI Summary

This paper addresses open-set domain generalization (OSDG) under label noise. To mitigate source-domain knowledge corruption caused by noisy labels and bridge inter-domain distribution shifts, we propose an uncertainty-aware knowledge transfer framework. Our method introduces an evidential reliability-aware mechanism to dynamically assess sample confidence; designs a residual flow matching strategy to align feature flows of known classes with rejection flows of unknown classes; and integrates two-stage unsupervised evidential clustering, clean-set-guided meta-learning, and high-quality pseudo-label optimization. Extensive experiments on multiple benchmarks demonstrate significant improvements over state-of-the-art methods, validating the framework’s robustness under label noise, as well as its simultaneous enhancement of known-class classification accuracy and unknown-class rejection capability.

Technology Category

• Machine Learning: Calibration & Uncertainty Quantification
• Reasoning under Uncertainty: Uncertainty Representations
• Data Mining & Knowledge Management: Anomaly/Outlier Detection

Application Category

• Economics, Online Markets and Human Computation: Trust and reliance of crowd workers and data experts on GenAI
• Semantics and Knowledge: Scalable techniques for the creation, curation, publication, maintenance, and consumption of large, Web-based, structured, reusable, knowledge graphs and ontologies
• Graph Algorithms and Modeling for the Web: Algorithms and analysis for incomplete, noisy, or partially observed Web-related graphs

Abstract

Open-Set Domain Generalization (OSDG) aims to enable deep learning models to recognize unseen categories in new domains, which is crucial for real-world applications. Label noise hinders open-set domain generalization by corrupting source-domain knowledge, making it harder to recognize known classes and reject unseen ones. While existing methods address OSDG under Noisy Labels (OSDG-NL) using hyperbolic prototype-guided meta-learning, they struggle to bridge domain gaps, especially with limited clean labeled data. In this paper, we propose Evidential Reliability-Aware Residual Flow Meta-Learning (EReLiFM). We first introduce an unsupervised two-stage evidential loss clustering method to promote label reliability awareness. Then, we propose a residual flow matching mechanism that models structured domain- and category-conditioned residuals, enabling diverse and uncertainty-aware transfer paths beyond interpolation-based augmentation. During this meta-learning process, the model is optimized such that the update direction on the clean set maximizes the loss decrease on the noisy set, using pseudo labels derived from the most confident predicted class for supervision. Experimental results show that EReLiFM outperforms existing methods on OSDG-NL, achieving state-of-the-art performance. The source code is available at https://github.com/KPeng9510/ERELIFM.

Problem

• Recognizing unseen categories in new domains under noisy labels
• Bridging domain gaps with limited clean labeled data
• Handling corrupted source-domain knowledge from label noise

Innovation

• Evidential loss clustering enhances label reliability
• Residual flow matching enables uncertainty-aware domain transfer
• Meta-learning optimizes clean set updates for noisy supervision