EviNet: Evidential Reasoning Network for Resilient Graph Learning in the Open and Noisy Environments

2025-06-08

AI Summary

Real-world graph learning is often constrained by the closed-world assumption, limiting its ability to detect misclassifications (i.e., erroneous predictions on known classes) and identify out-of-distribution (OOD) samples (i.e., instances from unknown classes) in open, noisy environments. To address this, we propose the first evidence-reasoning framework integrating Beta embeddings with subjective logic, featuring two novel modules: *Dissonance*, quantifying logical inconsistency among class-wise evidences, and *Vacuity*, measuring epistemic uncertainty. Our method employs differentiable Beta-distributed graph embeddings, uncertainty-aware graph neural networks, and an evidence-aggregation mechanism. Evaluated across multiple benchmarks, it consistently surpasses state-of-the-art methods: it achieves significant improvements in in-distribution classification accuracy, misclassification detection, and OOD identification—demonstrating superior predictive accuracy and robustness under distributional shift and label noise.

Technology Category

• Reasoning under Uncertainty: Graphical Models
• Machine Learning: Calibration & Uncertainty Quantification
• Knowledge Representation and Reasoning: Reasoning with Beliefs

Application Category

• Graph Algorithms and Modeling for the Web: Graph embeddings and representation learning for Web-related graphs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Graph learning has been crucial to many real-world tasks, but they are often studied with a closed-world assumption, with all possible labels of data known a priori. To enable effective graph learning in an open and noisy environment, it is critical to inform the model users when the model makes a wrong prediction to in-distribution data of a known class, i.e., misclassification detection or when the model encounters out-of-distribution from novel classes, i.e., out-of-distribution detection. This paper introduces Evidential Reasoning Network (EVINET), a framework that addresses these two challenges by integrating Beta embedding within a subjective logic framework. EVINET includes two key modules: Dissonance Reasoning for misclassification detection and Vacuity Reasoning for out-of-distribution detection. Extensive experiments demonstrate that EVINET outperforms state-of-the-art methods across multiple metrics in the tasks of in-distribution classification, misclassification detection, and out-of-distribution detection. EVINET demonstrates the necessity of uncertainty estimation and logical reasoning for misclassification detection and out-of-distribution detection and paves the way for open-world graph learning. Our code and data are available at https://github.com/SSSKJ/EviNET.

Problem

• Detects misclassification in known data classes
• Identifies out-of-distribution novel class data
• Enhances graph learning in noisy open environments

Innovation

• Integrates Beta embedding in subjective logic
• Uses Dissonance Reasoning for misclassification detection
• Employs Vacuity Reasoning for out-of-distribution detection