Rule Learning for Knowledge Graph Reasoning under Agnostic Distribution Shift

2025-07-07

AI Summary

This study addresses the violation of the independent and identically distributed (i.i.d.) assumption in knowledge graph (KG) reasoning—caused by unknown selection bias during training and distributional shifts during testing—and formally defines the “out-of-distribution (OOD) KG reasoning” task for the first time. To enhance model generalization and robustness under non-i.i.d. conditions, we propose StableRule, a stable rule learning framework that decouples spurious correlations via feature decorrelation, thereby separating logical rule learning from distribution shifts; it further introduces an end-to-end differentiable architecture that jointly optimizes rule discovery and invariant representation learning. Extensive experiments across seven heterogeneous KG benchmarks demonstrate that StableRule significantly improves OOD generalization performance and inference stability. Our work establishes a novel paradigm for KG reasoning in open-world, distributionally dynamic environments.

Technology Category

• Knowledge Representation and Reasoning: Other Foundations of Knowledge Representation & Reasoning
• Data Mining & Knowledge Management: Rule Mining & Pattern Mining
• Machine Learning: Statistical Relational/Logic Learning

Application Category

• Semantics and Knowledge: Scalable techniques for the creation, curation, publication, maintenance, and consumption of large, Web-based, structured, reusable, knowledge graphs and ontologies
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Graph Algorithms and Modeling for the Web: Graph embeddings and representation learning for Web-related graphs

Abstract

Knowledge graph (KG) reasoning remains a critical research area focused on inferring missing knowledge by analyzing relationships among observed facts. Despite its success, a key limitation of existing KG reasoning methods is their dependence on the I.I.D assumption. This assumption can easily be violated due to unknown sample selection bias during training or agnostic distribution shifts during testing, significantly compromising model performance and reliability. To facilitate the deployment of KG reasoning in wild environments, this study investigates learning logical rules from KGs affected by unknown selection bias. Additionally, we address test sets with agnostic distribution shifts, formally defining this challenge as out-of-distribution (OOD) KG reasoning-a previously underexplored problem. To solve the issue, we propose the Stable Rule Learning (StableRule) framework, an end-to-end methodology that integrates feature decorrelation with rule learning network, to enhance OOD generalization performance. By leveraging feature decorrelation, the StableRule framework mitigates the adverse effects of covariate shifts arising in OOD scenarios, thereby improving the robustness of the rule learning component in effectively deriving logical rules. Extensive experiments on seven benchmark KGs demonstrate the framework's superior effectiveness and stability across diverse heterogeneous environments, underscoring its practical significance for real-world applications.

Problem

• Addressing KG reasoning under unknown selection bias
• Solving OOD KG reasoning with agnostic distribution shifts
• Enhancing robustness in rule learning for KG reasoning

Innovation

• Feature decorrelation for stable rule learning
• End-to-end OOD KG reasoning framework
• Robust logical rule derivation under shift