Breaking the Reasoning Horizon in Entity Alignment Foundation Models
This work addresses the challenge of aligning unseen knowledge graphs without retraining, which is crucial for knowledge graph fusion, though it appears incremental as it builds on graph foundation models.
The paper tackles the problem of entity alignment in knowledge graphs by addressing the reasoning horizon gap that limits existing models, proposing a parallel encoding strategy that uses seed pairs as anchors to improve transferability and achieve strong generalizability to unseen KGs.
Entity alignment (EA) is critical for knowledge graph (KG) fusion. Existing EA models lack transferability and are incapable of aligning unseen KGs without retraining. While using graph foundation models (GFMs) offer a solution, we find that directly adapting GFMs to EA remains largely ineffective. This stems from a critical "reasoning horizon gap": unlike link prediction in GFMs, EA necessitates capturing long-range dependencies across sparse and heterogeneous KG structuresTo address this challenge, we propose a EA foundation model driven by a parallel encoding strategy. We utilize seed EA pairs as local anchors to guide the information flow, initializing and encoding two parallel streams simultaneously. This facilitates anchor-conditioned message passing and significantly shortens the inference trajectory by leveraging local structural proximity instead of global search. Additionally, we incorporate a merged relation graph to model global dependencies and a learnable interaction module for precise matching. Extensive experiments verify the effectiveness of our framework, highlighting its strong generalizability to unseen KGs.