![[2506.12362] HYPER: A Foundation Model for Inductive Link Prediction with Knowledge Hypergraphs](https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png){kind=logo}
[2506.12362] HYPER: A Foundation Model for Inductive Link Prediction with Knowledge Hypergraphs
Summary
The paper presents HYPER, a foundation model designed for inductive link prediction using knowledge hypergraphs, capable of generalizing to novel entities and relations.
Why It Matters
HYPER addresses limitations in existing link prediction methods that cannot adapt to new relational types, enhancing the capability of machine learning models in dynamic environments. This advancement is crucial for applications in AI that require robust generalization across diverse datasets.
Key Takeaways
- HYPER can generalize to novel entities and relations in knowledge hypergraphs.
- The model outperforms existing methods in both node-only and node-and-relation settings.
- 16 new inductive datasets were constructed to evaluate HYPER's performance.
- HYPER encodes entities and their positions in hyperedges for better learning.
- The approach enhances the adaptability of models to unseen relational structures.
Computer Science > Machine Learning arXiv:2506.12362 (cs) [Submitted on 14 Jun 2025 (v1), last revised 13 Feb 2026 (this version, v2)] Title:HYPER: A Foundation Model for Inductive Link Prediction with Knowledge Hypergraphs Authors:Xingyue Huang, Mikhail Galkin, Michael M. Bronstein, İsmail İlkan Ceylan View a PDF of the paper titled HYPER: A Foundation Model for Inductive Link Prediction with Knowledge Hypergraphs, by Xingyue Huang and 3 other authors View PDF Abstract:Inductive link prediction with knowledge hypergraphs is the task of predicting missing hyperedges involving completely novel entities (i.e., nodes unseen during training). Existing methods for inductive link prediction with knowledge hypergraphs assume a fixed relational vocabulary and, as a result, cannot generalize to knowledge hypergraphs with novel relation types (i.e., relations unseen during training). Inspired by knowledge graph foundation models, we propose HYPER as a foundation model for link prediction, which can generalize to any knowledge hypergraph, including novel entities and novel relations. Importantly, HYPER can learn and transfer across different relation types of varying arities, by encoding the entities of each hyperedge along with their respective positions in the hyperedge. To evaluate HYPER, we construct 16 new inductive datasets from existing knowledge hypergraphs, covering a diverse range of relation types of varying arities. Empirically, HYPER consistently outperforms all existing...
