This paper combines graph neural network encoders with box-embedding-based semantic losses that encourage learned representations to better satisfy ontology-derived class hierarchies, and compares this approach to both standard link prediction models and models incorporating subclass relations as graph edges.
Abstract
Knowledge graphs are often accompanied by ontological class hierarchies that encode valuable semantic information, yet many link prediction methods either ignore such hierarchies or incorporate them indirectly through additional graph edges. Recent work introduced hierarchy-aware graph neural networks (GNNs), which use semantic losses derived from box embeddings to encourage satisfaction of subclass relationships during GNN-based representation learning. While this approach has shown promise for biological regression tasks, its effectiveness for knowledge graph link prediction has not been investigated. In this paper we evaluate hierarchy-aware semantic losses on link prediction across three benchmark datasets: AIFB, CoDEx, and BioKG. We combine graph neural network encoders with box-embedding-based semantic losses that encourage learned representations to better satisfy ontology-derived class hierarchies, and compare this approach to both standard link prediction models and models incorporating subclass relations as graph edges. Across all datasets, hierarchy-aware semantic losses significantly improve mean reciprocal rank (MRR) and consistently outperform models that incorporate hierarchy information through additional subclass edges. Relative to the baseline GNN models, MRR improved by 7.6%, 2.4%, and 15.5% on AIFB, CoDEx, and BioKG, respectively. Furthermore, semantic losses consistently outperform the alternative of augmenting the graph with subclass edges. These results are consistent with ontology-derived class hierarchies providing complementary information to graph structure, and suggest that encouraging hierarchical consistency through semantic losses is an effective and comparatively parameter-efficient mechanism for improving knowledge graph link prediction.
A neural regression model (LitEm) that enables transductive knowledge graph embedding models to predict numerical attributes within knowledge graphs and a co-training framework that jointly trains state-of-the-art transductive knowledge graph embedding models with LitEm, which improves link prediction performance mainly for bilinear models and simultaneously enables them to predict numerical attributes.
Rupesh Sapkota, Louis Mozart Kamdem Teyou, Moshood Yekini et al.· 0 citations
Ontology-based Knowledge Graphs (KGs) augment entity representation through additional semantic information, facilitating link prediction for unseen entities through predefined ontology libraries. While most existing knowledge graph representation learning methods predominantly focus on co-optimizing both entities and ontologies to leverage ontological contexts, the structural-semantic discrepancies in ontology-based KGs have been largely overlooked. Through graph structure analysis, we identify two fundamental limitations: (1) structural incompatibility between entity subgraph semantics and multi-ontology mappings (1-N redundancy), and (2) missing explicit ontology link in subgraph contexts (1-0 absence). To resolve these issues, we propose a structural empowered module built upon link prediction backbones. First, we develop a subgraph-aware semantic expansion module that coordinates $k$-hop neighborhood information with LLM-generated descriptions to alleviate structural sparsity. Subsequently, a contrastive ontology matching mechanism resolves structural inconsistencies by computing adaptive similarity metrics between ontology embeddings and subgraph-derived semantic prototypes. Experimental results demonstrate that our model outperforms fourteen state-of-the-art models, maintaining robust performance across varying benchmarks and subgraph density conditions.
Hao Li, K. Liang, Lingyuan Meng et al.· IEEE Transactions on Pattern...· 0 citations
Knowledge Graph Completion (KGC) aims to predict missing links in knowledge bases. However, existing methods often struggle with reasoning over complex semantic relations and generate predictions that violate common-sense logic. In this paper, a knowledge graph completion model based on semantic constrained hybrid scoring (SCHS) is proposed, which integrates three synergistic mechanisms. First, an Adaptive Feature Gating (AFG) mechanism is designed for feature recalibration. Unlike static embeddings, this module dynamically recalibrates the importance of different semantic aspects based on the relation context. Second, a Relation-Concept Matrix Projection (RCMP) mechanism is introduced to act as an explicit semantic constraint. This mechanism maps triplets to a latent concept space to capture the inherent consistency between relations and concepts, providing a penalization mechanism to filter out noise that is topologically close but semantically invalid. Third, an Adaptive Fusion Strategy (AFS) utilizing a soft-start mechanism is introduced to effectively balance structural and semantic scores without early-training collapses. This tri-fold synergy essentially resolves the rigid trade-off between geometric alignment and flexible semantic matching. Extensive experiments on the WN18RR and FB15k-237 datasets demonstrate that SCHS achieves substantial performance improvements, verifying the robustness and effectiveness of the dynamically regulated fusion of semantics and structures in the alleviation of long-tail errors.
Pengsong Duan, Jia-Ke Zhao, Bo Zhang et al.· International journal of sof...· 0 citations
Inductive link prediction in knowledge graphs refers to the task of inferring known relations between entities unseen during training. Most existing approach-es are limited to predicting only known relations and struggle to generalize to un-seen relations, which restricts their utility in dynamic settings. To address this challenge, we propose a novel inductive link prediction approach named RGIILP. Specifically, we construct a relation graph from the source knowledge graph and design a neural network model that enables interactive feature propagation be-tween entities and relations. Furthermore, we introduce the mutual information maximization mechanism between global and local representations to capture the global structural information of the graph. Experiments on several benchmark da-tasets demonstrate that RGIILP outperforms existing state-of-the-art methods for inductive link prediction task.
Hong-Bo Liu· Poster Volume 0008 The 2026...· 0 citations
This framework performs LLM knowledge elicitation to extract factual knowledge from the model’s internal representations and transforms sentence-level representations into entity-level representations and aligns them within a unified space.
Deyu Chen, Qiyuan Li, Jinguang Gu et al.· 0 citations
Related blog posts
MIT News · Artificial Intelligence· news.mit.eduJul 15, 2026
Assistant Professor Pat Pataranutaporn describes a new interface that lets everyday users glimpse inside an AI's neural network before their chatbot ever says a word.
Microsoft Research Blog· microsoft.comJul 13, 2026
Cryptographic code supports vital protections in modern computing systems. Learn how a new method helps verify code as developers write it while preserving speed and adaptability as it gets implemented and evolves. The post Verifying Rust cryptography in SymCrypt, from standards to code appeared first on Microsoft Research.
MIT News · Artificial Intelligence· news.mit.eduJul 6, 2026
PhD student Rachel Sava, winner of the Envisioning the Future of Computing Prize, explores transformative improvements and dystopian risks of neural technology.
