knowledge-graph-completion

Proposes DuPLeR, a dual-path LLM reasoning framework for multimodal few-shot knowledge graph completion, combining LLM-derived type priors with factual structures to improve inductive KGC under data scarcity.

Proposes M2GDT, a novel MKGC framework that uses an MLLM-guided diffusion transformer with relation-adaptive mixture-of-experts to align and denoise multimodal features, outperforming baselines on three benchmark datasets.

Proposes a conditional diffusion-guided knowledge transfer framework for multi-domain knowledge graph completion, generating domain-general entity embeddings without suppressing domain-specific information, achieving 4.3% average MRR improvement over state-of-the-art methods.

Proposes RelBall, a KGC model that extends Rotate3D with modulus transformation for modeling hierarchies and a tail-centric relation ball to handle one-to-many relations, achieving competitive link prediction performance.

This paper introduces CORE, a new knowledge graph completion model that uses cyclic orthotope relation embeddings on a torus manifold to address boundary constraints in region-based models. Experiments show competitive performance in link prediction tasks.

This paper proposes M-Hyper, a novel multi-modal knowledge graph completion method that balances fusion and independence of modality representations using hypercomplex (biquaternion) algebra. The approach introduces Fine-grained Entity Representation Factorization and Robust Relation-aware Modality Fusion modules to achieve state-of-the-art performance with improved robustness.