​Integrating Graph-Structured Knowledge into Large Language Models
"How GNNs with Ontology Enhance LLMs' Attention"

• ​Bio

  • ​Ph.D. in Industrial Engineering, Yonsei Univ.  

  • ​B.T. in Theology & CS, Yonsei Univ.

• ​Presenter recent research
  1) Knowledge Graph as Pre-Training Corpus for Structural Reasoning via Multi-Hop Linearization, IEEE Access

  • ​Treats knowledge graphs as a pre-training corpus, enabling LLMs to acquire structural and multi-hop reasoning ability during pretraining.

• ​2) Graph Discrete Prompt Optimization for Knowledge Graph Question Answering, WWW 2026 Accept

  • ​Optimizes KG-to-text prompting as a discrete optimization problem, enabling effective KG utilization in closed LLMs.

• ​3) Addressing Information Bottlenecks in Graph Augmented Large Language Models via Graph Neural Summarization, Information Fusion

  • ​Identifies information bottlenecks as a core limitation in graph-augmented LLMs, showing that dense graph information degrades reasoning.