Knowledge Graph

The Knowledge Graph Research Group focuses on chemical information extraction, knowledge graph construction, and applications, aiming to advance artificial intelligence-driven innovations in chemistry. Our team consists of experts from interdisciplinary fields such as chemistry and computer science, leveraging Large Language Models (LLMs), Knowledge Graphs (KGs), and Retrieval-Augmented Generation (RAG) to support chemical research and enhance scientists' ability to explore unknowns.

Research Directions

• Chemical Information Extraction
  Utilizing proprietary LLMs and fine-tuned open-source LLMs, we extract, clean, and store structured chemical information from unstructured sources such as literature, building a high-quality data foundation.
• Chemical Knowledge Graph Construction
  Defining knowledge graph schemas, designing ontology models, and structuring vast amounts of chemical information into computable and inferable knowledge systems.
• Applications of Knowledge Graphs
  Integrating RAG, LLMs, and knowledge inference techniques to enable accurate and traceable Q&A, support chemical reaction pathway recommendations and node inference predictions, and provide intelligent decision-making tools for chemists.

Future Prospects

We aim to further explore the deep integration of large models and knowledge graphs in chemical research, enhancing AI-driven scientific discoveries. The research group looks forward to broad collaborations with academia and industry, jointly pushing the boundaries of intelligent chemical research.