Dr. Fuhai Li is an Associate Professor in the Institute for Informatics (I2), Department of Pediatrics and Computer Science & Engineering (CSE) of Washington University in St Louis (WashU). Our lab specializes in integrative and interpretable graph AI and LLN model development, and applications in precision health and medicine (AI4PHM).

Research Interest

My lab’s research interests lie in modeling and interpreting large-scale, complex graphs. I am particularly enthusiastic about developing interpretable graph AI (xGAI) models, graph representation learning, generating large-scale graph datasets, building graph foundation models, and graph understanding and reasoning via large language models (LLMs) and Graph AI models. I believe that graph foundation models, using GraphAI, LLMs, and text-graph multimodal models, are among the most important foundations paving the way for general AI to integrate and interpret massive growing graph datasets along with accumulated prior knowledge. Furthermore, by leveraging powerful graph foundation models and autonomous AI agents, nontrivial graph data analysis tasks can be automated, making graph AI computing power more accessible to users with domain knowledge but without training in AI or data mining. My long-term goal is to develop graph foundation models and AI agents to automate graph understanding and reasoning in real-world systems, thereby speeding up scientific discovery in a scalable, robust and trustworthy manner. Applications focus on AI for precision health and precision medicine (AI4PHM), which are one of the most challenging, fundamental and impactful scientific problems in AI research to cure lethal and painful diseases.

Motivation/Impact

AI is accelerating scientific discovery. Artificial Intelligence (AI) has emerged as a transformative tool, fundamentally reshaping the paradigm of scientific research. AI empowers researchers to explore new frontiers in science and push the boundaries of knowledge further. Developing innovative solutions to address fundamental scientific problems is more effective and powerful when we leverage AI’s capabilities. In the real world, many complex systems can be represented using interconnected graph-structured data, such as knowledge graphs, cell signaling graphs, and chemical structure graphs. However, understanding and reasoning on large-scale and complex graphs remains an open problem. Graph AI, based on graph neural networks (GNN) models, is one effective approach to decoding large and complex graphs for knowledge discovery. AI Applications in AI4PHM: AI for precision health and precision medicine (AI4PHM) is one of the most challenging, fundamental and impactful scientific problems in AI research. Most of us will die from diseases, which are irreversible, untreatable, and incurable. Some fundamental problems of AI4PHM are to decode complex cell signaling systems; and to predict precision medicines that can correct the dysfunctional parts of the signaling systems. If can decode cell signaling systems, we can better create effective precision medicines for disease prevention and treatment.

Welcome to join FuhaiLiAI lab

If you are interested in our research, please contact me: Fuhai.Li@wustl.edu