Graph Foundation Models
Pre-training and prompt-based adaptation for graph data, especially transfer across signed and directed graphs.
Ph.D. Student · Arizona State University
I study graph foundation models and reliable AI, with a focus on graph pre-training, prompt learning, and domain adaptation.
I am a Ph.D. student in Computer Science at Arizona State University. My research develops transferable and adaptive learning methods for structured and relational data, with particular interests in signed graphs, few-shot learning, multimodal evaluation, and large language models.

01
Pre-training and prompt-based adaptation for graph data, especially transfer across signed and directed graphs.
Few-shot, continual, and domain-adaptive learning methods for limited supervision and changing data distributions.
Evaluation of multimodal systems and their reasoning, robustness, and behavior in socially consequential settings.
02
Our first-author paper Signed Graph Pre-Training and Prompt Learning will appear at the Fifth Learning on Graphs Conference (LoG 2026).
GroupToM-Bench: Benchmarking Group Theory of Mind and Nonlinear Social Emergence in MLLMs was published at ACL 2026.
MGP: Integrating Pre-Training and Few-Shot Node Classification via Meta Graph Prompt was published in Knowledge-Based Systems.
03
Learning on Graphs Conference (LoG), 2026
SGPT combines a structural encoder with magnetic signed Laplacian representations to learn persistent-homology-aware graph features. The resulting embeddings transfer to prompt learning for downstream signed-graph tasks.
04
Prof. Yixuan He’s Group · Arizona State University
HPC-AI Lab · National University of Singapore (Remote)
Data Mining Lab · University of Electronic Science and Technology of China
05
Ph.D. in Computer Science
Dual B.Eng. (Hons) in Electronics & Information Engineering
06
Mathematical Contest in Modeling · Top 2% globally