Research Experiences

Master Dissertation, Prof. Xiaokui Xiao

Leverage LLM-assigned persona labels to construct a tripartite graph and apply rec-models like LGCN. Improve major accuracy metrics by 5-15% when intuitively combined with the sota GNN-based recommender system.

Graduation Project, Prof. Jiaping Gui

Quickly and effectively perform anomalous node detection and reliability analysis. Realize the general application from offline collected data sets to public data sets based on the graph theory algorithm. Form a journal article in Cyber Security and Data Governance.

Summer Research, Prof. Zhou Li

Start with basic Graph neural network model on anomaly detection like DeepWalk, NetWalk and Euler. Apply NetWalk on LANL dataset to help detect unusual online flows. Adapt existing model for anomaly detection, and make fair comparison.

College Innovation Program, Prof. Jie Guo

Use BERT, LDA and other models in the NLP field to help analyze harmful speeches on the Internet. Define anonymous Internet users, establish relationship networks, and identify key users. Our project won the 2022 Outstanding College Innovation Project in Shanghai.