PhD Dissertation Proposal by XIAO Hanhua | From Robustness to Insight: Advanced Methods for Outstanding Facts Mining

ABOUT THE TALK

Outstanding fact mining is crucial for extracting significant and exceptional information from large datasets, providing valuable insights in various domains such as knowledge discovery, data analysis, and artificial intelligence. This dissertation focuses on advanced methods to enhance the robustness and generalizability of outstanding fact mining techniques.

The first part of this work addresses measuring robustness of outstanding facts mined from knowledge graphs. Using perturbation techniques, we evaluate the stability of these facts when the graph structure is modified. To extend the generalizability of our methods, the second part of this dissertation proposal explores perturbation on general graph patterns and subgraph counting. By combining robustness evaluation and advanced perturbation techniques, this dissertation proposal aims to provide a comprehensive framework for outstanding facts mining, ensuring the reliability and applicability of extracted knowledge across different domains.

ABOUT THE SPEAKER

Xiao Hanhua is a PhD Candidate in Computer Science at the SMU School of Computing and Information Systems, supervised by Assistant Prof. Li Yuchen.  His research is focused on knowledge graph mining and LLM on graph analytics.