Local Post-hoc Explainers for Graph Neural Networks

PhD Qualifying Examination


Title: "Local Post-hoc Explainers for Graph Neural Networks"

by

Miss Linrui LI


Abstract:

Graph Neural Networks (GNNs) have emerged as powerful tools for learning and 
reasoning over graph-structured data, achieving remarkable success in 
applications such as healthcare, molecular analysis, and recommendation 
systems. However, their increasing use in critical domains highlights the 
need for explainability to ensure trust and transparency. This survey focuses 
on local post-hoc explanation methods for GNNs, categorizing them by how they 
measure feature importance. For each, we summarize the core idea, advantages, 
and limitations, offering a concise overview of GNN explainability. We also 
review commonly used datasets and evaluation metrics.


Date:                   Thursday, 19 December 2024

Time:                   12:00noon - 2:00pm

Venue:                  Room 2128C
                        Lift 19

Committee Members:      Prof. Lei Chen (Supervisor)
                        Prof. Qiong Luo (Chairperson)
                        Dr. Dan Xu
                        Dr. Linping Yuan