
Lecture 1.1 - Why Graphs Graphs : a general language for describing and analyzing entities with relations/interactions 많은 type의 graph가 존재한다 ; event graphs, computer networks, disease pathways, social...

Design features for nodes/links/graphs

Node Embeddings Input Graph → Sturucture Features → Learning Algorithm → Prediction by "Representiation Learning" (feature eng.를 대체하도록) Graph Representation Learning Goal : Efficient task-independent ...

We investigate graph analysis and learning from a matrix perspective.

Main Question: Given a network with labels on some nodes, how do we assign labels to all other nodes in the network?
Limitations of Graph Neural Networks A "Perfect" GNN Model A k-layer GNN embeds a node based on the K-hop neighborhood structure A perfect GNN should bhild an injective function b/w neighborhood struc...

Previous Lectures Message Passing Neural Nets Node features are updated from iteration t to t+1 via learnable permutation invariant neighborhood aggregate AGG and update UPD Graph Neural Networks Mess...

추천 readingsLIME (Local Interpretation)SHAP (attribution)GNNExplainerGNN Explainability TaxonomyTrustworthy Graph Neural NetworksGraphFramEx Evaluation