Chapter10. Deep Generative Models for Graphs
CS224W by 신윤종
Deep Generative Models for Graphs

오늘은 그 반대인 Decoder 모델을 배워보

The Problem: Graph Generation
Graph Generation Tasks



why is it Hard?



Machine Learning for Graph Generation


1. Make P_model(x;θ) close to P_data(x)

2. Sample from P_model(x;θ)
Auto-regressive models

Graph RNN : Generating Realistic Graphs





GraphRNN : Two levels of RNN





Issue : Tractability


Evaluating Graphs

Applications and Open Questions

Graph Convoliutional Policy Network



Open Problems
Last updated