Model extraction attacks on graph neural networks: Taxonomy and realisation

摘要: … studies about the model extracted attacks targeting at ordinary Deep Learning system’s privacy, … It demonstrates that obtaining more background knowledge can enhance our extraction …