LoadingByte/are-gnn-defenses-robust
Adaptive evaluation reveals that most examined adversarial defenses for GNNs show no or only marginal improvement in robustness. (NeurIPS 2022)
Jupyter NotebookMIT
Issues
- 1
How to design adaptive attacks against complex models or even non-differentiable models?
#4 opened by zhongjian-zhang - 2
About MongoDB
#2 opened by KanghoonYoon - 1
Why `Tapeable*`?
#3 opened by bwdeng20 - 3
About Soft-Median
#1 opened by chris-hzc