Wei Jin

There is no part missing but you may need to tune the parameters to get good performance.

Sorry that I was busy with job search. Thank you for the comments and I just resolved this issue.

Hi @likuanppd, thanks for the insightful work! I really like the observation about the connection between robustness and distribution shift. It is interesting to see how it explains some phenomena...

Hi, `ptb_rate` means perturbation rate, which is the ratio of changed edges to total edges. For example, if we have 1000 edges in the original graph, `ptb_rate=0.05` indicates that we...

Hi, the full graph of cora dataset has 2708 nodes while the largest connected component (LCC) of cora has 2485 nodes. Suppose you are using `deeprobust.graph.data.Dataset`. By setting `setting='gcn'`, you...

Hi, Thanks for your interest in DeepRobust. I wanted to clarify a few points here: 1. The most straightforward way to attack other models (GAT and GraphSAGE) is to use...

Hi, can you try it again? It works well on my side "pip install deeprobust==0.2.8"

Hi, we are currently working on it and plan to release the next version within the upcoming weeks, likely within the next two weeks.

I am not exactly sure about the described scenario, but we are supposed to use **the same data split** before and after attack. This ensures that the attacker uses the...

Hi, The input adjacency matrix already contains the injected nodes. Please see the file below. https://github.com/DSE-MSU/DeepRobust/blob/fb047ac918ca3525749b8d091898fc5670752e74/examples/graph/test_nipa.py#L60-L62