search

sigeisler / robustness_of_gnns_at_scale

This repository contains the official implementation of the paper "Robustness of Graph Neural Networks at Scale" (NeurIPS, 2021).
MIT License
25 stars 8 forks source link

Scalable Defense #7

Open MinoryHero opened 1 month ago

MinoryHero commented 1 month ago

How can I get scalable defense mentioned in the paper?

sigeisler commented 5 days ago

Hi,

Thank you for reaching out and sorry for the delay. If you refer to a specific setting please let me know. As we detail in the readme, e.g., 

python experiments/experiment_train.py with "dataset=cora_ml" "seed=0" "model_params={\"label\": \"Soft Median GDC (T=1.0)\", \"model\": \"RGNN\", \"do_cache_adj_prep\": True, \"n_filters\": 64, \"dropout\": 0.5, \"mean\": \"soft_median\", \"mean_kwargs\": {\"temperature\": 1.0}, \"svd_params\": None, \"jaccard_params\": None, \"gdc_params\": {\"alpha\": 0.15, \"k\": 64}}" "artifact_dir=cache" "binary_attr=False"  "make_undirected=True"

will execute the Soft Median GDC on cora_ml.

Best, Simon