Reproduce “Leakage-free evaluation of residual graph convolutional networks for network intrusion detection” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Zahid Iqbal
The authors evaluate a residual graph convolutional network for intrusion detection with training-only flow graphs; test flows attach to the training reference. On CICIDS2017, its binary F1 was 98.11%, 1.57 points below XGBoost and not measurably different from an MLP. Under Gaussian noise with standard deviation 0.5, its F1 was 96.16%, versus 0.67% for XGBoost, a result limited to that synthetic test.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.