ianupk/idsgan

How ianupk/idsgan works

IDSGAN’s main workflow turns NSL-KDD traffic into encoded records, trains black-box intrusion detectors, then trains a restricted-modification WGAN against their predictions to generate evasion examples. Evaluation measures detection and evasion rates across attack groups and IDS models; a separate baseline workflow compares other attacks. The graph also preserves dataset acquisition, feature policy, model persistence, and reporting. Relationships for unsampled modules are included only where the README or SOURCE INDEX supports them; the excerpts do not expose every runtime call or checkpoint detail.

Main entry points: data_preprocessing.py Preprocess and split · evaluate.py Evaluate evasion · idsgan.py IDSGAN trainer · train_idsgan.py Train IDSGAN

Components 17

Data preparation

Dataset download download_data.py
Preprocess and split src/data_preprocessing.py
Feature policy src/feature_utils.py
Configuration src/config.py

IDS models

Train IDS models train_ids.py
IDS model factory src/ids_models.py
IDS checkpoints

Adversarial generation

Train IDSGAN train_idsgan.py
IDSGAN trainer src/idsgan.py
Restricted generator src/generator.py
WGAN critic src/discriminator.py
GAN checkpoints

Evaluation

Evaluate evasion evaluate.py
DR and EIR metrics src/metrics.py
Result figures
Baseline comparison run_baselines.py

Other components

NSL-KDD data

Connections 27

Dataset download → NSL-KDD data: downloads README.md
Dataset download → Configuration: uses paths download_data.py
Preprocess and split → NSL-KDD data: loads records README.md
Preprocess and split → Feature policy: uses definitions src/data_preprocessing.py
Preprocess and split → Configuration: uses paths src/data_preprocessing.py
Train IDS models → Preprocess and split: prepares data train_ids.py
Train IDS models → IDS model factory: trains models train_ids.py
Train IDS models → IDS checkpoints: saves checkpoints README.md
Train IDSGAN → Preprocess and split: prepares data train_idsgan.py
Train IDSGAN → IDS model factory: loads IDS train_idsgan.py
Train IDSGAN → IDS checkpoints: loads checkpoints README.md
Train IDSGAN → IDSGAN trainer: starts training train_idsgan.py
Train IDSGAN → GAN checkpoints: saves checkpoints README.md
IDSGAN trainer → Restricted generator: generates records src/idsgan.py
IDSGAN trainer → WGAN critic: scores records src/idsgan.py
IDSGAN trainer → Feature policy: gets masks src/idsgan.py
IDSGAN trainer → IDS model factory: queries predictions src/idsgan.py
Restricted generator → Feature policy: uses feature indices src/generator.py
Evaluate evasion → Preprocess and split: prepares test data evaluate.py
Evaluate evasion → IDSGAN trainer: loads generator evaluate.py
Evaluate evasion → GAN checkpoints: loads checkpoints README.md
Evaluate evasion → IDS model factory: evaluates IDS evaluate.py
Evaluate evasion → DR and EIR metrics: computes metrics evaluate.py
Evaluate evasion → Result figures: produces plots README.md
Baseline comparison → Preprocess and split: prepares test data run_baselines.py
Baseline comparison → IDS model factory: loads target IDS run_baselines.py
Baseline comparison → DR and EIR metrics: compares detection run_baselines.py

More Python diagrams: darkshapes/zodiac · darpa-criticalmaas/uncharted-ta1 · davidteather/imc-prosperity-2024 · dev-essbee/iplt20-scraper · diamondlightsource/httomo

Drawn by GitDiagram on September 30, 2026 from the repository's file tree, README and source files. Read as Markdown ·