rogerpanel/robustidps.ai: v1.1.0 - Initial Release

RobustIDPS.ai v1.1.0 — Initial Release Research-grade web application demonstrating 7 + 7 novel ML methods for adversarially robust network intrusion detection, built as part of a PhD dissertation at MEPhI, Moscow. Highlights Web Dashboard — Real-time monitoring, threat overview, and SOC-style analytics Upload & Analyse — Drag-and-drop CSV/PCAP analysis with MC Dropout uncertainty quantification Live Traffic Streaming — WebSocket-powered row-by-row classification Ablation Studio — Toggle any of the 7 dissertation methods and measure accuracy impact Analytics & Benchmarks — Pre-computed metrics across 6 datasets, 4 adversarial attacks, privacy-robustness trade-offs 5 Detection Models — Surrogate Ensemble, Neural ODE, Optimal Transport, FedGTD, SDE-TGNN Architecture | Layer | Technology | |-------|-----------| | Frontend | React 18, TypeScript, Tailwind CSS, Recharts, Vite | | Backend | FastAPI, Python 3.10+, PyTorch 2.2, pandas, scikit-learn | | Streaming | WebSocket live per-flow classification | | Infrastructure | Docker Compose, nginx, Cloudflare SSL | Dataset Support Auto-detects 6 benchmark formats: CIC-IoT-2023, CSE-CIC-IDS2018, UNSW-NB15, Microsoft GUIDE, Container Security, Edge-IIoT, plus raw PCAP via NFStream. Citation @phdthesis{anaedevha2024robustidps, title={Advanced AI-Powered Intrusion Detection Systems: Neural ODEs, Optimal Transport, and Federated Learning}, author={Anaedevha, Roger Nick}, year={2026}, school={National Research Nuclear University MEPhI} } ## What's Changed * setup robustidps by @rogerpanel in https://github.com/rogerpanel/robustidps.ai/pull/1 ## New Contributors * @rogerpanel made their first contribution in https://github.com/rogerpanel/robustidps.ai/pull/1 **Full Changelog**: https://github.com/rogerpanel/robustidps.ai/commits/v1.1.0