ReAct-IDS: reproducibility artifact (predictions, analysis scripts, provenance ledger)
Year: 2026
Type: dataset
Abstract: Per-sample predictions and the scripts that turn them into the numbers reported in the paper. Unpack anywhere and run; nothing needs configuring. Python 3.10 or later. Version 1.0.2 re-establishes the C6 zero-day cohort: the repeated runs behind the C6 row of Table VIII are now 14 runs per dataset on a single reference host, replacing a cohort that spanned five hosts, and three shipped scripts that still carried superseded statistics or the withdrawn seed attribution have been corrected (CHANGES.md, Section 6); use this version.
Authors: Di, Qianli, Du, Keju, Chen, Gang, Deng, Hai
Institution: Nanjing University of Aeronautics and Astronautics
Language: English
Subfield: Computer Networks and Communications
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W7214036675
