Cyber Grand Challenge - Datasets
The autonomous systems' performance was measured using the same competition rating metrics used to quantify the performance of human analysts. Competitors successfully created cyber reasoning software that automatically detected and fixed intended and unintended flaws in never-before-seen software at machine speed. The results quantify the systems' ability to reason about and mitigate novel software flaws.
Archived below are the datasets that were utilized in the creation and support of this competition, from qualifiers and finals. These materials are available freely for use in research or to assist in hosting a similar event.
The following data sets are intended for public distribution and utilize public licenses to that effect. Each file link below will include further detailed instructions.