More Than Zero - Diagnosing Unknown Attacks Using Abductive Reasoning

This dataset was used to implement and validate our technique that uses abductive reasoning to diagnose unknown attacks. The repository contains:

datasets.zip -  This contains files from two datasets, CICIoT2023 and IoT-23, that have been cleaned and prepped for our experiments. The original contributors of the individual datasets can be found in the link provided.

anomalies.zip - This contains all the anomalies we detected from the datasets and encoded programmatically as Answer Set Programs. This representation is useful for logic-based reasoning techniques (to identify violated security requirements), and symbolic learning techniques (to evolve those security requirements).

The source code of our technique is available in our GitHub repo.