Security

Intrusion detection in Unmanned Aerial Vehicle (UAV) networks is crucial for maintaining the security and integrity of autonomous operations. However, the effectiveness of intrusion detection systems (IDS) is often compromised by the scarcity and imbalance of available datasets, which limits the ability to train accurate and reliable machine learning models. To address these challenges, we present the "CTGAN-Enhanced Dataset for UAV Network Intrusion Detection", a meticulously curated and augmented dataset designed to improve the performance of IDS in UAV environments.

To provide a standardized approach for testing and benchmarking secure evaluation of transformer-based models, we developed the iDASH24 Homomorphic Encryption track dataset. This dataset is centered on protein sequence classification as the benchmark task. It includes a neural network model with a transformer architecture and a sample dataset, both used to build and evaluate secure evaluation strategies.

This unlabeled dataset reflects the network activity of a real branch office with 29 active machines connected to the same broadcast domain for four hours. To achieve this, a Network Intrusion Detection System (NIDS) called BCAST IDS listened to network traffic every 10 seconds. During this time, various types of activities were carried out (browsing, emailing, file transfers, etc.) on each machine to ensure the dataset reflected a wide range of benign behavior.

Our dataset is constructed by leveraging existing malware samples and utilizing both UTRDCL and traditional DCL techniques to load the malicious components, thereby launching attacks. In addition to the malware samples themselves, we also provide online detection reports from reputable sources, including VirusTotal, MobSF, and Bazaar (Pithus). These reports offer a comprehensive analysis of the malware samples, enabling researchers to gain a deeper understanding of the attacks and their characteristics.