Project Grant 2334997

This $150,000 Project Grant, awarded by the National Science Foundation's (NSF) Computer and Information Science and Engineering (CISE) program (CFDA 47.070), is supporting research on improving anomaly detection techniques for the Internet of Things (IoT). The project aims to address technical gaps in the widespread adoption of novelty detection models by developing methods to generate labeled datasets and enable the transfer of models from controlled lab settings to real-world IoT deployments. Specifically, the researchers at Columbia University are creating a comprehensive dataset of IoT network traffic, user activity, and device information, including instances of anomalies such as network attacks. They are also developing robust machine learning techniques to handle deviations from learned traffic patterns, particularly in unbalanced datasets. The broader impacts include the release of an open-source software library for machine learning on network traffic data, as well as educational initiatives to promote the use of machine learning for networking. This project is expected to be completed by April 30, 2025.