Mitigating the Insider Threat with High-Dimensional Anomaly Detection

摘要

In this project, we explored new techniques for detecting the threat of insider attacks in enterprise networks. In particular, we explored the use of high-dimensional search techniques such as Latent Semantic Indexing to mitigate the problem of high dimensionality that is inherent in intrusion detection. This new technique can be used for both labeled and unlabeled detection, and shows promise for detecting attacks and anomalies earlier than previously possible and detecting attacks that are similar to past ones.