LLM-based Security-Compliance in Event Streams (Bachelor Thesis, Ongoing)

Author

Valentin Neher

Description

Investigation of the use of K-means clustering to establish a network baseline for anomaly detection in Operational Technology (OT) environments, whereby deviations from learned normal behavior are identified as potential security incidents. A large language model (LLM) will then contextualize the incidents, describe them in natural language, and perform a risk assessment.

Start / End Dates

2026/04/20 - 2026/08/19

Supervisors