Hackers hoping to disrupt the power grid, water or natural gas service may be foiled by an intrusion detection system developed by researchers at Oak Ridge National Laboratory. The system is an expanded version of Oak Ridge Cyber Analytics, which boosts the performance of existing cyber attack systems by filtering noise and quickly making sense of massive amounts of data. The system uses machine learning, a branch of artificial intelligence, to classify critical infrastructure communications as malicious or benign and discriminate man-made from natural events. “With this approach we can trigger alarms to operators if there is something attacking the network systems,” said Raymond Borges, one of the developers. Within a few years, Borges and colleagues plan to combine data from other sources—such as physical device logs – on the network to improve the detection rate.
Menu
