| CPC H04L 63/1425 (2013.01) [G06F 11/327 (2013.01); G06N 20/00 (2019.01)] | 20 Claims |
|
1. A method for detecting an anomalous state in a machine performed by an anomaly detection server, the method comprising:
receiving a plurality of training analysis vectors associated with a monitored machine during a training period, wherein each of the plurality of training analysis vectors describe a condition of the monitored machine at a corresponding point in time;
applying the plurality of training analysis vectors to a machine learning model to create a trained machine learning model configured to describe normal patterns of states in the monitored machine across sequences and timing as indicated by the plurality of training analysis vectors;
receiving a plurality of monitoring analysis vectors associated with the monitored machine during a monitoring period, wherein each of the plurality of monitoring analysis vectors describe a condition of the monitored machine at a corresponding point in time;
applying the plurality of monitoring analysis vectors to the trained machine learning model to identify monitored patterns of states in the monitored machine across sequences and timing as indicated by the plurality of monitoring analysis vectors and scan for at least one discrepancy indicating an anomalous state in the monitored machine by comparing the monitored patterns of states across sequences and timing to the normal patterns of states across sequences and timing; and
on condition that a discrepancy exists between the monitored patterns of states across sequences and timing and the normal patterns of states across sequences and timing, transmitting an alert indicating that an anomaly is detected in the monitored machine.
|