US Patent Application 18222340. ABNORMALITY SENSING DEVICE AND ABNORMALITY SENSING METHOD simplified abstract
Contents
ABNORMALITY SENSING DEVICE AND ABNORMALITY SENSING METHOD
Organization Name
Nippon Telegraph and Telephone Corporation
Inventor(s)
Takeshi Nakatsuru of Tokyo (JP)
ABNORMALITY SENSING DEVICE AND ABNORMALITY SENSING METHOD - A simplified explanation of the abstract
This abstract first appeared for US patent application 18222340 titled 'ABNORMALITY SENSING DEVICE AND ABNORMALITY SENSING METHOD
Simplified Explanation
The patent application describes an anomaly detection device that can detect anomalies in a communication apparatus.
- The device has a learning unit that uses a communication log as learning data to generate a detection model.
- An anomaly detection unit then uses this model to detect any anomalies in the communication apparatus.
- The device also has a data acquisition unit that acquires a second communication log generated during a later period.
- A determination unit instructs relearning using the second communication log if there is a difference between the current detection model and the second communication log.
- The relearning is only triggered if the difference includes a certain number of additional or deletion information, based on predetermined evaluation criteria.
Original Abstract Submitted
An anomaly detection device is provided with a learning unit that generates a detection model using a communication log during normal operation of a communication apparatus as learning data and an anomaly detection unit that detects anomaly of the communication apparatus using the generated detection model. The anomaly detection device is further provided with a data acquisition unit that acquires a communication log (second communication log) generated during a predetermined period later than a first communication log and a determination unit that instructs relearning using the second communication log when there is difference information between the learning data (first communication log) of the current detection model and the second communication log and when the number of pieces of additional information (information on the additional flow) or the number of pieces of deletion information (information on the delete flow) included in the difference information satisfies predetermined evaluation criteria.
