Control device, control program, and control method for anomaly detection

A non-transitory computer-readable storage medium storing a control program that realizes a control device for controlling a control target by being executed by a computer, the control device including a control logic configured to execute computations for the control target and a machine learning engine, the control program, when read and executed, causes the computer to perform:

• generating, by the control logic, a feature amount suitable for detecting an anomaly that occurs in the control target from data that relates to the control target;
• data-compressing, by the control logic, the generated feature amount, and providing the data to the machine learning engine,
• instructing, by the control logic, the machine learning engine to carrying out machine learning using the generated feature amount to detect the anomaly;
• detecting the anomaly, by the machine learning engine, based on the generated feature amount and an anomaly detection parameter determined based on a learning result of the machine learning and used in detection of the anomaly that occurs in the control target;
• wherein the detecting the anomaly comprises carrying out, after an instruction for the detection of the anomaly is transmitted, detection of the anomaly, without sending a response to the request, such that the control logic carries on with subsequent processing without waiting for the response.

A control method that is executed by a control device for controlling a control target, the control device comprising a control logic configured to execute computations for the control target and a machine learning engine, the control method comprising:

• generating, by the control logic, a feature amount suitable for detecting an anomaly that occurs in the control target from data that relates to the control target;
• data-compressing, by the control logic, the generated feature amount, and providing data to the machine learning engine,
• instructing, by the control logic, the machine learning engine to carrying out machine learning using the generated feature amount to detect the anomaly;
• detecting the anomaly, by the machine learning engine, based on the generated feature amount and an anomaly detection parameter determined based on a learning result of the machine learning and used in detection of the anomaly that occurs in the control target;
• wherein the detecting the anomaly comprises carrying out, after an instruction for the detection of the anomaly is transmitted, detection of the anomaly, without sending a response to the request, such that the control logic carries on with subsequent processing without waiting for the response.