Online Cyber-Attack Detection in the Industrial Control System: A Deep Reinforcement Learning Approach

Figure 3

The value of accurate varies with the number of training sessions.