An Insider Threat Detection Approach Based on Mouse Dynamics and Deep Learning

Algorithm 1

Adam optimization algorithm.

1t <- t + 1

2lr_t <- learning_rate sqrt(1 - betat) / (1 - betat)

3m_t <- beta1 m_ + (1 - beta1) g

4v_t <- beta2 v_ + (1 - beta2) g g

5variable <- variable - lr_t m_t / (sqrt(v_t) + epsilon)