About the agent updating and initialization

[PkuDavidGuan]

I have two questions about the nice paper "Learn to Augment: Joint Data Augmentation and Network Optimization
for Text Recognition":

1.  In LIne 9 of the Algorithm 1, why the Agent network update towards -S'? I don't understand why -S' is a harder moving state.
2. As for the agent initialization, what is the initialization direction of the 2*(N+1) fiducial points?

[Canjie-Luo]

1. Note that for the modified point, S == -S'. The S' decrease the difficulty, so we use S (-S') to update the agent.
   
2. The direction is randomly initialized.