Exploration Entropy for Reinforcement Learning

Figure 2

A navigation problem in a 5 × 5 maze. (a) Maze A. (b) Maze B.

(a)

(b)