All-to-All Self-Supervised Reinforcement Learning Neural Network