Evolving deep unsupervised convolutional networks for vision-based reinforcement learning
Dealing with high-dimensional input spaces, like visual input, is a challenging task for reinforcement learning (RL). Neuroevolution (NE), used for continuous RL problems, has t...