Submitted by Timo K. 5 Understanding and Improving Hyperbolic Deep Reinforcement Learning University of Vienna 2