Reinforcement Learning in Hyperbolic Spaces: Models and Experiments