Transfer and Reinforcement Learning for Hedging