The MineRL 2020 Competition on Sample Efficient Reinforcement Learning using Human Priors