Talk:Deep reinforcement learning

make an orange character }}give him legs,arms and a head }}name him John }}make John learn to walk

"Training" Section
The current "training" section is a mixture of a lot of different but very specific topics. It would make more sense to have it be an overview of deep RL algorithms, and then have a separate section on broad research directions that are being investigated: off-policy RL, inverse RL, meta-RL, goal-conditioned RL. Happy to do this myself if there is agreement. Anair13 (talk) 20:36, 24 November 2020 (UTC)