Comment on End-to-end differentiable learning of protein structureparentComments−eutectic8yAny environment with finite action and state-spaces.−rytill8yNo, the key requirement which makes it difficult to use on real-world tasks is that you must be able to do a forward rollout of your environment in your decision-making process.
Comments
Any environment with finite action and state-spaces.
No, the key requirement which makes it difficult to use on real-world tasks is that you must be able to do a forward rollout of your environment in your decision-making process.