If you want to know more about it, Kumar et al. Note that Imitation Learning is not always more efficient than RL. This technique is also called Behavior Cloning, which is the simplest form of imitation. We don’t have that much time to spend, so we’re going for the Imitation Learning solution. According to the authors of the MineRL 2021 competition, it takes 8 hours for the pure RL solution and 15 minutes for the imitation learning agent to reach the same level of performance. The two approaches have the same outcome, but they’re not equivalent. In this case, it is a sequence of actions to chop trees made by a human.
#We need to go deeper minecraft how to
Imitation learning: the agent learns how to chop trees from a dataset.It is rewarded every time it chops a tree. Pure deep RL: the agent is trained from scratch by interacting with the environment.More specifically, deep RL seems to be the solution since we’re processing images to select the best actions. Naturally, reinforcement learning is a pertinent framework to train this agent.
Instead of scripting orders, we want an AI that knows how to chop trees. This approach is too static for our requirements: we need something that can adapt to new environments. Our bot works well in a fixed environment, but what happens if we change the seed or its starting point?Įverything is scripted so the agent would probably try to chop a non-existent tree. This is a good start, but we would like to do it in a more automated way… 🧠 II. The agent efficiently chopped the entire tree.
0 kommentar(er)
