Objetivo

Analisar, reproduzir e possivelmente incrementar soluções existentes de Reinforcement Learning para os desafios propostos na competição MineRL.

Refêrencias

1. MineRL
2. MineRL Docs
3. MineRL 2021
4. The MineRL competition
5. Behavioural cloning baseline for the Intro track
6. Keras
7. Deep Q-Learning for Atari Breakout
8. Image classification from scratch
9. KerasRL
10. Deep Q-learning from Demonstrations
11. Boosted Bellman Residual Minimization Handling Expert Demonstrations
12. Meta Learning Shared Hierarchies
13. Discriminator Soft Actor Critic without Extrinsic Rewards
14. minerl_imitation_learning