Reinforcement Learning Projects

In this repository, you'll find my solutions to the CS294-112 course homework about DeepRL offered by UC Berkeley in Fall '18. Specifically, this course covers implementation offerings of the following algorithms:

1. Imitation Learning with DAgger
2. Policy Gradient with variants: vanilla, reward to go, reward normalized, baseline average, baseline estimation w/ neural network
3. Q-Learning + Actor-Critic incl. Double DQN
4. Model-based RL with a dynamics model and MPC
5. Advanced topics
   1. Exploration with histogram, RBF-Kernel, EX2
   2. Soft Actor-Critic with Maximum-Entropy method
   3. Meta-Learning with feed-forward / GRU based contextual policies