Decision Making
Note: I took the core codebase for environment simulation and related infrastructure in this repository from my assignments in the Deep Decision Making and Reinforcement Learning (DDRL) course at New York University. I only implemented the Behavior Cloning, Goal-Conditioned Behavior Cloning, DAgger, Double Q-Learning, Dueling DQN, and PPO algorithms. Their performance are shown in the GIFs below.
Environment
Expert Dataset
Changing Goal
Episode 1
Episode 2
Episode 3
Fixed Goal
Episode 1
Episode 2
Episode 3
Behavior Cloning
Changing Goal
Fixed Goal
Multimodal
Goal Conditioned Behavior Cloning
Changing Goal
Fixed Goal
