Explore the decomposition of the Q value function (which acts as the critic) in actor-critic algorithms for a singular learning agent

Interface Baxter robot simulation over Docker and ROS, to be deployed on server

Train a reinforcement learning agent (for Sawyer robot simulation) on server through Docker

Testing of composition reinforcement learning (Soft Actor Critic) agent on 7 DOF robot

Solving the visual question and answer dataset with a proposed network.