RL Series - DDPG

This is part of my RL-series posts.

This algorithm is from the “Continuous Control with Deep Reinforcement Learning” paper and uses the ideas from deep q-learning in the continuous action domain and is a model-free method based on the deterministic policy gradient.

In Deterministic Policy Gradient (DPG), for each state, we have one clearly defined action to take (the output of policy is one value for action and for exploration we add a noise, normal noise for example, to the action). But in Stochastic Gradient Descent, we have a distribution over actions (the output of…