# DDPG

An off-policy actor-critic algorithm that learns a deterministic policy over continuous action spaces.

## What it is used for

Early continuous-control problems; now mostly replaced by TD3 and SAC.