·3 min·Deep RL with Unity 3/3
DDPG on Reacher: Twenty Arms Chasing a Balloon
DDPG solves Unity's Reacher in 101 episodes with twenty arms sharing one replay buffer, and why the noise that failed on Tennis worked here.
Reinforcement Learning
3 posts
·3 min·Deep RL with Unity 2/3
Two DDPG Agents Learning to Play Tennis
Two DDPG agents learn to keep a tennis rally going in Unity. What failed for 20,000 episodes, and the three changes that finally solved it.
·3 min·Deep RL with Unity 1/3
Training a DQN Agent to Collect Bananas in Unity
A DQN agent learns to collect bananas in Unity. Here are the Q-network, the replay buffer, the target network and how it reached +13 in 481 episodes.