Original article was published by Nickodemus Richard on Artificial Intelligence on Medium
I have said before about Artificial Intelligence. In a brief explanation, artificial intelligence is where the machine can learn by itself and reinforcement learning is part of it.
Have you train a pet or heard about training a pet? When we train a pet we give set of instruction to follow, our pet will act on that command, if it is right we give it reward, and punishment if wrong. The pet will learn from experience what it should do to get more reward and avoid punishment. This type of learning is what we try to mimic in reinforcement learning. Taking this analogous, the pet is called an agent. It will explore the environment and do some action. Based on its action it will be given reward, the reward is either positive or negative. The goal of the agent is to get as many positive reward as possible.
So before we try to implement our model on real car, which can be dangerous and cost a lot. AWS gives us opportunity to implement and also learn reinforcement learning at fly. Then we also have AWS DeepRacer league where we can race with others. Here is the AWS DeepRacer definition from its website.
Get hands-on with a fully autonomous 1/18th scale race car driven by reinforcement learning, 3D racing simulator, and global racing league.