Learning Artificial Intelligence by Making (Reinforcement Learning)
This course explains reinforcement learning without math. You can learn the concepts easily and clearly. In addition, you can implement and run an actual tic-tac-toe game by coding RLkit, which is written in Python, the most accessible language.
100 learners
Level Basic
Course period Unlimited
News
1 articles
Hello. I am Kwang, an instructor of "Learning Artificial Intelligence through Making (Reinforcement Learning) in inflearn."
First of all, thank you for taking the class.
The lecture "Learning AI by Making" started with my impulse to make something like AlphaGo. In fact, AlphaGo was not a simple Q-learning algorithm, but a complex one that used a neural network called DQN (Deep Q-learning network). Tensorflow was a huge barrier for me to learn and use DQN. So I trained Tic-Tac-Toc only with simple Q-learning.
Now, I am trying to make a cool guy using DQN, the flower of reinforcement learning. And I am making this process into a lecture. There is so much to say and learn, so I will divide it into small lecture units and make it. The first lecture, "Math of Tensorflow that you must know" has been released.
Tensorflow network learning is ultimately a process of executing gradient descent. Gradient descent? We have heard of that unknown mathematics, but we do not understand it with mathematics. I think that if we do not understand the algorithm with mathematics, we do not understand it clearly. So, before going into Tensorflow in earnest, I created a lecture to learn about gradient descent.
The ultimate goal of this series is to implement the original tic-tac-toe with Tensorflow.
https://www.inflearn.com/course/%ED%85%90%EC%84%9C%ED%94%8C%EB%A1%9C%EC%9A%B0-%EC%88%98% ED%95%99#
I hope you will join us on this ship. Thank you.
