Udacity Deep Learning Nanodegree Foundation - Exercises

(11.2017 - 04.2018)

About

Here you can find some selected exercises from the course

Part 6: Reinforment Learning

Project: Frozen Lake with Dynamic Programming

• the environment is completely know to the agent

Project: Black Jack with Monte Carlo Method

• the environment is not known to the agent
• the agent learns by interaction
• the agent learns in episodes

Project: Cliff Walk with Sarsa

• the environment is not known to the agent
• the agent learns by interaction
• the agent learns in each step

Project: Q-Learning-Cart

• discrete action space
• continous state space
• Deep Learning applied
• Keras Implementation
• Pytorch Implementation