| Technical details | |
|---|---|
| Submit file | ** |
| Languages | It needs to be completed in the language you are working on right now. If you are doing Bootcamp Javascript, then javascript (file extension will be .js). If you are doing Bootcamp Ruby, then Ruby (file extension will be .rb). It goes the same for Python, Java, C++, Rust, ... |
Let's play Atari games, but with deep learning :)
It's time to dive in and learn about neural networks and deep learning! This is a significant but fun project that contributes value to your technical portolio and software development experience.
Your Mission Your mission is to build an AI that "plays" Atari video games. You will build 3 diffeent models that solve three different Atari games.
You will use the help of neural networks to solve problems using reinforcement learning and the Deep Q-Network (which represents the optimal action-value function as a neural network instead of a table).
Atari games have a large variety of screens, among other things, rendering it unsolvable using a Q-table. Learners must create an algorithm that "plays" Atari video games better than humans (and using all 49 Atari games to train the model is recommended!!). GPUs are becoming indispensable for learning problems that involve large neural networks. We will be using GPUs for training networks on larger-scale tasks.
It would be best if you started with OpenGym and I would start with CartPole
Your deliverables:
- a model that plays Cart Pole
- a model that plays Space invaders
- a model that plays Pacman
- A blog post explaining your approach