Some machine learning exercises. Learning goals below.
draw_tensor_model:
- Learn to build a Flask app.
- Build some basic AI models using MNIST data.
- Have an interface for testing and tweaking the models using spontaneously created data.
game_learner:
- Write some general classes for training an AI to play games using Q learning. (done)
- Train an AI to play tic-tac-toe using Q learning.
- TODO: Train an AI to play Connect 4 using deep Q learning.
- TODO: Do the same but with Monte Carlo search trees.
tensor_puzzles:
- Learn how to broadcast with numpy arrays.
gpu_puzzles:
- Learn the basics about GPU programming.