Intro to AI
A semester of AI implementations from UW–Madison's CS540.
The course worked through PCA, hierarchical clustering, linear regression, neural networks, deep learning, and game AI with minimax and A*. Everything was implemented from the ground up in Python. Here are three of the ones I found most interesting.
A few in more detail
PCA image projection
Using PCA on the Yale face dataset (32×32 images): compute the covariance matrix and its eigenvectors, then reconstruct each face from its top components for a smoother, compressed result.
Neural network clothing classifier
A neural network trained on a labeled clothing dataset (shirts, pants, jackets, shoes, boots) that predicts the item in a new image, and reports its accuracy on held-out data.
A* on the 7-tile puzzle
An A* solver that finds the optimal, fewest-move path from any board state to the goal, sliding one tile at a time into the open space.