The puzzle solver is a classic and a pretty neat AI demonstration which shows how an agent can solve a "sliding puzzle" like the one below. This project compares several different algorithms including Best First Search and A* with three different heuristics: Manhattan distance, misplaced tiles, and row/column misplacement. I first wrote the solver for the puzzle with 8 pieces and later added a script to solve the larger 15 piece puzzle, where a sample solution is shown on the right.

Another well-known AI puzzle, a solution to the 8-queens puzzle involves finding a way to place 8 queen pieces on an 8x8 chessboard in such a way that none are "attacking" another. This project uses a genetic algorithm to find a solution.

Reinforcement Learning Robot

This is a project I'm currently working on.

Euphonomimesis is an experiment in the application of a genetic algorithm to audio samples in the time and frequency domains. The name came about as a portmanteau of euphonic (“pleasant sounding”) and mimesis (“imitation of the real”). We developed this project based on our interest in applying AI/ML algorithms to digital audio for creative purposes. Created by a team of three students.