Python application that lets you play the card game hearts. Supports up to 5 agents with any number of AI. AI either uses monte carlo learning or multi-layer random search
This project is a novel maze searching and coin-collecting game with semi intelligent enemies. The player agent is trained purely through Q-learning or REINFORCE with PPO. Agent regularly beats game on various map levels with different enemy conditions.
Computer vision project that recognizes board states and pieces, convert them into chess piece and position info for input into a chess API. Uses Hough lines and K-means clustering for board recognition and a VGG16 model for image classification.
Project for Brown’s HCI Lab, takes various influential CS department rankings and a novel PageRank metric that evaluates academic placement and combines them into an easy to use and customizable meta ranking.
Personal project that takes in input of a Settlers of Catan board state and gives the user a list of the most optimal settling locations.