Course projects related to Artificial Intelligence (COMP 424), Computer Vision (ECSE 415), and Applied Machine Learning (COMP 551).
Analyze, visualize, and compare the performance of k-nearest neighbors, decision trees, and time-series analysis in predicting COVID-19 hospitalization cases from related symptoms search, across time and regions in the US.
Implement from scratch the Softmax Regression model fitted with a batch gradient descent optimizer, tuning hyperparameters using grid search, and early-stopping capability to avoid overfitting.
Deep Neural Network trained by automatic differentiation and implemented in PyTorch, to classify image data using a modified MNIST dataset. Reached a 99.4% accuracy.
Computer Vision algorithm for modeling and analyzing movements of people in a surveillance video of a mall, based on a Support Vector Machine person detection technique.
Implemented from scratch and compared the performance of AI algorithms minimax, alpha-beta, and Monte Carlo Tree Search to develop an AI agent for a variant of the Saboteur board game in Java.