This repository contains my solutions for the 2025 Probabilistic Artificial Intelligence course at ETH Zürich. The course was held by Prof. Dr. Andreas Krause. The solutions are not solely mine. They contain the prewritten code that was written by the lecturers, and also the work was done in cooperation with my teammates.
- Project 1: We used Gaussian Processes for regression to predict air pollution given other data.
- Project 2: We implemented a Bayesian Neural Network to classify satellite images. For the approximation of the posterior we used Stochastic Weight Averaging-Gaussian (SWAG).
- Project 3: We implemented Bayesian Optimization to tune the structural features of a drug candidate.