This repository contains my solutions for the 2025 Introduction to Machine Learning course at ETH Zürich. The course was held by Prof. Dr. Andreas Krause and Prof. Dr. Fanny Yang. The solutions contain my (and my team's) implementation on top of the prewritten code that we received from the lecturers
- Project 1: Consists of an A and a B part. In part A, we implement ridge regression with cross-validation, and in part B, we perform linear regression again, but with gradient descent instead of the closed-form solution.
- Project 2: Also contains linear regression, but here, the data contains missing values. Therefore, we first deal with the problem, then cross-validate different kernels and choose the best one for the final model.
- Project 3: Implements an encoder-decoder model for the reconstruction of images. The input numbers are cropped at the middle of the image, and the trained model reconstructs the original image.
- Project 4: Fine-tunes a pre-trained GPT2 model for binary classification.