Open Source

Hands-on assignments for the Advances in AI lecture covering PyTorch, ResNet, GANs, VAEs, embeddings, perceptual/adversarial loss, and TensorBoard.

Practical exercises and examples for the Machine Perception & Tracking lecture, providing course tasks and supporting material for students.

Tracking assignment for the Machine Perception & Tracking lecture, including evaluation code and baseline tracker implementations.

Rust-based bot controller using async Tokio tasks, behaviour-tree logic, a neural-network controller, and a genetic algorithm for training.

Online textbook for the Programming for Data Science & AI lecture with programming examples, exercises, and supporting material.

Natural-language image retrieval demo using OpenAI's CLIP model - scans image directories, computes embeddings, and retrieves images from text queries.

Wasserstein GAN implementation for generating artificial images.

Variational autoencoder for learning a data distribution and sampling new images, demonstrated on a flower dataset.

Chess-AI programming assignment for the Programming for Data Science & AI course, with chess engine/UI code, tests, sprites, and documentation.

Image upscaling experiment that trains step-wise 2x upscalers and stacks them for 8x upscaling from 64x64 to 512x512 pixels.

Reinforcement learning example using Q-learning for the Lunar Lander environment.

Image colorisation experiment covering grayscale-to-colour reconstruction with YUV preprocessing, discussing regression vs. classification as the learning objective.

NLP model that predicts which Star Trek character said a given phrase, trained on official episode scripts.

Semester assignment for Machine Perception & Tracking where student teams analyse football matches and present a working algorithmic solution.