Explore projects

From a 'root' Wikipedia article, a tree is constructed, where the child nodes are random subset of the articles linked in the parent node article. Used to generate datasets of varying size and similarity to the root article for NLG models.

Spatially invariant integrative voxel embedding for morphologically variable HCR stained samples

The purpose of this Jupyter Notebook is to summarise theory and demonstrate the implementation of variational autoencoders, such that I should be able to use this as a VAE cheat sheet in a few months from now.

Public repository for the submission "Building Floorspace in China: A Dataset and Learning Pipeline" (Egger, Rao and Papini, 2023).

A lightweight and portable toolchain that allows to systematically quantify, optimize, and analyze NNs (at layer granularity) by using TensorFlow (+Lite, +Lite Micro) and ARM Mbed OS.

This repository contains the data and analysis-, evaluation- and plotting scripts.