On this Website

This website provides visualizations and demos of work done during my PhD in natural language processing. One project focuses on generation of natural language landmark navigation instructions and the other on vision and language navigation. To realise these projects I first gathered a dataset of human written navigation instructions for given routes.


My name is Raphael Schumann and I am a PhD student at Heidelberg University. The work presented on this website was made possible by a Google Focused Research Award.
Please cite the following work when using or referencing the map2seq data and models presented on this website:
title= "Generating Landmark Navigation Instructions from Maps as a Graph-to-Text Problem",
author= "Raphael Schumann and Stefan Riezler",
year= "2021",
publisher= "Association for Computational Linguistics"}