Doing Legal Research Differently

Where ( offers three very challenging NLP thesis topics. We are an Amsterdam bases startup incubated by Deloitte. While writing your thesis you will become part of the team and help out with various CS related problems. Depending on your skills and wishes, we can make it work.

Legal argumentation miner

Within court cases, different parties bring in arguments to defend their legal position. For this topic you will write a methodology to automatically mine these arguments from court cases. Data and infrastructure is available. You are to use a combination of NLP, process and data mining.

Multilingual classification

European Court of Justice case law is classified according to a standard called EuroVoc. For this topic you will attempt to classify document from each member states in that EuroVoc classification. Challenge: this is in the local language of that member state. You will most likely focus on multi-lingual embedding techniques. Data and infrastructure is available.

Multilingual document association

With each European country having their case law published in its own language, it is very hard to find (very) similar court cases in for example The Netherlands and Germany. However, since all EU memberstates have to comply with EU Law, this will become more important over time. For this topic you will attempt to link millions of legal documents to their most similar ones (in a different language). Data and infrastructure is available.

An example of a TU/e Master project completed at in 2021:
Michael van der Veen. Mining Local Process Models from Dutch Legislative Texts


For more information feel free to reach out to Roderick Lucas ( and Natalia Sidorova (

Leave a Reply