Meet our interdisciplinary PhD Students - Part 2: Masha Medvedeva

For decades, summaries of court judgements were published in written journals, which were not easily accessible for the public. Nowadays, courts publish their judgements online. For example, through www.rechtspraak.nl (NL) or through http://hudoc.echr.coe.int (European Court of Human Rights). The goal of our project is to automatically identify linguistic patterns in this type of legal big data and use these to predict the outcome of the judgements.

Coming from a computational linguistics background, I wanted to learn something new and see how I can apply all that knowledge in a new area. I got interested in this project, because of how new the topic is and how relevant its application is. Combining computational linguistics and law opens so many doors for statistical analysis in law and automation in legal studies.

Right now I am at working with trials of European Court of Human Rights. After a few weeks of work we can already predict how the court will rule on the case correctly 81% of the time. And I expect that number to get higher soon. To achieve such results, we look at parts of the trial that describe the circumstances of the case, for instance: who is the victim, where they are from, which of their right they believe has been violated. Then we use a machine learning algorithm that counts how many times different words were used in that text and evaluates which words make the most impact and thus learns to predict the decision. Using that information we can predict the results of future trials in the European Court of Human Rights.

Later, we are planning to move on to other international courts and then to the Dutch courts. By using more linguistic information extracted from the texts of the trials we can learn to predict the decisions better as well as detect potential flaws in the legal systems.

I feel that the interdisciplinary nature of the YAG projects is a very unique opportunity to get out of the comfort zone, learn something completely new and find what lies in the intersection of different disciplines. "To Boldly Go Where No Man Has Gone Before…”