UM Researchers propose model to automatically retrieve relevant legal provisions

DlS researchers Antoine Louis, Gijs van Dijck and Gerasimos Spanakis from Maastricht University analysed how to automatically identify relevant legal provisions for a legal question.

This project forms part of their research on whether and how machines can help providing basic legal assistance to people. Their publication reveals the potential of statutory article retrieval (SAR), the task of retrieving statute law articles relevant to a legal question. It could improve the work efficiency of legal professionals and provide basic legal assistance to citizens in need at no costs.

While traditional approaches to searching for legal articles disregard the way that legal provisions are connected to each other, their approach uses this interconnectedness to significantly improve the search results compared to existing models. For this, they propose a novel graph-augmented dense statute retriever (G-DSR) model that incorporates the structure of legislation via a graph neural network to improve dense retrieval performance.

You can read the paper here.

Antoine Louis presented their work at the 17th Conference of the European Chapter of the association for Computational Linguistics, held from May 2-6 in Dubrovnik, Croatia.