Structuring legal information at Doctrine: Challenges and solutions - Naver Labs Europe
loader image

Speaker: Nicolas Fiorini – Arnaud Miribel @ Doctrine

doctrine logo imageAbstract: Every year in France, millions of legal documents (decisions, legislations, commentaries) are published and play en eminent role in our country’s evolution. Being quickly aware of what’s relevant among all this data, for example, is critical to lawyers who need up-to-date and precise information to defend their clients in court., a French legal search engine, is an AI-powered solution to fulfil this need. In this talk, we will present what it takes to use this amount of data and build a legal search engine upon it. First, we will outline the major inherent challenges we face in dealing with domain-specific, legal data. Second, we will describe projects where machine learning and deep learning help us address these challenges and extract knowledge from our documents. Finally, we will thoroughly describe one use-case, namely how we detect the structure of decisions on Doctrine (i.e., the table of contents) to enable users navigate through them more easily.

Date: 13th May 2019

Ceci correspond à une petite biographie d'environ 200 caractéres