This master thesis aims at developping neural networks for text-mining in justice decisions. The desired final outcome is a neural network capable of summarizing justice decision's text into a feature-vector. This neural network will enable us to explore the relationships between trademark logos, with the legal aspect of trademark similarity, and ultimately, to propose novel tools for legal offices for their daily operations (trademark similarity assessment, search engines, ...).

Contact : Thomas Vandamme (thomas.vandamme@ulb.ac.be)

Promoter : Prof. Olivier Debeir