The Magnet project aims to design new machine learning based methods geared towards mining information networks. Information networks are large collections of interconnected data and documents like citation networks and blog networks among others. For this, we will define new structured prediction methods for (networks of) texts based on machine learning algorithms in graphs. Such algorithms include node classification, link prediction, clustering and probabilistic modeling of graphs. Envisioned applications include browsing, monitoring and recommender systems, and more broadly information extraction in information networks. Application domains cover social networks for cultural data and e-commerce, and biomedical informatics.
Specifically, our main objectives are:
- Learning graphs, that is graph construction, completion and representation from data and from networks (of texts)
- Learning with graphs, that is the development of innovative techniques for link and structure prediction at various levels of (text) representation.
Each item will also be studied in contexts where little (if any) supervision is available. Therefore, semi-supervised and unsupervised learning will be considered throughout the project.
International and industrial relations
- UCLA, UCL, Aalto University
- Clic and Walk, Orange, Keycoopt