Thierry Nagellen




Ubiquitous big data for knowledge extraction from secured silo data

One of the main obstacles of new knowledge creation is the reluctance of organizations to share data, fear of losing control on potential value being most important that the promise to gain additional value or business. While cloud is a major trend, new distributed architectures might emerge to topple this barrier revealing new technical challenges: supervised or unsupervised learning without accessing all data, detailed semantic description of data, algorithms for knowledge extraction.

First the presentation will introduce the different types of distributed machine learning architecture and second explain how some architecture might be interesting in the area of semantics and knowledge creation.