Cloud-based architecture to train models from third-parties

A cloud-based architecture was implemented to train models from third-parties without sharing datasets. This allow to assure that intelectual property and privacy issues are assured, giving flexibility to train and retrain models that are nto being developed under NOS infrastructures. This achievement is now assessed to a potential European patent.

Train models without sharing datasets and source code. The framework supports any programming language.

The code should be delivered according to NOS contracts.

Researchers, Start-ups

Pay per use.