Session A2: Addressing Challenges in Data Protection: Is Federated Learning the Solution?

Customs administrations often face different constraints when it comes to the sharing of data generated in trade transactions or Customs operations. On the other hand, the sufficient quantity and quality of data are one of the main prerequisites for developing efficient AI-based algorithms in support of Customs operations.

This session will explore whether federated learning, the process of training AI algorithms collaboratively without actually sharing the data on which the algorithms are trained, is a viable solution for developing AI-based algorithms for Customs.