Protecting Model Updates in Privacy-Preserving Federated Learning: Part Two

The problem The previous post in our series discussed techniques for providing input privacy in PPFL systems where data is horizontally partitioned. This blog will focus on techniques for providing input privacy when data is vertically partitioned . As described in our third post , vertical partitioning is where the training data is divided across parties such that each party holds different columns of the data. In contrast to horizontally partitioned data, training a model on vertically partitioned data is more challenging as it is generally not possible to train separate models on different