D6.2 - Preliminary conclusions about Federated Learning applied to clinical data

This report comprises the first contributions from different partners on Federated Learning (FL). Aftera preliminary introductory section where the fundamental procedures and limitations are described,we detail the well-known mathematical foundation of Federated Learning for convex problems. In thiscase, we present a key algorithm, Alternating Direction Multipliers Method (ADMM), which is ableto implement in a distributed way some fundamental problems such as regression (Ridge and LASSO)and classification (Logistic Regression and Support Vector Machines (SVM)). This procedure sharesthe fundamental approach of FL, which consists of performing some local processing, sharing someintermediate information and updating the local information with some global innovation. In a secondstep we introduce the extension of this approach to non-convexproblems using Bayesian Neural Networks(BNN) where the update is based on the cooperative construction of the posterior of weightsfrom different architectures. Several sections follow where different partners provide different contributionsdescribing our first initiatives on the topic. Some preliminary code from all partners hasbeen uploaded to a common repository to start creating a pool of methods and tools to foster incomingsynergies.