In this deliverable, we have presented the initial progress of the distributed AI-driven decision engine (DE),
proposed in Work Package 4. We identified the key axes of novelty of the proposed mechanisms, mainly as
the co-design of: (a) a stack of algorithms that is fully data-driven, and based on state-of-the-art machine
learning methods that are revisited taking into account the intricacies specific to cellular networking setups,
that force these algorithms to become network-aware or network-friendly; (b) a distributed implementation
of such algorithms that is flexible enough to adapt from hierarchical to fully decentralized “federated”
operation.
