Federation Boosting Tree for Originator Rights Protection

The problem of data island hinders the application big in artificial intelligence model training, so researchers propose a federated learning framework. It enables training without having to centralize all central storage point. In current horizontal scheme, each participant gets final jointly trained model. No solution is proposed for scenarios where participants only provide exchange benefits, but do not care about Therefore, this paper proposes new boosted tree algorithm, called RPBT (the originator Rights Protected Boosted Tree algorithm). Compared with federal will obtain can guarantee that local be leaked, while cannot obtained. worth mentioning that, from perspective participants, scheme uses batch idea make participate random batches. more suitable large number are modeling. Furthermore, small actually joint process. secure. Theoretical analysis and experimental evaluations show secure, accurate efficient.