Network Scalability

DeAI leverages the internet infrastructure for facilitating communication among diverse parties. This communication framework is fundamental for Federated Learning (FL) and Multi-party Computation (MPC) within the DeAI paradigm. Both FL and MPC heavily rely on communication protocols that facilitate multiple rounds of interaction among participating nodes.

Efficient network communication is critical for the success of FL and MPC protocols within DeAI settings\cite{li2020federated}. However, the communication overhead associated with transmitting checkpoints and model updates can significantly impact the overall efficiency. The optimized design of computation protocols minimizes the number of communication rounds required between nodes, enhances the efficiency and scalability of DeAI systems.