Federated Myopic Community Detection with One-shot Communication
This addresses the challenge of decentralized network analysis for applications like social networks or distributed systems, though it appears incremental as it builds on federated and myopic learning paradigms.
The paper tackles the problem of recovering network community structure under federated myopic learning, where clients observe small subgraphs and send censored evidence to a central server, showing that exact recovery is possible in polynomial time under certain topological and signal-noise conditions.
In this paper, we study the problem of recovering the community structure of a network under federated myopic learning. Under this paradigm, we have several clients, each of them having a myopic view, i.e., observing a small subgraph of the network. Each client sends a censored evidence graph to a central server. We provide an efficient algorithm, which computes a consensus signed weighted graph from clients evidence, and recovers the underlying network structure in the central server. We analyze the topological structure conditions of the network, as well as the signal and noise levels of the clients that allow for recovery of the network structure. Our analysis shows that exact recovery is possible and can be achieved in polynomial time. We also provide information-theoretic limits for the central server to recover the network structure from any single client evidence. Finally, as a byproduct of our analysis, we provide a novel Cheeger-type inequality for general signed weighted graphs.