Quantum Federated Learning Based on Bures--Uhlmann Geometry for Heterogeneous Noisy Clients

The paper proposes a quantum federated learning framework that extends parameter-space geometry to mixed states, using the Bures metric as a local preconditioner and the mean Uhlmann curvature to create an aggregation rule that down‑weights unreliable clients. It provides theoretical convergence guarantees and demonstrates through trapped‑ion quantum emulator experiments that the method retains high accuracy under device heterogeneity and outperforms standard federated averaging, which suffers under strong noise.

Machine-generated by The Flow from the publisher's headline and feed description — not written or checked by a human. The full article lives at arXiv Machine Learning.