A Typed Tensor Language for Shared-State Federated Computation

Excerpt

arXiv:2605.21103v2 Announce Type: replace Abstract: Shared-state federated computations combine client-local tensor computation, mergeable aggregation into shared state, and shared-only post-processing. We introduce a typed tensor language for this class of computations. Its two tensor sorts separate client-partitioned data from globally available values, and typing tracks the partitioned axis. A virtual global tensor serves as a semantic reference for centralized evaluation. We show that typed