Quantum advantage for distributed computing without communication

Understanding the role that quantum entanglement plays as a resource in various information processing tasks is one of the crucial goals of quantum information theory. Here we propose a new perspective for studying quantum entanglement: distributed computation of functions without communication between nodes. To formalize this approach, we propose identity games. Surprisingly, despite of no-signaling, we obtain that non-local quantum strategies beat classical ones in terms of winning probability for identity games originating from certain bipartite and multipartite functions. Moreover we show that, for majority of functions, access to general non-signaling resources boosts success probability two times in comparison to classical ones, for number of outputs large enough.