Quantum Nonlocality under Latency Constraints
Bell inequalities are bounds on the correlations between different parties obeying a local hidden variable theory. Here, "local" refers to spacetime locality: the parties cannot communicate their inputs because they must produce their outputs faster than the speed-of-light delay between them. In other words, the parties must satisfy a certain latency constraint. In this work, we explicitly incorporate spacetime locality into the formulation of Bell inequalities by imposing such a latency constraint. When the latency constraint is sufficiently tight such that no parties can communicate, this becomes a standard Bell scenario. When the latency constraint is relaxed such that a subset of the parties can communicate, we no longer have a Bell scenario, but we can again find a divide between classical and quantum behaviors. Hence, we observe that the classical-quantum gap should actually be a function of time. To study these more general scenarios, we introduce the mathematical framework of latency-constrained games, which models time-evolving input and output processes for spatially separated parties subject to finite communication speeds. This framework allows us to systematically study the weirdness of quantum mechanics in the "low-latency regime" where the speed-of-light delay is non-negligible. Latency-constrained games can describe real-time decision-making in real-world settings that are latency-sensitive, such as high-frequency trading and distributed systems, and can reveal the utility of quantum correlations in these settings.