Speaker
Description
Can the direction of time and causal structure be inferred from operational quantum principles, rather than assumed? Causal models and tensor networks both formulate a notion of quantum causality without reference to spacetime, in complementary ways. Causal models are operationally well-defined, yet typically assume acyclicity, which induces a global direction. Tensor networks assume no directionality, but their causal content is not operationally well understood. Connecting the two promises progress on both fronts. However a key challenge in the bridge is to consistently describe general quantum causal models beyond the acyclic case, allowing for cyclic scenarios where central causal modelling tools, such as graph-separation theorems breakdown.
This talk will first present a framework for cyclic quantum causal models equipped with a robust probability rule, and show how the graph-separation problem is resolved in this framework. We will then construct two-way mappings between such (possibly cyclic) quantum causal models and tensor networks, giving an operational meaning to causal influence in tensor networks by linking it to the concept of signalling between agents in causal models. Finally, the talk will discuss applications of these results for defining discrete space-time rotations of causal models that preserve signalling relations and some implications for the emergence of a time direction.
Based on joint works with Carla Ferradini, Victor Gitton and Giulia Mazzola (https://arxiv.org/abs/2502.04168, https://arxiv.org/abs/2603.12283).