Optimization of the Breeding Protocol for Preparation of GKP States

Abstract:

Ideal target states in continuous-variable quantum information frequently require infinite energy, which makes them unphysical and impossible to generate exactly. This presentation focuses on identifying finite-energy states that maximally approximate these theoretical resources and developing optimal protocols for their physical preparation.
By generalizing existing methods into a tree structure of mixing steps with Gaussian Boson Sampling at the leaves, the proposed protocol reduces optimization complexity and allows for exploration of larger physically relevant systems. The core of the work involves a massive GPU-accelerated multi-objective optimization process to simultaneously balance the quality of the state approximation against the probability of successful preparation and overall circuit complexity.