arXiv · 2611.23514
Quantum resources and the onset of the quantum Monte Carlo sign problem
Abstract
The Monte Carlo sign problem poses a fundamental obstacle to classical simulation of quantum many-body systems, yet its connection to quantum resources remains poorly understood. We probe this connection by systematically tuning the degree of nonstoquasticity in random k-local Hamiltonians. In numerical tensor-network studies, we find the sign complexity and the quantum resources of entanglement and nonstabilizerness of the ground state to be strongly suppressed in a broad regime up to substantial nonstoquasticity. As nonstoquasticity is further increased, the considered quantum resources experience sharp crossovers into a regime where they are strongly enhanced. Surprisingly, numerical quantum Monte Carlo (QMC) simulations drastically deteriorate in the same region, even though the Hamiltonian is already considerably nonstoquastic well before it, indicating that the failure of ground-state QMC is correlated more with the resource content of the state than the sign structure of the Hamiltonian. We corroborate these findings analytically by constructing an approximate ground state of k-local Hamiltonians in the stoquastic limit. Moreover, we identify a distinct behaviour in graph states. Their nonstabilizerness always vanishes, but, as we demonstrate analytically, on bipartite graphs the sign can be easily cured while in multipartite graph states it cannot. Our results thus shed light on the highly subtle interplay between quantum resources and the computational difficulty encountered by QMC methods.
Explore related subjects
Keep this discovery
Explore connections, maps & timelines
Naga Dileep Varikuti, Leela Ganesh Chandra Lakkaraju, Philipp Hauke. 2026-11-17. Quantum resources and the onset of the quantum Monte Carlo sign problem. https://arxiv.science/abs/2611.23514
Citations retain the original authors and DOI and include this site’s record URL. If this discovery service contributed to your research, you can also cite arXiv Science by Indiciae: BibTeX · RIS.