arXiv:2610.13609
Genuine multipartite nonlocality detection in graph states
Owidiusz Makuta, Flavio Baccari, Remigiusz Augusiak
Abstract
Experimental detection of multipartite Bell nonlocality faces many challenges. The increase in the size of the system demands a higher fidelity of state preparation, a higher detection efficiency, and a larger number of correlators that need to be estimated. These issues are even more pronounced if one wishes to detect a more restrictive variant of Bell nonlocality called genuine multipartite nonlocality (GMNL). In this work, we introduce two classes of multipartite Bell inequalities that are capable of detecting GMNL correlations and address different experimental hurdles in GMNL detection. The first class consists of the number of correlators that scales linearly with the number of parties. The second class increases the number of correlators and, in return, provides higher noise robustness of GMNL detection. Both classes of inequalities are constructed to detect GMNL in graph states, and in fact both can be used for self-testing of graph states and the performed measurements.
AI Summary
The paper asks how to detect genuine multipartite nonlocality (GMNL) in graph states with few correlators and two binary measurements per party, while retaining self-testing. GMNL is defined against mixtures of bipartition-local models whose within-group correlations are nonsignalling. The first construction reweights stabilizer-based BASTA inequalities, which cannot themselves certify quantum GMNL. For any connected $n$-vertex graph and a chosen rotated vertex $r$ of degree $d$, it uses $O(n)$ correlators and has exact bounds $\beta_{BL}=2(n-1)$ and $\beta_Q=2\sqrt{1+d^2}+2(n-d-1)$. The graph state attains the quantum bound with degree-dependent measurements. Smaller $d$ improves the violation ratio; local complementation can further optimize it. The ratio nevertheless approaches one as system size increases. A second construction replaces the core by a Svetlichny-type expression when the closed neighbourhood of $r$ is a clique of size $m$. Its bounds are $\beta_{BL}=2^{m-1}(n-m+1)$ and $\beta_Q=2^{m-1}(n-m+\sqrt{2})$, trading exponentially many core correlators for greater white-noise tolerance. Bounds follow from nonsignalling correlator inequalities, path-based extensions, and sum-of-squares decompositions. Both families support graph-state self-testing at maximal violation; numerical robust-self-testing estimates cover GHZ states with three to six parties. Limitations include exponential local-complementation optimization, restricted applicability of the clique construction, and no general robust-self-testing bounds. Clique applicability and the possible benefit of multiple rotated parties remain open for larger graphs.
Central claim may not hold: Appendix C applies its key correlator lemma to incompatible measurement settings for an allowed graph, leaving a substantive gap in the proof of the clique-based family's general bi-local bound. The clique-based example's stated all-+1 saturation strategy yields 4 rather than 8, although another deterministic strategy does attain 8.
Written by gpt-6.1-sol from the paper's full text. Automated: it can be wrong.
- Comments
- 29 pages, 3 figures