REVIEW 2 cited by
Noncommutative polynomial optimization under symmetry
Not yet reviewed by Pith; the record is open.
This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.
SPECIMEN: schema-true, not a live event
T0 review · schema-true
One-sentence machine reading of the paper's core claim.
pith:XXXXXXXX · record.json · timestamp
Signed reviews
read the original abstract
We present a general framework to exploit the symmetries present in the Navascu{\'e}s-Pironio-Ac{\'i}n semidefinite relaxations that approximate invariant noncommutative polynomial optimization problems. We put equal emphasis on the moment and sum-of-squares dual approaches, and provide a pedagogical and formal introduction to the Navascu{\'e}s-Pironio-Ac{\'i}n technique before working out the impact of symmetries present in the problem. Using our formalism, we compute analytical sum-of-square certificates for various Bell inequalities, and prove a long-standing conjecture about the exact maximal quantum violation of the CGLMP inequalities for dimension 3 and 4. We also apply our technique to the Sliwa inequalities in the Bell scenario with three parties with binary measurements settings/outcomes. Symmetry reduction is key to scale the applications of the NPA relaxation, and our formalism encompasses and generalizes the approaches found in the literature.
Forward citations
Cited by 2 Pith papers
-
Trading symmetry for Hilbert-space dimension in Bell-inequality violation
Some symmetric Bell inequalities can only be maximally violated by asymmetric minimal-dimension quantum strategies, while the symmetric CGLMP family admits symmetric maximizers up to dimension 19.
-
PCPOP.jl: A Julia package for partially commutative polynomial optimization
PCPOP.jl implements partially commutative polynomial optimization with clique-based arithmetic, algebraic/symmetry/Jordan reductions, and benchmarks showing advantages on quantum-information SDPs.
Discussion (0). Continue with ORCID to comment.