This work investigates how the choice of computational basis impacts neural quantum states performance in quantum systems, suggesting implications for optimal basis selection.
Key Points
This work aims to understand how the performance of neural quantum states is affected by the choice of computational basis.
Examined the dependence of neural quantum states on computational basis using restricted Boltzmann machines.
Analyzed a family of rotated Hamiltonians related to the transverse-field Ising model.
Investigated ground state properties such as degeneracies and amplitude uniformity.
Identified basis-dependence of NQS performance linked to ground state degeneracies.
Noted the relationship between convergence properties of multi-spin operators and NQS performance.
Provided a framework for connecting physical properties to computational effectiveness.