Speaker
Description
In this talk, we extend the Variational Quantum Eigensolver (VQE) approach to improve the estimate of the ground state of a quantum system by minimizing the expectation value of a target Hamiltonian on a k-frame—a set of k linearly independent orthonormal states—that define a k-dimensional subspace within the full Hilbert space. This search is then supplemented by an exact diagonalization in the optimal subspace. We find that this method significantly improves ground state estimation accuracy and optimization efficiency. We provide theoretical justification for these improvements by investigating the correlation between ground state infidelity and the loss function, as well as analyzing the expressivity of the k-frame formulation in comparison to the standard version of VQE.
