Changing the Basis


I am attempting to use a VQE algorithm to find the ground state of a deuterium nucleus by applying a constructed hamiltonian to an ansatz state with one parameter created by a circuit. While I am getting a different expectation value for different parameters, it does not line up with what I am supposed to get (from the paper I am attempting to replicate). I think the basis is wrong. How should I change the basis to fit the hamiltonian I am using? Should I change the basis of the hamiltonian (which I got from tensor products of pauli operators)? Or should I change the basis of the ansatz itself to fit that of the hamiltonian?

PS: this is the paper whose results I am trying to replicate (The H2 graph to be specific). They are using ↑ and ↓ as their basis while I am still on qiskit's default 0 and 1

Tareq Hamarneh

Posted 2020-08-25T20:35:36.073

Reputation: 11

Hello, I think you should try to change the basis at the beginning of your circuit, and check if you have the values you are supposed to get :) – Lena – 2020-08-26T08:06:35.107

But how do I change the basis? That is my question. – Tareq Hamarneh – 2020-08-26T15:17:24.470

1can you add context? What is the Hamiltonian you mention? What basis are you saying is wrong? Please note that questions should be self-contained whenever possible, i.e. not require to read a paper to understand the question – glS – 2020-08-26T23:53:38.017

No answers