Quantum Natural Gradient on single Qubit and H2

PennyLane Help
1

Hello PennyLane help team,

I am currently attempting to reproduce results from Naoki Yamamoto’s 2019 paper on Quantum Natural Gradient, focusing on a single qubit and the H₂ molecule. I would like assistance in identifying any missing or incorrect components in my code, as I am not achieving the expected results. Could you please advise on which parts may need adjustment or any specific parameters I should double-check? Thank you!

Single qubit:

result_single qubit
result_single qubit1064×435 66.8 KB

H2:

H2_cost
H2_cost1062×435 41 KB

H2_theta
H2_theta1061×443 57.2 KB

The code to reproduce the figure can be found here

  1. Naoki Yamamoto. “On the natural gradient for variational quantum eigensolver.” arXiv:1909.05074, 2019.
2

Hi @Dwi_Cahyo_Mariyanto, welcome to the Forum!

Can you please be more specific about what you got and what you were expecting? To me the graphs look very similar to the paper. Also, note that software evolves over time so the graphs obtained in 2019 may differ from the ones obtained today with the same function. Finally, I noticed that you’re using approx='block-diag'. This means that you’re using an approximation which the paper may not use.

I hope this helps!

1 Like