0
$\begingroup$

I'm playing around with the IBM Q to demonstrate the qunatum Zeno effect.

If we prepare a qubit in the $|0\rangle$ state and apply 5 consecutive $R_y(\pi/5)$ gates, we should end up in state $|1\rangle $ with 100% probability since $R_y(\pi/5)^5=R_y(\pi)$.

enter image description here

This gives output 00100, as expected.

If we now add couplings between each of these gates and measure them (using deferred measurement), then after the first $\pi/4$ rotation we collapse the q[2] qubit into either $|0\rangle $ with probability $\cos^2(\pi/8)=0.905$ or into $|1\rangle $ with probability $\sin^2(\pi/8)=0.095$.

What happens after all 5 have been applied, with measurements after each rotation?

My intuition tels me to create a probability tree such as

enter image description here

which gives us a total probability of 67.4% chance of ending up with a |0> state.

The IBM Q simulator gives a 66.3% chance of measuring $|0\rangle $. Is this just statistical error or is there something wrong with my circuit?

My circuit is available at https://quantumexperience.ng.bluemix.net/share/code/5c3890a72f408b005a0d9f06

enter image description here

$\endgroup$
  • $\begingroup$ What are your thoughts? $\endgroup$ – Norbert Schuch Jan 14 at 10:51
  • $\begingroup$ @NorbertSchuch I think it is probably just statistical error, but it just depends whether I've made an error anywhere! $\endgroup$ – James Jan 14 at 12:29
  • $\begingroup$ Did you run the circuit on a simulator which does not make statistical errors? $\endgroup$ – Norbert Schuch Jan 14 at 15:11
  • $\begingroup$ @NorbertSchuch as far as I'm aware there doesn't seem to be away to turn off the randomness on the simulator $\endgroup$ – James Jan 19 at 15:57
  • 1
    $\begingroup$ Then why don't you answer your own question? Might help future users with the same question! (And you'll get my upvote :) ) $\endgroup$ – Norbert Schuch Jan 19 at 19:27
1
$\begingroup$

If you want to know whether the deviation of the output of your circuit from the analytically derived value is a statistical error, you can use a circuit simulator which simulates the exact noise-free circuit, such as quirk.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.