Fix PVQD loss function

[gentinettagian]

Summary

The loss function in PVQD is defined as 1 - fidelity, in the current implementation it was 1 - |fidelity|².

[qiskit-bot]

Thank you for opening a new pull request.

Before your PR can be merged it will first need to pass continuous integration tests and be reviewed. Sometimes the review process can be slow, so please be patient.

While you're waiting, please feel free to review other open PRs. While only a subset of people are authorized to approve pull requests for merging, everyone is encouraged to review open pull requests. Doing reviews helps reduce the burden on the core team and helps make the project's code better for everyone.

One or more of the the following people are requested to review this:

• @Qiskit/terra-core
• @manoelmarques
• @woodsp-ibm

[coveralls]

Pull Request Test Coverage Report for Build 3912389631

• 2 of 2 (100.0%) changed or added relevant lines in 1 file are covered.
• 3 unchanged lines in 1 file lost coverage.
• Overall coverage decreased (-0.03%) to 84.818%

---

Files with Coverage Reduction	New Missed Lines	%
src/vf2_layout.rs	3	94.74%

Totals
Change from base Build 3908401991:	-0.03%
Covered Lines:	65562
Relevant Lines:	77297

---

💛 - Coveralls

[ElePT]

Hi @gentinettagian!!! Thanks for catching this!! I was surprised that the unit tests did not find the bug before, but now I see that the only unit test that actually calls the loss function is with fidelity = 1. Given that you are already working with the algorithm, do you have any simple use case we could add as a relevant unit test here?

[Cryoris]

It's probably tricky to find a test that covers this, since the minimum is the same whether we square or not: if the optimizer used is good enough it would work with both this and the previous version. One option would be to test get_loss directly, but since there is some freedom in how the loss function is defined I think it would even be ok without a dedicated test 🙂

[ElePT]

It's probably tricky to find a test that covers this, since the minimum is the same whether we square or not: if the optimizer used is good enough it would work with both this and the previous version. One option would be to test get_loss directly, but since there is some freedom in how the loss function is defined I think it would even be ok without a dedicated test 🙂

That would simplify things for sure :)