Tutorial: Deuteron binding energy

[jvscursulim]

Summary

A tutorial about how to use Qiskit Nature tools to tackle the specific problem of deuteron binding energy.

Details and comments

This PR is linked with the Issue #559. In this Issue, was proposed the idea of add a new tutorial in Qiskit Nature to explain how to use FermionicOp and others tools to determine the value of the binding energy between a proton and a neutron in the Deuteron nucleus, as was described in https://arxiv.org/pdf/1801.03897.pdf.

[CLAassistant]

All committers have signed the CLA.

[woodsp-ibm]

In regards of the the present CI failures here.

1. One is that all code here should be formatted with black and this includes notebooks. This can be done with 'make black' in you project root that covers the whole repo and will change what is needed.

2. The other, I expected, this is because we run spell checks and often names and other new technical words need adding to the custom dictionary - its the .pylintdict file in the project root which as you will see is a list, in alphabetic order, of all the custom words we have so far. New ones would be added in the right place in lowercase as per the existing ones.

One final note is that all the tutorials should carry a version and copyright at the end produced from a cell like this (you can see this in the existing tutorials).

import qiskit.tools.jupyter

%qiskit_version_table
%qiskit_copyright

[jvscursulim]

About the problems

1. I reformatted the jupyter notebook with black.
2. I included some new technical words in the .pydictlint file. (annihilation, deuteron, deuterium, eta and others)
3. Added a cell in the notebook with the version table and copyright.

I'm uploading the updated files, hope this time everything works.

[woodsp-ibm]

For the Dirac notation isn't it usually bra-ket with a hyphen in there? Which would pass the spell check without the custombraket addition I think since since bra would be a known word and ket is already there as a custom one.

[jvscursulim]

For the Dirac notation isn't it usually bra-ket with a hyphen in there? Which would pass the spell check without the custombraket addition I think since since bra would be a known word and ket is already there as a custom one.

Yeah, that is true! There is a hyphen in there, I just added the word without thinking about the typo, sorry about that. I'll fix this in a new commit.

[coveralls]

Pull Request Test Coverage Report for Build 2845100980

• 0 of 0 changed or added relevant lines in 0 files are covered.
• No unchanged relevant lines lost coverage.
• Overall coverage increased (+0.01%) to 85.305%

---

Totals
Change from base Build 2842472941:	0.01%
Covered Lines:	17572
Relevant Lines:	20599

---

💛 - Coveralls

[mrossinek]

I don't have time to go through the entire tutorial but at least I got started
with this. I left some more or less detailed comments below. Some of them can be
extended towards sections which I did not yet review in detail. Particular the
explanations should be proof-read.

I am wondering whether we can find a better way to actually compute the energy.
Here, you are constructing a function which creates and runs the VQE but there
you are not really making any use of Qiskit Nature.
The GroundStateEigensolver would come to mind here but I am aware that it has
the "limitation" of taking a Problem instance as its input to solve(). But
maybe constructing a DummyProblem would be a shorter and more Nature-oriented
way of doing things, than the current implementation of the
compute_binding_energy method.

Just wanted to put this out there for you to consider.

[jvscursulim]

@mrossinek Thanks for your comments and suggestions, I'll work to make these changes as soon as possible.

[mrossinek]

Sorry for the long delay on reviewing your PRs!

We just merged #746 which majorly restructured the layout of our code base.
Effectively, all parts of the code which you are using in this tutorial are now residing under qiskit_nature/second_q. Please update your tutorial accordingly.

I will soon get around to doing a more in-depth review of this tutorial! 👍

[jvscursulim]

Sorry for the long delay on reviewing your PRs!

We just merged #746 which majorly restructured the layout of our code base. Effectively, all parts of the code which you are using in this tutorial are now residing under qiskit_nature/second_q. Please update your tutorial accordingly.

I will soon get around to doing a more in-depth review of this tutorial! 👍

No problem and thanks for letting me know about the changes due #746, I will update the code according to these changes.

I'll wait for more comments and suggestions.

[mrossinek]

Okay I finally managed to do a full review.
There are quite a lot of comments here, so please let me know if I need to clarify anything further.

Overall, I think this is a nice tutorial. There are some aspects which need to be changed but you can see the comments below for more details.

At first, I was wondering about the first person perspective of the writing style. Maybe an impersonal perspective would be better suited to guide a user through this, but you can also leave it as it is now.

If you disagree with my comments (especially the one about the Qiskit Terra aspects), maybe the Qiskit Community Tutorials (https://github.com/qiskit-community/qiskit-community-tutorials) are a better place for this. But let's see how this progresses 👍

[mrossinek]

I think the new changes look very good 👍
I left some replies in the threads above and I cannot seem to find a comment which I thought I had left yesterday, so I'll repeat it here:
I think it would be nice, if instead of having one plot after the other (which makes the tutorial very long and makes it impossible to view plots all at once), they were done as a group of subplots. So e.g. the four SLSQP plots could be arranged in four quadrants. The same could be done for the COBYLA plots.

And one final comment w.r.t. writing: ansatz should basically always go with an article. You frequently omit that. I tried correcting this yesterday but I suggest you read through everything again and make sure those articles are in place 👍

[jvscursulim]

I think the new changes look very good 👍 I left some replies in the threads above and I cannot seem to find a comment which I thought I had left yesterday, so I'll repeat it here: I think it would be nice, if instead of having one plot after the other (which makes the tutorial very long and makes it impossible to view plots all at once), they were done as a group of subplots. So e.g. the four SLSQP plots could be arranged in four quadrants. The same could be done for the COBYLA plots.

And one final comment w.r.t. writing: ansatz should basically always go with an article. You frequently omit that. I tried correcting this yesterday but I suggest you read through everything again and make sure those articles are in place 👍

I changed the plots, now I'm using subplots with nrows = 3 for COBYLA and nrows = 4 for SLSQP plots. I hope that now is better to observe the plots.

[mrossinek]

Thanks for the great work on this! I think this tutorial is now in very good shape! Maybe some phrasing could be improved here and there, but I think this will always be the case to some degree.

So I will tentatively approve this and will give @woodsp-ibm a chance to review this before merging anything 👍

Again, thank you @jvscursulim, for the great and patient work here!

[jvscursulim]

Thanks for the great work on this! I think this tutorial is now in very good shape! Maybe some phrasing could be improved here and there, but I think this will always be the case to some degree.

So I will tentatively approve this and will give @woodsp-ibm a chance to review this before merging anything 👍

Again, thank you @jvscursulim, for the great and patient work here!

Thanks @mrossinek for the comments and coding advice, they definitely contribute to improving my coding skills and my writing in english, once it is not my native language.

[woodsp-ibm]

Hi Victor, looks great. Thanks for continuing with this. I have one minor comment to the text. Can we change the intro a little - talking about noisy quantum computers is fine but we prefer to refrain from using NISQ to describe things. So maybe the following - and feel free to do something different if it does not work for you. I tried to preserve what I took from the meaning of what you had. The first part below is changed text, the part after continues with your original text as the latter part of the paragraph (i.e. the text starting with However, there are some problems... . I kept it all here so it can be seen/read at once but still distinguish what I changed

---

At the current stage of quantum computing the quantum processors we have available do not yet have error correcting codes implemented and thus still have noise that must be contended with. This noise can adversely impact the results and hence can limit the sort of problems we can currently tackle when, due to the noise, a high level of confidence and precision cannot
be expected from the results.

However, there are some problems which demand only few qubits to be solved, like the computation of the binding energy between the proton and neutron in the Deuteron nucleus [1]. This problem is an example of a scenario where we can achieve good results despite the challenges mentioned earlier.

[jvscursulim]

Hi Victor, looks great. Thanks for continuing with this. I have one minor comment to the text. Can we change the intro a little - talking about noisy quantum computers is fine but we prefer to refrain from using NISQ to describe things. So maybe the following - and feel free to do something different if it does not work for you. I tried to preserve what I took from the meaning of what you had. The first part below is changed text, the part after continues with your original text as the latter part of the paragraph (i.e. the text starting with However, there are some problems... . I kept it all here so it can be seen/read at once but still distinguish what I changed

At the current stage of quantum computing the quantum processors we have available do not yet have error correcting codes implemented and thus still have noise that must be contended with. This noise can adversely impact the results and hence can limit the sort of problems we can currently tackle when, due to the noise, a high level of confidence and precision cannot be expected from the results.

However, there are some problems which demand only few qubits to be solved, like the computation of the binding energy between the proton and neutron in the Deuteron nucleus [1]. This problem is an example of a scenario where we can achieve good results despite the challenges mentioned earlier.

Hi Steve, thanks for the comment. I read your suggestion and I agree with the changes. I'll update the tutorial text.

[woodsp-ibm]

Victor, thanks for your perseverance and work in contributing this notebook to Nature. I am very happy to be able to approve it - congrats on your contribution to Qiskit!

[jvscursulim]

Victor, thanks for your perseverance and work in contributing this notebook to Nature. I am very happy to be able to approve it - congrats on your contribution to Qiskit!

Thanks Steve, I'm happy with the acceptance of this work as a contribution to Qiskit. I hope that I can make more contributions to Qiskit in the future.