Guides pages for OBP

[kaelynj]

Closes #2107 and closes #2108

I'll include the files in the toc and satiate CI once #2148 is merged to avoid dealing with conflicts.

[review-notebook-app]

Check out this pull request on 

See visual diffs & provide feedback on Jupyter Notebooks.

---

Powered by ReviewNB

[caleb-johnson]

Thanks Kaelyn, looks great! I'll just drop a few thoughts below for the getting started guide:

• Paragraph 1

  • For clarity can we change "slices" --> "circuit slices" when we first mention it since this is kind of a made-up term
  • I like that you highlight preparation of the slices and operator budget. I think we need to include the truncation budget here, as it is arguably more important than the operator budget (can explain this in more detail if you're interested). Some of the language in your OBP landing page intro could be moved over here regarding this.

• Prepare inputs to backpropagate

  • qiskit_addons_utils.slice_by_gate_types() --> qiskit_addon_utils.slicing.slice_by_gate_types

[caleb-johnson]

Looks great, just a couple of minor comments/requests

[BryceFuller]

I have a few suggested edits to qiskit-addons-obp.mdx:

This method has the benefit that the overhead incurred by backpropagating various gates is determined by the non-Cliffordness of that gate (i.e. how much of a particular portion of $U_C$ which contains non-Clifford instructions).

I would change to

This method has the benefit that the overhead incurred by backpropagating various gates scales with the non-Cliffordness of $U_C$ (i.e. how much of $U_C$ is comprised of non-Clifford instructions).

Also, later in that section I would append an additional sentence

Each of these slices are then analytically applied in sequence to measure the back propagated operator $O'$ and may or may not contribute to the overall size of the sum, depending on if the slice is a Clifford vs non-Clifford operation.

would become

Each of these slices are then analytically applied in sequence to measure the back propagated operator $O'$ and may or may not contribute to the overall size of the sum, depending on if the slice is a Clifford vs non-Clifford operation. If an error budget is allocated, truncation will then occur between the application of each slice.

Here are some suggested edits to qiskit-addons-obp-get-started.ipynb:

there is a space missing in one of the print messages for the cell that runs the actual OBP method

New observable has 17 terms, which can be combined into7 groups.

should print out

New observable has 17 terms, which can be combined into 7 groups.

Also, the value of t here has been changed from 0.2 in the obp tutorial, to 2. This results in the optimal expval being very close to 0, and the final answers in the notebook are all within statistical accuracy from zero. Can you return t to it's earlier value so we know there is a larger expval.

EDIT: looks like reps also changed. It's ok for these values to differ from the addon documentation, as long as the target expval is not very close to 0.

[abbycross]

copyediting suggestions, otherwise good to go

[jenglick]

@kaelynj -- pending getting Bryce's suggestions folded in, are you also able to also add in the results for running the backprop and backprop w truncation circuits run on the StatevectorEstimator to isolate the OBP-specific effects? Beyond that, this looks good to me!

[kaelynj]

@kaelynj -- pending getting Bryce's suggestions folded in, are you also able to also add in the results for running the backprop and backprop w truncation circuits run on the StatevectorEstimator to isolate the OBP-specific effects? Beyond that, this looks good to me!

This should be included now (along with @BryceFuller's suggestion).

@Eric-Arellano I'm having an issue with CI on the checkPatternsIndex script. NPM claims that the pages are not present in the toc when they actually are. Could this be because the pages are nested one level further than the rest of ours?

[jenglick]

@kaelynj The tutorial for OBP in Learning Platform just dropped -- would you be able to add a link to this in the "next steps" section of the getting started page?

https://learning.quantum.ibm.com/tutorial/improving-estimation-of-expectation-values-with-operator-backpropagation

[kaelynj]

@kaelynj The tutorial for OBP in Learning Platform just dropped -- would you be able to add a link to this in the "next steps" section of the getting started page?

https://learning.quantum.ibm.com/tutorial/improving-estimation-of-expectation-values-with-operator-backpropagation

Ended up adding this link in the AQC pr #2291 to save time for QDC