GQSP-Based Hamiltonian Simulation for the Hubbard-Holstein Model — Paper Implementation Challenge

[achebiyam]

Paper Implementation Challenge Submission

GitHub Issue: #1504
Author: @achebiyam

Summary

First application of Generalized Quantum Signal Processing (GQSP) to an
electron-phonon coupled system (Hubbard-Holstein model) on the Classiq platform.

What's included

• GQSP simulation verified on 5-qubit Hubbard-Holstein (overlap 0.9999999917)
• Eigenbasis verification on full 8-qubit model (N_max=2, 41 Pauli terms)
• Resource comparison: GQSP vs Suzuki-Trotter (orders 1, 2, 4) with depth and CX counts
• Time-dependent physical observables (polaron dynamics, CDW order)
• VQE comparison showing GQSP's advantage in systematic accuracy
• Parameter regime study across coupling g, Hubbard U, phonon frequency ω
• Documented Classiq LCU synthesis limitation at ≥16 Pauli terms

Primary reference

D. Motlagh and N. Wiebe, "Generalized Quantum Signal Processing",
PRX Quantum 5, 020368 (2024). arXiv:2308.01501

Notes

• The notebook runs end-to-end on Google Colab with pip install "classiq[qsp]"
• The .qmod file can be exported from the Classiq platform circuit link included in the notebook
• The parameter sweep section uses analytical α estimates to avoid long Pauli decomposition runtimes

[review-notebook-app]

Check out this pull request on 

See visual diffs & provide feedback on Jupyter Notebooks.

---

Powered by ReviewNB

[github-actions]

🔥 New notebook just dropped!

@amir-naveh , @TomerGoldfriend — come check out this shiny new addition to our repo.

[achebiyam]

Note: The .qmod export fails due to a serialization issue in lcu_pauli (TypeError: 'CParamScalar' object cannot be interpreted as an integer). This appears to be an SDK bug in write_qmod when used with lcu_pauli. The synthesized circuit is viewable on the Classiq platform here: https://platform.classiq.io/circuit/3B1F4EOqyt9Dqwju9Ik37lJQvc4

Happy to add the .qmod file once this serialization issue is resolved, or if the team can suggest a workaround.

[TomerGoldfriend]

Note: The .qmod export fails due to a serialization issue in lcu_pauli (TypeError: 'CParamScalar' object cannot be interpreted as an integer). This appears to be an SDK bug in write_qmod when used with lcu_pauli. The synthesized circuit is viewable on the Classiq platform here: https://platform.classiq.io/circuit/3B1F4EOqyt9Dqwju9Ik37lJQvc4

Happy to add the .qmod file once this serialization issue is resolved, or if the team can suggest a workaround.

Thank you @achebiyam , I will go over the PR. I will pass this bug to the team. There is no need to include .qmod anymore, so you can delete it and the write_qmod call.

[achebiyam]

Note: The .qmod export fails due to a serialization issue in lcu_pauli (TypeError: 'CParamScalar' object cannot be interpreted as an integer). This appears to be an SDK bug in write_qmod when used with lcu_pauli. The synthesized circuit is viewable on the Classiq platform here: https://platform.classiq.io/circuit/3B1F4EOqyt9Dqwju9Ik37lJQvc4
Happy to add the .qmod file once this serialization issue is resolved, or if the team can suggest a workaround.

Thank you @achebiyam , I will go over the PR. I will pass this bug to the team. There is no need to include .qmod anymore, so you can delete it and the write_qmod call.

Thanks @TomerGoldfriend! The submitted notebook already uses show(qprog_gqsp)
instead of write_qmod, so no changes needed on that front. Happy to address
any other feedback from your review.

[orsa-classiq]

Line #1.    !pip install "classiq[qsp]" keyrings.alt pennylane -q

remove any pennylane use from the notebook

Reply via ReviewNB

[orsa-classiq]

Line #3.    import keyring

remove all keyring related stuff from the notebook

Reply via ReviewNB

[orsa-classiq]

Line #2.    classiq.authenticate()

remove authentication from the notebook. It is a pre assumption for the notebook

Reply via ReviewNB

[achebiyam]

Done. I removed pennylane, keyring, and authenticate. Also removed the VQE section which depended on pennylane. Happy to address any further feedback.

[orsa-classiq]

@achebiyam
I'm sorry, it looks like almost the entire thing is generated by AI. Also the sectioning is weird and the explanations are not enough. If you want us to take a serious look at the notebook it needs a serious work from you. Take a look on other notebooks in the repo \ paper implementation project as an example.

[achebiyam]

I appreciate the feedback. This notebook represents several weeks of original research, including diagnosing a previously unreported platform limitation at exactly 16 Pauli terms and implementing the first GQSP simulation of an electron-phonon system on Classiq. I used AI tooling for polish and presentation, not for the research or implementation itself.

That said, I do understand the style doesn't match the existing notebooks in the repo. I'll rework the markdown and formatting to better align with your conventions and resubmit. Thanks again.

[achebiyam]

Hi @orsa-classiq, @TomerGoldfriend, just checking in on this PR. I addressed all the feedback from late March (removed pennylane/keyring/authenticate, reworked the markdown style and tone to match the existing notebooks in the repo). Happy to make any additional changes if there's anything else you'd like adjusted. Thanks!

[TomerGoldfriend]

Hi @orsa-classiq, @TomerGoldfriend, just checking in on this PR. I addressed all the feedback from late March (removed pennylane/keyring/authenticate, reworked the markdown style and tone to match the existing notebooks in the repo). Happy to make any additional changes if there's anything else you'd like adjusted. Thanks!

OK @achebiyam, I will examine it. Sorry for missing your last commits and thanks for the reminder.