SciPy 2026 CFP Readiness¶
Date prepared: 2026-05-06
This note prepares a possible SciPy 2026 talk or poster proposal. It does not claim that a CFP is open, that a proposal has been submitted, or that the conference has accepted the work. Before submission, verify the current SciPy CFP page, tracks, deadlines, length limits, and eligibility requirements.
Recommended Angle¶
Primary angle:
Artefact-first quantum software: reproducible Kuramoto-XY NISQ workflows with Python/Qiskit orchestration, Rust hot-path kernels, benchmark regeneration, and raw hardware-count provenance.
Why this fits SciPy:
- Python-first scientific workflow with a narrow Rust acceleration layer.
- Reproducibility discipline: JSON/CSV artefacts generated from scripts, not hand-edited tables.
- Open scientific data lineage: raw counts, job IDs, SHA256 hashes, and claim-boundary documents.
- Practical benchmark story across local CPU, ML350, Vertex CPU, and Vertex T4.
- Honest negative and bounded hardware results, useful as a reproducible scientific-computing case study.
Suggested Formats¶
| Format | Fit | Boundary |
|---|---|---|
| Talk | Strong if the track welcomes scientific Python infrastructure or reproducibility workflows. | Emphasise software engineering, artefacts, and reproducibility rather than speculative physics. |
| Poster | Strong fallback if a full talk is not suitable. | Use figures/tables from generated artefacts only. |
| Birds-of-a-feather | Possible if focused on reproducible NISQ/scientific benchmark packaging. | Avoid making it a project advertisement. |
Draft Title Options¶
Artefact-first quantum workflows in Python: reproducible Kuramoto-XY NISQ studies with Rust hot pathsFrom oscillator networks to auditable hardware counts: a Python/Rust workflow for reproducible NISQ experimentsMaking small-N quantum experiments reproducible: generated benchmarks, raw-count provenance, and bounded claims
Draft Abstract¶
scpn-quantum-control is a Python/Rust research-software package for
reproducible Kuramoto-XY quantum-control workflows. It maps oscillator coupling
matrices and natural frequencies to Qiskit Hamiltonians, topology-informed
ansatze, simulator workflows, IBM hardware artefact packages, and
benchmark-regeneration scripts. This proposal presents the software-engineering
layer behind the project: an artefact-first rule where numerical tables are
regenerated from committed scripts into JSON/CSV summaries; selected Rust/PyO3
hot-path kernels for deterministic coupling-matrix and workflow support;
cross-machine benchmark provenance; and raw hardware-count packaging with job
identifiers, metadata, SHA256 hashes, and claim boundaries. The talk uses
small-N NISQ studies as a concrete scientific-computing case study while
explicitly avoiding quantum-advantage claims. The goal is to show how Python
scientific workflows can make hardware-adjacent research auditable, rerunnable,
and honest about negative or backend-sensitive results.
Evidence to Cite¶
| Evidence | Path |
|---|---|
| Methods benchmark dashboard | docs/methods_benchmark_dashboard.md |
| Rust/VQE methods checklist | docs/rust_vqe_methods_submission_checklist_2026-05-06.md |
| JOSS/software checklist | docs/joss_software_submission_checklist_2026-05-06.md |
| DLA parity submission checklist | docs/dla_parity_submission_checklist_2026-05-06.md |
| FIM submission checklist | docs/scpn_fim_submission_checklist_2026-05-06.md |
| Coverage gap audit | docs/coverage_gap_audit_2026-05-06.md |
| Behavioural test audit | docs/behavioural_test_audit_closure_2026-05-06.md |
| Source packaging readiness | docs/arxiv_source_packaging_readiness_2026-05-06.md |
Claims to Avoid¶
- Do not claim quantum advantage.
- Do not claim backend-stable DLA parity protection.
- Do not claim FIM hardware coherence protection.
- Do not claim Rust accelerates the whole Qiskit/VQE/transpilation workflow.
- Do not present opportunistic shared-machine timings as isolated microbenchmarks.
- Do not imply that new QPU time is needed for the SciPy proposal.
Pre-Submission Gate¶
Before submitting a proposal:
- verify the current CFP page and track descriptions;
- check word/character limits;
- decide talk vs poster vs BoF;
- confirm whether papers/preprints are public yet;
- use only generated benchmark numbers from committed artefacts;
- ensure no private paths, credentials, or internal coordination files appear;
- update this note with the final submitted title, abstract, and submission URL if submission is completed.
QPU Boundary¶
No QPU time is required for a SciPy proposal. Any hardware claims should cite already committed raw-count artefacts and manuscript claim boundaries only.