Phase 1 Operational Task Checklist¶

This checklist aggregates the outstanding Phase 1 tasks called out across the roadmap, experiment plans, and validation guides so the execution team can run them without cross-referencing multiple documents. Use it alongside the detailed procedures in experiments/shadows and the runtime runbook when scheduling IBM Quantum jobs.

Shared infrastructure & preparation¶

[x] Establish a reusable readout calibration workflow (circuit templates, manifest storage, reuse cadence) that precedes every hardware run.
[x] Draft the runtime budgeting checklist (shot counts, batching, queue timing) to stay within the 10-minute IBM Quantum free-tier window.
[x] Implement shared analysis utilities for shot-saving ratio (SSR), confidence-interval (CI) coverage, and variance tracking so experiments share the same metrics code.
[x] Document how to generate and store high-statistics reference datasets (simulators or large-shot baselines) whenever analytical ground truth is unavailable.

Readout calibration cadence and artifacts¶

Invoke quartumse calibrate-readout --backend <provider:name> --qubit <i> ... before each hardware session to refresh confusion matrices. The CLI will reuse an existing archive unless --force is set or --max-age-hours expires.
Calibrations live under validation_data/calibrations/<backend>/q<indices>/confusion_matrix.npz with a sibling .manifest.json capturing metadata (backend_descriptor, shots_per_state, reuse flag, etc.). The manager in experiments/shadows/common_utils.py uses the same layout so GHZ, Bell, Clifford, Ising, and H₂ scripts automatically share matrices.
Surface the cached path in manifests via MitigationConfig.confusion_matrix_path; the CLI output and experiment metadata record this location directly for provenance linking.
Recommended cadence: refresh for new backend calibrations, topology changes, or when the cached artifact exceeds its --max-age-hours threshold (default is unlimited reuse). Force a regeneration whenever qubit mappings change or significant readout drift is observed.

Shadows workstream (S)¶

Extended GHZ (S-T01/S-T02 bridge)¶

[ ] Simulate connectivity-aware GHZ(4–5) preparation circuits for target backends.
[ ] Integrate MEM-calibrated measurement routines and demonstrate mitigated fidelity ≥ 0.5.
[ ] Expand observable estimation to full stabilizer + Mermin terms and compare against grouped measurement baselines.
[ ] Run ≥10 hardware trials to study CI coverage/heavy tails; apply robust estimators if needed.
[ ] Archive procedures, raw logs, processed metrics, and discussion notes in the experiment subfolders.

Parallel Bell pairs (S‑BELL)¶

[ ] Build 4-qubit (optionally 6/8-qubit) disjoint Bell-pair circuits and verify pairwise entanglement.
[ ] Measure $ZZ$, $XX$, and CHSH combinations with MEM such that mitigated $S>2$.
[ ] Quantify SSR gains for simultaneous subsystem observables versus per-pair grouped runs.
[ ] File methodologies, datasets, and analyses under the dedicated directories.

Random Clifford benchmarking (S‑CLIFF)¶

[ ] Generate depth-limited random Clifford circuits (≥5 qubits) and capture stabilizer references from simulation.
[ ] Estimate ≥50 Pauli observables with shadows and grouped baselines, reporting MAE distributions and CI coverage.
[ ] Execute direct fidelity estimation (DFE) and compare shot requirements to shadows.
[ ] Store scripts, calibration notes, stats summaries, and interpretation artifacts.

Ising chain (S‑ISING)¶

[ ] Assemble first-order Trotter circuits for the 6-qubit transverse-field Ising model and validate expected observables in simulation.
[ ] Collect hardware data comparing grouped vs. shadow measurements for equal shot budgets (track energy error/variance).
[ ] Extract auxiliary observables (magnetization, correlators, energy variance) from the same shadows datasets.
[ ] Document procedures, execution logs, analyses, and interpretation materials.

H₂ energy (S‑CHEM)¶

[ ] Implement the 4-qubit H₂ ansatz and benchmark ideal expectations for validation.
[ ] Run shadow+MEM versus grouped-measurement comparisons, targeting 0.02–0.05 Ha accuracy and ≥30% uncertainty reduction.
[ ] Monitor and correct estimator bias from locally weighted sampling if uncertainties are exceeded.
[ ] Capture methodology narratives, raw/processed data, and publication-ready discussion notes.

Cross-experiment reporting¶

[ ] Aggregate the high-impact findings (Hamiltonian efficiency gains, entanglement recovery, multi-observable accuracy) into a consolidated Phase 1 report.
[ ] Ensure each experiment’s discussion notes document success criteria, SSR achievements, and limitations for manuscript prep.

Validation gating tasks¶

[ ] Complete the extended IBM hardware validation campaign (SSR ≥ 1.1×; manifests saved under validation_data/).
[ ] Run the hardware validation post-checks: manifests present, calibration snapshot archived, v0 vs v1 MAE comparison, and results summarised in the Phase 1 status log section below.
[ ] Verify validation CI coverage ≥ 80% and document Phase 1 completion once criteria are met.

Cross-workstream starters (C/O/B/M)¶

[ ] Execute C/O/B/M starter experiments on simulator and generate first data drops (manifests + shot data for Phase 1 closure).
[ ] Confirm patent theme shortlist drafting continues ahead of the Phase 2 gate review.

Note: Advanced shadow variants (v2 Fermionic, v3 Adaptive, v4 Robust) are Phase 2+ scope and intentionally excluded from this checklist to keep Phase 1 focused on v0/v1 hardware validation.

How to use this checklist¶

Copy relevant tasks into your sprint tracker, linking back to their detailed procedure files (design docs in experiments/shadows/**/ and the hardware validation design note).
Before each IBM Quantum run, execute quartumse runtime-status --json --backend <backend> and record queue depth/runtime budget in the runbook.
After completing an experiment, attach manifests, calibration data, and summary notebooks to the appropriate results/ and discussion/ folders and check the corresponding box here.
When preparing a public release or milestone summary, update CHANGELOG.md with the scope of work that shipped before tagging.

Update this document whenever a task is completed or a new Phase 1 dependency is identified. Use the log below to capture concise status updates instead of scattering notes across ad-hoc documents.

Phase 1 status log¶

Date	Update
2025-10-22	IBM ibm_torino smoke test complete. GHZ v0/v1 + MEM validated, SSR ≥ 1.2× on simulator, CI coverage 100%.