Skip to content

VALIDATION.md

Latest commit

 

History

History
123 lines (92 loc) · 6.5 KB

VALIDATION.md

File metadata and controls

123 lines (92 loc) · 6.5 KB

Validation

Test Suite

679 unit, integration, property-based, and regression tests across ~80 test files. All pass on Python 3.9–3.12 with Qiskit 1.0+. 100% line coverage.

pytest tests/ -v

Test Categories

Unit Tests (~540 tests, ~70 files)

Cover individual modules: Hamiltonian construction, Trotter evolution, VQE, QAOA, QSNN neurons/synapses, crypto protocols, QEC decoder, error mitigation (ZNE + PEC), trapped-ion backend, ITER disruption classifier, quantum advantage benchmark, SNN adapter, SSGF adapter, identity binding spec, QSNN training, fault-tolerant UPDE. Each test runs in <1s on statevector simulator.

Integration Tests (21 tests, 4 files)

End-to-end pipeline validation from K_nm coupling matrix to quantum observables:

File Tests What It Validates
test_integration.py 5 Quantum Trotter vs classical exact evolution (N=2,3,4,6), ZNE on noiseless circuit, energy conservation under Trotter
test_integration_pipeline.py 4 Full pipeline: K_nm → Hamiltonian → VQE ground state (4q), K_nm → Trotter → energy (4q), 8q spectrum properties, 16q Hamiltonian construction
test_cross_module.py 5 Solver ↔ bridge Hamiltonian identity, classical_exact_diag vs numpy.eigvalsh, classical R ∈ [0,1], Z-parity conservation
test_regression_baselines.py 7 K_nm calibration anchors (Paper 27 Table 2 ±0.001), ω values, 4q ground energy E₀ = -6.303 ± 0.01, R range guards

Property-Based Tests (12 tests, 3 files)

Hypothesis-driven fuzzing of invariants:

File Tests Properties
test_bridge_properties.py 5 K_nm symmetry/positivity/diagonal, Hamiltonian Hermiticity (2-6 qubits), probability ↔ angle roundtrip
test_crypto_properties.py 4 CHSH S-parameter bound, key generation roundtrip, QKD sifting preserves key length
test_qec_properties.py 3 Syndrome length, decoder output shape, correction preserves code space

Identity Continuity Tests (43 tests, 4 files)

File Tests What It Validates
test_identity_ground_state.py 11 VQE attractor basin, robustness gap, binding spec input
test_identity_coherence_budget.py 15 Fidelity monotonicity, budget bounds, hardware param propagation
test_identity_entanglement.py 13 Bell state CHSH violation (S≈2√2), product state respects bound
test_identity_key.py 9 Spectral fingerprint, challenge-response verification

v1.0 Module Tests (74 tests, 9 files)

File Tests What It Validates
test_pec.py 9 PEC quasi-probability coefficients, Monte Carlo sampling, overhead scaling
test_trapped_ion.py 8 MS gate noise model, transpilation to {cx,ry,rz,sx,x}, unitarity preservation
test_q_disruption_iter.py 10 ITER 11-feature normalization, synthetic data generation, classifier benchmark
test_quantum_advantage.py 8 Classical vs quantum timing, crossover extrapolation, memory guard at n>14
test_snn_adapter.py 8 Spike-to-rotation conversion, measurement-to-current, bridge forward pass
test_ssgf_adapter.py 8 W→Hamiltonian, phase encoding/recovery roundtrip, state extraction
test_binding_spec.py 7 6-layer 18-oscillator topology, K/omega compilation, VQE attractor
test_qsnn_training.py 8 Parameter-shift gradient, epoch training, loss decrease
test_fault_tolerant.py 8 Repetition code encoding, transversal RZZ, syndrome extraction, qubit count

Cross-Repo Wiring Tests (17 tests, 1 file)

File Tests What It Validates
test_cross_repo_wiring.py 17 ArcaneNeuronBridge (6, skip without sc-neurocore), SSGFQuantumLoop (4), orchestrator mapping roundtrip (4), fusion-core shot adapter (3)

Hardware Smoke Tests (34 tests, 3 files)

All 20 experiment circuits validated on AerSimulator (no IBM credentials needed).

Physics Verification Gates

1. Quantum-Classical Parity

Module Classical Reference Parity Check
qlif.py Bernoulli(sin²(θ/2)) Spike rate within 2σ
xy_kuramoto.py classical_kuramoto_reference() R(t) within 5% for K >> Δω
trotter_upde.py classical_exact_evolution() Per-layer phase tracks classical at n={2,3,4,6}
phase_vqe.py classical_exact_diag() Ground energy within 0.1% (simulator)
qaoa_mpc.py classical_brute_mpc() Optimal action for small horizons
pec.py Analytical quasi-prob coefficients q_I + 3·q_XYZ = 1, overhead = γ^n_gates
quantum_advantage.py Classical expm timing Exponential fit crossover at n>>14
fault_tolerant.py Distance-d repetition code Syndrome detects injected bit-flip

2. Numerical Invariants

Invariant Where Checked
Hamiltonian Hermiticity test_knm_hamiltonian.py, test_bridge_properties.py (hypothesis)
K_nm symmetry, positivity, diagonal test_bridge_properties.py (hypothesis)
K_nm calibration anchors (Paper 27 Table 2) test_regression_baselines.py
4q ground energy E₀ = -6.303 ± 0.01 test_regression_baselines.py
Order parameter R ∈ [0, 1] test_xy_kuramoto.py
Synapse weight bounds [w_min, w_max] test_qsynapse.py
Angle-probability roundtrip test_sc_to_quantum.py
Energy conservation under Trotter test_integration.py
Trotter order-2 convergence test_trotter_error.py
Z-parity conservation test_cross_module.py
Bell state CHSH S > 2 test_identity_entanglement.py
Product state CHSH S ≤ 2 test_identity_entanglement.py
Fidelity monotonically decreasing with depth test_identity_coherence_budget.py
PEC overhead = (Σ q_k
Repetition code: d qubits encode 1 logical test_fault_tolerant.py
Orchestrator phase roundtrip (mod 2π) test_cross_repo_wiring.py
Fusion-core features normalized to [0,1] test_cross_repo_wiring.py

3. Hardware Validation (ibm_fez)

12-point decoherence curve (depth 5 to 770):

  • Readout noise floor: 0.1% at depth 5
  • Linear decoherence regime: depth 85–400
  • Coherence wall: depth 250–400
  • VQE hardware result: 0.05% error on 4-qubit subsystem

Coverage

100% line coverage locally. CI enforces --cov-fail-under=95 (excludes hardware runner/experiments which require IBM credentials).

pytest tests/ --cov=scpn_quantum_control --cov-report=term-missing