quantum-poc

name	quantum-poc
description	Quantum computing POC standards for QAE, QAOA, and hybrid quantum-classical algorithms. Use for quantum circuit design, simulator benchmarking, and honest reporting of quantum results.
chains_with	["performance","quality"]

Quantum POC Skill — Honest Simulation, Clear Caveats

Mandate

Every quantum result MUST be labeled as simulator-only. No claims of quantum advantage without fault-tolerant hardware evidence.

Core Truths to Embed in Every Output

1. Simulator only: All quantum results are from classical CPU simulation (AerSimulator)
2. No quantum advantage: Quadratic speedup is theoretical — requires fault-tolerant hardware
3. Small scale: Current circuits limited to 3 assets, 16 qubits due to simulator constraints
4. Noise-free: Simulated circuits assume perfect gates and zero decoherence

Quantum Primitives (Project-Q)

QAE — Quantum Amplitude Estimation (Primary)

• Purpose: Estimate tail probability P(return ≤ -VaR)
• Speedup: Theoretical quadratic (O(1/ε) vs O(1/ε²) classical)
• Provable: YES — rigorous lower bound (Brassard et al. 2002)
• Hardware needs: Fault-tolerant, ~1,000+ logical qubits
• Code: src/project_q/quantum/models/risk/qae_estimator.py

IQAE — Iterative QAE (Practical)

• Purpose: Same as QAE but with shallower circuits
• Speedup: Near-quadratic (O(log(1/ε)/ε))
• Advantage: Shallow circuits, no controlled-Grover chains
• Code: src/project_q/quantum/models/risk/iqae.py

QAOA — Quantum Approximate Optimization (Exploratory)

• Purpose: Portfolio optimization (QUBO encoding)
• Speedup: NONE proven — heuristic algorithm
• Status: Exploratory POC, 4 assets, not in UI
• Code: outputs/quantum/qaoa_portfolio_results.json

Benchmark Protocol

# Run QAE vs Classical benchmark
python scripts/benchmark_qae_vs_classical.py --synthetic

# Run with real data
python scripts/benchmark_qae_vs_classical.py --assets SPY QQQ BTC-USD

Output Schema

{
  "metadata": {
    "assets": ["SPY", "QQQ", "BTC-USD"],
    "simulator_only": true
  },
  "classical_mc": {
    "var": 20690.48,
    "es": 27171.79,
    "n_samples": 100000
  },
  "qae_fast_backend": {
    "qae_var": 0.03806,
    "runtime_seconds": 0.002,
    "simulator_only": true
  },
  "qae_aer_backend": {
    "qae_var": 0.03806,
    "runtime_seconds": 8.873,
    "circuit_depth": 34,
    "simulator_only": true
  },
  "disclaimer": "All results are from classical QPU simulation..."
}

Reporting Standards

Do Say

• "Theoretical quadratic speedup for QAE"
• "Simulator validates algorithm correctness"
• "Wasserstein distance = 0.0001 — distributions match"
• "Pipeline ready for fault-tolerant hardware"
• "Circuit compiles at 16 qubits, depth 18"

Don't Say

• "Quantum advantage demonstrated"
• "Faster than classical" (runtime comparison is simulator overhead)
• "We ran on quantum hardware"
• "Production-ready quantum risk engine"
• "Beats classical models at [anything]"

Output Files

outputs/quantum/
  summary.json              — QAE circuit metadata
  comparison.json           — Quantum vs classical distribution comparison
  benchmarks.json           — Runtime comparison
  circuit.json              — Circuit diagram data
  qubit_states.json         — Basis state probabilities
  aer_qae_results.json      — AerSimulator QAE results
  benchmark_results.json    — Full benchmark comparison
  qaoa_portfolio_results.json — QAOA POC results

Known Caveats

• 3-asset limit (qubit budget: 6 state qubits = 64 bins)
• AerSimulator at 17 qubits is near practical limit for classical simulation
• QAOA results not hooked to UI
• No entanglement or interference effects beyond QAE