| name | vqe-active-space-drug-benchmark |
| description | Systematic benchmark methodology for active space selection in VQE-driven quantum drug discovery pipelines. Classifies molecule suitability for quantum computing using chemically grounded metrics, evaluates VQE across UCCSD and HEA ansatze with both simulation and QPU execution on drug-like molecules (lovastatin, oseltamivir, morphine). |
| version | 1.0.0 |
| category | quantum-medical |
| activation_keywords | ["VQE active space","quantum drug discovery","active space benchmark","UCCSD drug molecule","HEA ansatz drug","molecular suitability quantum","VQE pipeline benchmark","lovastatin quantum","oseltamivir quantum"] |
| last_updated | 2026-06-17T00:00:00.000Z |
| source | arxiv:2512.18203 |
VQE Active Space Selection Benchmark for Drug Discovery
Source
Paper: "Benchmarking the Impact of Active Space Selection on the VQE Pipeline for Quantum Drug Discovery"
arXiv: 2512.18203 (December 2025)
Authors: Zhi Yin, Xiaoran Li, Zhupeng Han, Shengyu Zhang, Xin Li, Zhihong Zhang, Runqing Zhang, Anbang Wang, Xiaojin Zhang
Problem Statement
Applying Variational Quantum Eigensolvers (VQE) to realistic drug-like molecules on NISQ hardware remains challenging. Active space selection is a key strategy to leverage current hardware effectively, yet remains under-benchmarked. This work introduces the first systematic benchmark for active space-driven VQE in quantum drug discovery.
Methodology
Molecule Suitability Classification
Heuristic criteria based on chemically grounded metrics to classify whether a molecule is suitable for quantum computing:
- Electronic Structure Complexity: Number of correlated electrons, orbital degeneracy
- Active Space Size: Feasible qubit count for NISQ hardware (typically 10-30 qubits)
- Chemical Significance: Presence of pharmacophores, reactive centers, metal centers
- Classical Intractability: Cases where classical methods (DFT, CCSD(T)) struggle
Benchmark Molecules
- Lovastatin: Statin drug, complex ring system
- Oseltamivir: Antiviral (Tamiflu), amine-containing
- Morphine: Opioid analgesic, complex polycyclic structure
VQE Evaluation Setup
Ansätze Tested:
- UCCSD: Unitary Coupled Cluster with Singles and Doubles (chemistry-accurate but deep circuits)
- HEA: Hardware-Efficient Ansatz (shallower, hardware-friendly but less chemically motivated)
Execution Modes:
- Statevector Simulation: Noiseless baseline
- QPU Execution: Real quantum hardware with noise
Evaluation Metrics:
- Chemistry metrics: Energy error vs FCI/CCSD(T) reference, dissociation curve accuracy
- Architecture metrics: Circuit depth, parameter count, gate count, hardware compatibility
Active Space Selection Strategies
- Chemically Motivated: Domain-expert selection based on molecular orbital analysis
- Automated Heuristic: Algorithmic selection based on orbital energy gaps, occupation numbers
- Comparative Analysis: Systematic comparison across active space sizes
Implementation Steps
- Molecule Selection: Choose drug-like molecules with known pharmacological relevance
- Classical Pre-computation: Run HF/DFT to obtain molecular orbitals
- Active Space Selection: Apply heuristic criteria to select active orbitals
- VQE Setup: Configure UCCSD and HEA ansätze for each active space
- Execution: Run on simulator and QPU
- Evaluation: Compare energy accuracy, circuit complexity, hardware performance
- Benchmark Report: Classify molecules by quantum readiness
Key Insights
UCCSD vs HEA Trade-offs
- UCCSD: Higher accuracy but deeper circuits, more prone to noise on real hardware
- HEA: Shallower circuits, better hardware performance but less chemical accuracy
- Active space size critically affects both: larger spaces need UCCSD for accuracy but may exceed hardware limits
Hardware-Algorithm Co-Design
- Active space selection should consider target hardware constraints (qubit count, connectivity, coherence time)
- Different molecules require different active space strategies based on their electronic structure
Pitfalls
1. Active Space Too Small
Problem: Missing important correlation effects, inaccurate energies.
Solution: Use chemical intuition + automated heuristics to ensure all relevant orbitals are included.
2. Active Space Too Large
Problem: Circuit depth exceeds NISQ capabilities, noise destroys results.
Solution: Benchmark molecule suitability before committing to VQE.
3. HEA Ansatz Chemical Accuracy
Problem: HEA may not capture correct chemical physics even with enough parameters.
Solution: Use UCCSD for accuracy-critical applications, HEA for hardware-feasibility studies.
4. Reference Method Selection
Problem: Need reliable classical reference (FCI, CCSD(T)) to evaluate VQE accuracy.
Solution: Use CCSD(T) for drug-sized molecules where FCI is intractable.
Verification
- VQE energy should converge to classical reference as active space increases
- UCCSD should be more accurate than HEA on simulators
- HEA should be more robust on real QPU hardware
- Active space selection criteria should correctly predict molecule suitability for quantum computing
Related Skills
- vqe-active-space-benchmarking
- quantum-drug-discovery
- quantum-chemistry
- dft-embedded-quantum-chemistry