
Researchers from Sorbonne Université (CNRS Laboratoire de Chimie Théorique) and Qubit Pharmaceuticals have introduced VULCAN (Variational Upbuilding, Layered Compression, Adaptive N-body perturbation), a hybrid quantum-classical electronic-structure framework designed to eliminate the Hamiltonian gradient-measurement bottleneck of adaptive variational quantum eigensolvers (ADAPT-VQE). Detailed in a preprint on arXiv, VULCAN combines dynamic Lie-algebraic operator pool generation, state-overlap circuit compression, active-subspace projection, and GPU-accelerated second-order perturbation theory to reach chemical accuracy with substantially lower quantum resource demands.
| VULCAN Pipeline Architecture & Benchmark Performance | ||
|---|---|---|
| Pipeline Stage / Metric | Technical Mechanism | Key Benchmark Results |
| Gradient-Free Layered Growth | Generates pool via Variational Double Bracket Flow (VDBF) commutator + 2N − 4 MCP; appends complete layers without energy-gradient screening. | Pools 4×–25× smaller than QEB (e.g., 1,480 vs. 36,456 on 32-qubit C3H6); reaches chemical accuracy at 5%–20% of batched ADAPT-VQE measurements. |
| Overlap-Guided Compression | Prunes circuit depth by ranking operators via single-measurement state-overlap gradients against the uncompressed layer. | Cuts CNOT counts below batched ADAPT-VQE on 4 of 7 benchmarks, reaching 0.14× baseline CNOTs on stretched H10. |
| Adaptive Subspace Projection | Projects Hamiltonian onto 2k dominant determinants guided by a perturbative stopping criterion. | Reduces required qubits by 23%–50% (e.g., 26 → 20 qubits on H2O) and Pauli terms by 60%–88%, with 40%–74% fewer CNOTs after basis optimization. |
| Stochastic ENPT2 Post-Processing | Importance-sampled second-order Epstein–Nesbet perturbation theory on classical GPUs enables early VQE termination. | Reaches chemical accuracy in 1 layer on 26-qubit H2O (0.05% baseline measurements) and 32-qubit C3H6 (1.0 × 10−3 Ha error), where batched ADAPT-VQE fails to converge. |
Rather than screening thousands of candidate operators via energy gradients at every iteration, VULCAN derives physically relevant excitations in seconds from a one-body Variational Double Bracket Flow (VDBF) commutator with the molecular Hamiltonian, completes them into a linearly scaling Minimal Complete Pool, and appends full layers simultaneously. To control circuit depth and register size, users can apply single-measurement state-overlap compression or project the Hamiltonian onto a compact dominant-determinant subspace before running quantum optimization.
In GPU-accelerated statevector benchmarks up to 32 qubits, VULCAN achieved chemical accuracy at 5% to 20% of the measurement cost of batched ADAPT-VQE across hydrogen chains, BeH2, and H2O, while pairing a single quantum layer with stochastic Epstein-Nesbet perturbation theory achieved chemical accuracy on 32-qubit C3H6 where batched ADAPT-VQE stalled. The authors also reported an initial proof-of-principle hardware validation on a Quantinuum System Model H2 trapped-ion processor for molecular hydrogen, measuring a ground-state energy within 1.7 × 10−3 Ha of the exact reference.
Read the full preprint on arXiv here.
In GQI Portal
The players behind the news
The team that writes QCR tracks every company, deal and technology in the GQI Factory, GQI's verified database of the quantum industry. Next up: every story linked to its players, coming to QCR's paid plans.
- Players Companies and institutions across the quantum industry, by segment.
- Scorecards How the players compare on hardware, software, funding and more.
Newsletter
QCR Alerts in your inbox
The latest reporting and analysis from Quantum Computing Report. Free, unsubscribe any time.
