arXiv:2609.31400
Counterdiabatic quantum optimization for efficient state preparation in the Schwinger model
Ethan Laval, Bipasha Chakraborty, Stefano Cipolla
Abstract
The Quantum Approximate Optimization Algorithm (QAOA), whose design is underpinned by the adiabatic theorem, is one of the leading variational quantum algorithms for preparing the ground states of gauge theories. A recently proposed variant, DC-QAOA, incorporates counterdiabatic driving to accelerate the adiabatic process, reducing both circuit depth and runtime. In this work, we present a systematic analysis, from both theoretical and computational perspectives, of DC-QAOA and related counterdiabatic variants for preparing the ground state of the Schwinger model, the paradigmatic (1+1)-dimensional lattice gauge theory. Benchmarking these methods against standard QAOA, we show that counterdiabatic driving improves ground-state preparation while lowering circuit depth requirements. Our results establish counterdiabatic protocols as a practical route towards efficient quantum simulation of gauge theories.
AI Summary
The paper examines whether counterdiabatic (CD) variants of QAOA improve ground-state preparation for the lattice Schwinger model. It compares standard QAOA, DC-QAOA, and three reduced ansätze retaining CD evolution alongside both, one, or neither of the problem and mixer evolutions. It introduces the problem-plus-CD variant, “CD-prob,” and a layerwise operator-selection procedure that tests a truncated nested-commutator pool and fixes the best-performing CD generator at each layer. Benchmarks use exact-expectation-value classical simulation on $4,6,8,10$ qubits and $1$–$5$ layers, at $m=0.35$, $a=5/6$, $g=1$, and $\theta=0$. Optimization uses gradient descent for 300 iterations, with 20 random initializations per candidate operator and layerwise parameter warm starts. Performance is evaluated through energy ratios and fidelities against exact diagonalization. With an XY mixer, CD-only evolution (“CD-inspired”) consistently gives the strongest average performance; with an X mixer, CD-inspired and CD-mixer perform best. Initializing DC-QAOA from optimized QAOA parameters reverses its otherwise poorer average performance, indicating initialization sensitivity. Three-layer circuits transpiled to an IBM-derived backend generally have lower depth for CD-inspired; at ten qubits with an X mixer, its depth is 246 versus 468 for QAOA. The stated limitations include additional optimization runs for CD-operator selection, fidelity deterioration relative to energy ratios as system size increases, and backend-dependent resource estimates. Results cover one model parameter point and small noiseless simulations; resource counts exclude repeated measurements and the full optimization workload.
Central claim may not hold: The reported advantage over QAOA depends strongly on the counterdiabatic variant, mixer, and initialization, rather than establishing a general DC-QAOA advantage. DC-QAOA has greater transpiled depth than QAOA in every three-layer comparison, and a lower-depth advantage at matched preparation accuracy is not established for DC-QAOA itself. The resource analysis excludes optimization and measurement costs, although CD-term selection uses five or eight times as many training runs per added layer as QAOA. Exact-expectation simulations of 4–10 qubits at one Schwinger-model parameter point do not establish practical efficiency or generalization to broader gauge-theory simulation. The reported CD averages reuse the runs that selected the best operator, introducing selection optimism and preventing an independent assessment of average performance and reliability.
Written by gpt-6.1-sol from the paper's full text. Automated: it can be wrong.
- Comments
- 19 pages, 9 figures, 2 tables. The code used in the paper can be found at https://github.com/EthanLaval/Counterdiabatic-quantum-optimization-for-efficient-state-preparation-in-the-Schwinger-model