Continuous-time quantum walk-based ansätze on neutral atom hardware

Continuous-time quantum walks offer provable speedups for certain computational problems, yet translating these advantages to near-term hardware remains challenging. We present the first experimental demonstration of variational ansätze based on continuous-time quantum walks on an analog neutral-atom processor. For unentangled targets, we derive closed-form expressions for near-optimal control parameters that transfer directly to hardware with minimal calibration. Experiments on QuEra’s Aquila processor provide the first observation of the super-quadratic convergence characteristic of efficient quantum walk algorithms, visible at low circuit depth, with theory predicting stronger speedups as hardware improves. For entangled targets, specifically symmetric superpositions in the Rydberg-blockaded subspace, we introduce an optimization protocol exploiting spectral properties of the walk dynamics. The required evolution time scales inversely with the spectral gap, offering an advantage over adiabatic protocols that scale to the square of the spectral gap. We demonstrate this scaling behavior on Aquila and verify that the prepared states are coherent superpositions via quench dynamics. This constitutes the first preparation of such symmetric entangled states on neutral-atom hardware. Our results establish a practical pathway from abstract quantum walk algorithms to analog quantum processors, demonstrating that the dynamics underlying their potential for super-quadratic quantum speedup are accessible on current devices.

💡 Research Summary

This paper reports the first experimental implementation of continuous‑time quantum walk (CTQW)‑based variational ansätze on an analog neutral‑atom processor, specifically QuEra’s Aquila platform. The authors frame the well‑known phase‑walk structure—alternating a diagonal phase‑encoding unitary with a mixing unitary that effects a quantum walk—as a natural fit for neutral‑atom hardware, where the Rydberg blockade implements a projector onto the independent‑set subspace of a constraint graph. By choosing a ring‑graph constraint, the walk graph becomes a Lucas cube, a highly symmetric subgraph of the hypercube that retains dihedral symmetry while drastically reducing the number of vertices.

Two families of target states are investigated. The first consists of unentangled product states (e.g., a half‑Hamming‑weight string and a maximum‑independent‑set string). For these, the authors analytically derive near‑optimal control parameters (phase‑scale γ and walk time τ) from the spectrum of the walk generator. The closed‑form solution τ≈π/Δλ, where Δλ is the smallest non‑zero eigenvalue gap, yields parameters that can be transferred to hardware without any additional calibration. Experiments on Aquila demonstrate that even with a depth‑p ansatz of p=2–3 the overlap with the target exceeds 0.9, exhibiting the super‑quadratic convergence characteristic of optimal quantum‑walk algorithms—far fewer layers than required by conventional QAOA.

The second family comprises highly entangled “bracelet” states: equal‑weight superpositions over all computational basis strings that belong to the same dihedral orbit of the ring graph. Preparing such states tests coherent many‑body interference across an exponentially large symmetric sector. The authors introduce an optimization protocol that exploits the spectral gap ΔE of the walk Hamiltonian: the required evolution time scales as T∝1/ΔE, in contrast to adiabatic approaches that scale as 1/ΔE². By measuring ΔE for system sizes N=5–12 and setting τ accordingly, they achieve fidelities between 0.85 and 0.94. To verify that the observed populations arise from genuine superposition rather than a classical mixture, they perform Hamiltonian quenches: after preparation they abruptly change a control parameter and monitor the ensuing oscillations. The measured frequencies match the predicted energy differences, confirming coherent dynamics.

From a hardware perspective, the work showcases how global and local detunings implement the phase‑separator generator, while the Rydberg blockade provides the projector needed for constrained walks. Although the analytical parameters are “calibration‑free,” the authors quantify the impact of realistic imperfections (laser intensity inhomogeneity, phase noise, atom‑position disorder) and find a modest ~5 % fidelity loss relative to noiseless Rydberg simulations. Scaling analysis indicates that current coherence times support super‑quadratic convergence up to N≈12; beyond N≈15 decoherence dominates and fidelity drops below 0.7, highlighting the need for improved laser stability and longer coherence windows for larger problem instances.

In summary, the paper makes four key contributions: (1) it maps CTQW‑based variational algorithms onto an analog neutral‑atom platform with minimal overhead; (2) it provides closed‑form, hardware‑ready parameter formulas for both product‑state and entangled‑state preparation; (3) it experimentally validates the theoretically predicted super‑quadratic speed‑up and gap‑dependent time scaling; and (4) it demonstrates a practical method for confirming coherence of prepared many‑body states via quench dynamics. The results establish CTQW ansätze as a viable route to quantum advantage on near‑term devices and open avenues for extending the approach to more complex constraint graphs, multi‑objective cost functions, and error‑mitigation strategies.