Quantum Simulation of a Hyperbolically Driven Quantum System with Superconducting Circuits

The gist: This work simulates a hyperbolically driven quantum system using superconducting circuits to investigate how intrinsic geometry dictates observable physical quantities in parameter space.

Construction of a Hyperbolically Driven Quantum System

To simulate quantum dynamics on a hyperbolic manifold, the authors implement a parameter-space driving scheme where the system’s Hamiltonian is modulated along trajectories on a curved geometry As illustrated in Fig. 1(a), the top panel depicts the theoretical framework: an ensemble of classical particles evolving on a hyperbolic parameter space M. Theoretically, a parametrically driven system is defined by the triple (M, Hˆ (x), x(t)). They utilize the Bolza surface Σ2—a closed Riemann surface with genus g = 2 and constant negative curvature—as their parameter manifold M. The Bolza surface is constructed as the quotient of the Poincar´e disk D = [z ∈ C z < 1] under the action of a Fuchsian group. The metric is given by gij = 4δij/(1 − z2) squared. In this article, they synthesize a two-level Hamiltonian Hˆ (z) = omega(z) · Λ, parameterized by the coordinate z on the manifold, where Λ are spin operators. The parameter ϵ controls the topological phase of the system. Experimentally, they set omega=20MHz.

Topological Response and Adiabatic Limits

A key question addressed is whether global topological properties of the quantum system can be recovered by driving the system along specific trajectories. While local driving is generally insufficient, the Hopf–Anosov theorem guarantees that geodesic flow on a closed, negatively curved surface is ergodic. This ergodicity implies that a particle moving along a geodesic will eventually explore the entire phase space, allowing for the extraction of global invariants. A topological response function ω(T) is defined based on the time-averaged expectation value of a geometric operator Oˆλ (t). Under the dual conditions of ergodicity (T → ∞) and adiabaticity (λ → 0), ω(T) converges to the first Chern number C1. They experimentally investigate this relationship using a superconducting qubit. For a slow particle (λ = 1/20), the driving is sufficiently adiabatic, and as the geodesic length increases, ω(T) converges to C1 = 1, as shown in Fig. 2. Conversely, as the velocity increases, λ = 1/5, the breakdown of adiabaticity introduces dynamic error. This result highlights the trade-off: exploring the geodesic flow on a closed, negatively curved surface requires long times, but doing so too quickly violates adiabaticity.

Ensemble-Accelerated Topological Characterization

To overcome the intrinsic limitations of single-particle trajectory tracking—most notably the pronounced sensitivity to initial conditions characteristic of chaotic hyperbolic dynamics—the authors introduce a multiparticle ensemble approach. This approach is rooted in ergodic theory, where statistical properties are inferred from a sufficiently large collection of independent trajectories rather than from a single path. The theoretical foundation justifies the use of ensemble equivalence and ensures that independent evolutions with distinct initial conditions yield statistically consistent measures of dynamical quantities in the longtime limit. By mapping each member of the ensemble to a separate quantum register and leveraging the intrinsic parallelism of quantum processors, their scheme permits simultaneous forward propagation of a large set of initial conditions. They distribute N virtual particles uniformly on the Bolza surface according to the induced volume measure p det(g)dx1∧dx2. The generalized topological response is averaged over the ensemble: ω˜(T) = 1/N X N i=1 1λ 2i T Z T 0. With the increase in particle number and the extension of evolution time, the distribution of the dynamical response becomes more concentrated, and the standard deviation gradually decreases, converging to the first Chern number of the quantum system. Increasing N reduces the time required for ˜ω(T) to stabilize at the quantized value.

Observation of Topological Phase Transitions

The authors demonstrate the universality of their method by probing topological phase transitions. By tuning the parameter ϵ in the Hamiltonian, they drive the system across distinct topological phases. For ϵ < 1, the system resides in a topologically nontrivial phase, where the ground-state bundle corresponds to a generator of the cohomology group H2 (Σ2). Outside this range, the topology becomes trivial (C1 = 0). The two spin-texture patterns plotted in Fig. 5(a) demonstrate that the system resides in two distinct topological phases. The excellent agreement between experiment and theory confirms that the HDQS can accurately capture global topological invariants through local dynamical responses. Extending the hyperbolically driven quantum system to accommodate higher-spin degrees of freedom—yielding the Chern number scaling C1 = 2n—unlocks a rich landscape of underlying physics.

Experimental Implementation and Characterization

The experiment was conducted on a superconducting quantum processor with transmon qubits. The continuous time evolution operator U(T) = P R T 0 e −iHˆ (xt)tdt was calculated by numerically integrating the Hamiltonian along the precomputed geodesic. This unitary was then decomposed into a sequence of standard single-qubit gates and implemented on the hardware. Quantum state tomography was performed to reconstruct the density matrix at each time step. The specific vector field used in the Hamiltonian is defined by a complex expression involvingomega, f(z), and ϵ. They chose parameters A = πe1/ρ2 and B = π/2, and ρ = 0.6 to ensure the smoothness of the mapping across the manifold.

Discussion

The authors successfully simulated a hyperbolically driven quantum system (HDQS) on a superconducting circuit platform. By mapping the geometric properties of a Bolza surface onto the Hamiltonian parameters, they revealed how the intrinsic geometry of the parameter space governs the dynamical observables of the quantum system. Their results elucidate the critical interplay between adiabaticity and ergodicity. While Hopf–Anosov theorem ensures that single-particle geodesics are ergodic, practical extraction is constrained by the need for adiabatic driving. This ensemble strategy effectively mitigates numerical instabilities inherent to long-time integration of chaotic systems while enhancing statistical significance. Furthermore, this work demonstrates the unique advantage of quantum simulation in parameter space. By leveraging the flexible controllability of superconducting qubits and the efficiency of parallel computing, this framework offers a novel avenue for exploring physical systems on exotic manifolds and provides fresh perspectives for manifold-based quantum machine learning.

Appendix

The trajectory of the classical particle in the phase space (z, p) = (x1 + ix2, p1 + ip2) ∈ C2 is governed by the canonical equations of motion. They numerically integrate these equations using a fourth-order Runge–Kutta method.

1. This system can perform quantum simulation of a Hyperbolically Driven Quantum System (HDQS) by mapping hyperbolic properties onto the dynamically controllable Hamiltonian parameters. This enables researchers to elucidate how the intrinsic geometry of the parameter space dictates observable physical quantities by probing system response under geometric quantum driving.

2. The AI system can realize ensemble simulation for topological characterization, leveraging superconducting qubits to capture essential geometric characteristics, establishing a versatile platform for exploring quantum dynamics on hyperbolic manifolds. This allows for efficient capture of geometric features, effectively transcending temporal constraints by using parallel computing to track a large set of initial conditions.

3. The system can perform ensemble-accelerated topological characterization by employing a scheme rooted in ergodic theory, where ensemble averages converge to the same value as time averages under appropriate conditions. This mitigates numerical instabilities inherent to chaotic systems and enhances statistical significance by averaging the generalized topological response: ω˜(T) = 1/N X sum N i=1 1/λ 2i T sum T=0.

4. The AI system can accurately probe topological phase transitions by comparing experimental results with theoretical predictions, as shown in Fig. 5(a) compares the experimentally extracted Chern number (red triangles) with theoretical predictions (blue line). This capability allows it to confirm that the HDQS can accurately capture global topological invariants through local dynamical responses.

5. The system can extend its physics modeling to higher-dimensional degrees of freedom, such as spin-1 or higher qudits, yielding Chern numbers C1 = 2n, where n denotes the spin. This unlocks the ability to study non-Abelian gauge fields and complex Berry connections within the negatively curved parameter space.