We present experimental validation of a software-defined geometric noise suppression protocol on a commercial superconducting quantum processor (Rigetti Ankaa-3 via AWS Braket). The protocol combines three elements: (1) time-reversal symmetric Figure-8 control pulses that create closed trajectories in parameter space, (2) node-synchronized stroboscopic readout that gates measurements at loop-closure times, and (3) adaptive phase tracking that compensates for environmental drift. Key experimental results include: break-even threshold exceeded with +10 percentage point improvement over baseline idle; phase-dependent noise suppression achieving +82 percentage point improvement through optimal phase selection; and discovery of amplitude-phase coupling (APC) where optimal phase shifts by ≈ 0.9 rad per unit amplitude change. This coupling, interpretable as an environmentally-induced Berry connection in control parameter space, explains why conventional fixed-parameter calibration fails at varying drive strengths. The adaptive protocol successfully tracks environmental drift, recovering from 13% to 85% fidelity autonomously. These results demonstrate that geometric structure in quantum control can be exploited for robust, hardware-agnostic noise suppression without physical device modification.
We present experimental validation of a software-defined geometric noise suppression protocol on a commercial superconducting quantum processor (Rigetti Ankaa-3 via AWS Braket). The protocol combines three elements: (1) time-reversal symmetric Figure-8 control pulses that create closed trajectories in parameter space, (2) node-synchronized stroboscopic readout that gates measurements at loop-closure times, and (3) adaptive phase tracking that compensates for environmental drift. Key experimental results include: break-even threshold exceeded with +10 percentage point improvement over baseline idle; phase-dependent noise suppression achieving +82 percentage point improvement through optimal phase selection; and discovery of amplitude-phase coupling (APC) where optimal phase shifts by ≈ 0.9 rad per unit amplitude change. This coupling, interpretable as an environmentally-induced Berry connection in control parameter space, explains why conventional fixed-parameter calibration fails at varying drive strengths. The adaptive protocol successfully tracks environmental drift, recovering from 13% to 85% fidelity autonomously. These results demonstrate that geometric structure in quantum control can be exploited for robust, hardware-agnostic noise suppression without physical device modification.
We present an experimental demonstration of a software-defined noise suppression protocolon a commercial superconducting quantum processor (Rigetti Ankaa-3 via AWS Braket).The protocol combines three layers: (1) Figure-8 control pulses whose 2:1 quadrature frequencyratio gives each loop a vanishing first Fourier moment, (2) node-synchronized readoutthat gates measurements at loop-closure times, and (3) adaptive phase tracking that compensatesfor environmental drift. Key experimental results include: a break-even thresholdexceeded with a +10 percentage-point improvement over baseline idle; phase-dependent suppressionyielding +82 percentage-point variation through optimal phase selection; and theexperimental observation of amplitude–phase coupling (APC), in which the optimal controlphase shifts by ≈ 0.9 rad per unit amplitude change. We report APC here as an operationalfeature of the measured control landscape; its microscopic interpretation in termsof average-Hamiltonian cancellation in a three-level Duffing model has been developed insubsequent work [2, 3]. The adaptive layer recovers fidelity from 13% to 85% autonomouslyafter environmental drift. These results demonstrate that the proposed protocol providesrobust, hardware-agnostic noise suppression without device modification.
Quantum computers can significantly speed up the solution of certain classical tasks and even provide solutions to problems that are classically intractable. However, achieving this requires a fault-tolerant architecture with components and quantum operations (gates) whose physical error rates are below a specific threshold. Advanced quantum control methods can suppress environmental noise and implement below-threshold gates, but there are infinitely many ways to do so. These methods can rely entirely on analytical solutions or combine them with numerical optimization, where pulse properties are treated as optimization goals. Analytical constraints are always met exactly, but numerically optimized goals may only be satisfied approximately, leading to tradeoffs. Therefore, a modular analytical framework is needed that fixes some objectives in advance while optimizing only a selected subset. In this dissertation, I employ Space Curve Quantum Control (SCQC), a method that generates noise-resistant quantum gates using differential geometry. In SCQC, a space curve's global geometric properties dictate the gate's noise-robustness, while the controls that realize it are extracted from its local quantities. The first part of the dissertation is devoted to extending SCQC for arbitrary and simultaneous noise suppression, with experimentally-friendly pulses. I first derive the geometric conditions and curve examples for suppressing two separate noise types simultaneously. Then, I extend SCQC by deriving the adjoint representation equations of the associated quantum evolution. This allows one to describe the entire control design in terms of space curve elements and propose a systematic control design approach called the Bézier Ansatz for Robust Quantum (BARQ) control method. BARQ can encode the target gate upfront and allow optimization only for noise-robustness properties. This approach solves the usually unavoidable tradeoff between gate fidelity and noise-sensitivity and is capable of producing control pulses with experimentally-friendly features. The second part of the dissertation applies BARQ in two leading qubit modalities. I first describe my participation in experiments on IBM's superconducting devices, where doubly-robust BARQ pulses demonstrate at least one order-of-magnitude error reduction in the presence of engineered noise. The second platform is a four-qubit trapped-ion register hosted in the Quantum Scientific Computing Open User Testbed (QSCOUT) ecosystem. I show that utilizing BARQ pulses can reduce gate errors up to two times and effectively combat motion-induced decoherence.
We present a public summary of Informational Fiber Synchronization (IFS), a geometriccontrol protocol designed to suppress low-frequency parameter drift (1/f noise) in superconductingquantum processors. By synthesizing a ”Figure-8” loop trajectory in the control parameterspace, IFS generates a topological protection mechanism akin to Virtual Anyons, effectively cancellingdynamic phase errors via geometric echo. Simulation results demonstrate gate fidelity > 0.99under strong (15%) control jitter where standard protocols fail. We also report preliminary pulsesynthesis on IBM Quantum hardware. Note: Specific algorithmic details regarding the phase refinementand estimator logic are redacted in this public version to preserve intellectual property. Fulldetails are available for review under a non-disclosure agreement (NDA) This is a concept paper. Detailed algorithms and source codes are available for collaborative research partners.
We report phase-resolved measurements of amplitude-phase coupling (APC) in strongly driven superconducting transmon qubits, a symmetry-based mechanism analysis, and follow-up experiments on a second device. On a Rigetti Ankaa-3 transmon (1000 shots per point, five drive amplitudes, matched root-mean-square drive), the fitted optimal control phase
ϕ∗(A) drifts by more than 3 rad across the explored amplitude range for both a rectangular reference pulse and a self-intersecting Figure-8 trajectory, and the phase contrast of the two waveform families exhibits a reproducible low-/high-amplitude crossover. Theoretically, we prove that in the resonant rotating frame of the three-level Duffing model a constant-phase rectangular drive cannot produce any phase-dependent population response:
CRect=0 exactly. Any measured rectangular phase contrast is therefore a probe of physics outside this model (beyond-rotating-wave carrier-phase dependence, control-chain distortion, readout imperfections). The Figure-8 waveform breaks the symmetry intrinsically through its 2:1 Fourier content; a third-order resonance
ω+ω−2ω=0 mediated by the anharmonic level
∣2⟩ predicts
CF8∝A6N2 at weak drive, confirmed by direct numerical propagation (fitted log-log slope
5.97±0.02; loop-number ratio 2434 vs. predicted 2500). On the successor device Cepheus-1-108Q, the predicted phase contrast lies below the binomial shot-noise floor at the accessible amplitudes, but the same dataset reveals a stronger signature of the underlying first-order average-Hamiltonian cancellation: the phase-averaged survival probability of the Figure-8 train remains pinned at
⟨P(∣0⟩)⟩ϕ≈0.985 across all tested conditions, while the isenergetic rectangular reference sweeps the full range 0.24-0.93—an approximately 70-fold suppression of the Rabi-drive response. Together these results identify APC as a measurable control-landscape feature with distinct intrinsic and extrinsic components, and demonstrate a calibration-free, amplitude-insensitive idling property of Figure-8 pulse trains.
High-fidelity logical \emph{T}-gate realization constitutes a core prerequisite for large-scale fault-tolerant quantum computing. However, conventional magic state distillation requires massive physical qubit overhead across successive distillation rounds, alongside sophisticated measurement and feedback control, thereby inducing considerable spatial and temporal resource consumption. Herein, we propose a controlled superconducting geometric logical \emph{T} gate scheme that achieves high-order suppression of universal errors, by integrating decoherence-free subspace encoding with multi-loop optimized composite geometric pulse engineering. Guided by tailored trajectory design, we systematically establish unified gate construction frameworks for conventional geometric, composite geometric, and optimized composite geometric protocols. By flexibly controling additional parametric degrees of freedom, the proposed scheme achieves substantially enhanced robustness against diverse noise sources. Numerical simulations reveal that, within tunable superconducting quantum circuits, our geometric logical \emph{T} gate outperforms both conventional composite geometric and dynamical gates in suppressing Rabi frequency, detuning, and residual inter-qubit crosstalk errors that can all be suppressed to the fourth order, while additionally providing inherent suppression of collective dephasing errors. The present strategy alleviates intrinsic limitations of mainstream approaches and opens a promising avenue toward robust high-fidelity logical \emph{T} gate construction.