Subscribe to the gold package and get unlimited access to Shamra Academy
Register a new userMonitoring fast superconducting qubit dynamics using a neural network
74
0
0.0
(
0
)
Ask ChatGPT about the research
No Arabic abstract
Weak measurements of a superconducting qubit produce noisy voltage signals that are weakly correlated with the qubit state. To recover individual quantum trajectories from these noisy signals, traditional methods require slow qubit dynamics and substantial prior information in the form of calibration experiments. Monitoring rapid qubit dynamics, e.g. during quantum gates, requires more complicated methods with increased demand for prior information. Here, we experimentally demonstrate an alternative method for accurately tracking rapidly driven superconducting qubit trajectories that uses a Long-Short Term Memory (LSTM) artificial neural network with minimal prior information. Despite few training assumptions, the LSTM produces trajectories that include qubit-readout resonator correlations due to a finite detection bandwidth. In addition to revealing rotated measurement eigenstates and a reduced measurement rate in agreement with theory for a fixed drive, the trained LSTM also correctly reconstructs evolution for an unknown drive with rapid modulation. Our work enables new applications of weak measurements with faster or initially unknown qubit dynamics, such as the diagnosis of coherent errors in quantum gates.
rate research
Read More
At its core, Quantum Mechanics is a theory developed to describe fundamental observations in the spectroscopy of solids and gases. Despite these practical roots, however, quantum theory is infamous for being highly counterintuitive, largely due to its intrinsically probabilistic nature. Neural networks have recently emerged as a powerful tool that can extract non-trivial correlations in vast datasets. They routinely outperform state-of-the-art techniques in language translation, medical diagnosis and image recognition. It remains to be seen if neural networks can be trained to predict stochastic quantum evolution without a priori specifying the rules of quantum theory. Here, we demonstrate that a recurrent neural network can be trained in real time to infer the individual quantum trajectories associated with the evolution of a superconducting qubit under unitary evolution, decoherence and continuous measurement from raw observations only. The network extracts the system Hamiltonian, measurement operators and physical parameters. It is also able to perform tomography of an unknown initial state without any prior calibration. This method has potential to greatly simplify and enhance tasks in quantum systems such as noise characterization, parameter estimation, feedback and optimization of quantum control.
Demonstrating a quantum computational advantage will require high-fidelity control and readout of multi-qubit systems. As system size increases, multiplexed qubit readout becomes a practical necessity to limit the growth of resource overhead. Many contemporary qubit-state discriminators presume single-qubit operating conditions or require considerable computational effort, limiting their potential extensibility. Here, we present multi-qubit readout using neural networks as state discriminators. We compare our approach to contemporary methods employed on a quantum device with five superconducting qubits and frequency-multiplexed readout. We find that fully-connected feedforward neural networks increase the qubit-state-assignment fidelity for our system. Relative to contemporary discriminators, the assignment error rate is reduced by up to 25% due to the compensation of system-dependent nonidealities such as readout crosstalk which is reduced by up to one order of magnitude. Our work demonstrates a potentially extensible building block for high-fidelity readout relevant to both near-term devices and future fault-tolerant systems.
We demonstrate fast two-qubit gates using a parity-violated superconducting qubit consisting of a capacitively-shunted asymmetric Josephson-junction loop under a finite magnetic flux bias. The second-order nonlinearity manifesting in the qubit enables the interaction with a neighboring single-junction transmon qubit via first-order inter-qubit sideband transitions with Rabi frequencies up to 30~MHz. Simultaneously, the unwanted static longitudinal~(ZZ) interaction is eliminated with ac Stark shifts induced by a continuous microwave drive near-resonant to the sideband transitions. The average fidelities of the two-qubit gates are evaluated with randomized benchmarking as 0.967, 0.951, 0.956 for CZ, iSWAP and SWAP gates, respectively.
While relatively easy to engineer, static transverse coupling between a qubit and a cavity mode satisfies the criteria for a quantum non-demolition (QND) measurement only if the coupling between the qubit and cavity is much less than their mutual detuning. This can put significant limits on the speed of the measurement, requiring trade-offs in the circuit design between coupling, detuning, and decoherence introduced by the cavity mode. Here, we study a circuit in which the qubit-cavity and the cavity-feedline coupling can be turned on and off, which helps to isolate the qubit. We do not rely on the rotating-wave or dispersive approximations, but solve the full transverse interaction between the qubit and the cavity mode. We show that by carefully choosing the detuning and interaction time, we can exploit a recurrence in the qubit-cavity dynamics in a way that makes it possible to perform very fast, high fidelity, QND measurements. Here, the qubit measurement is performed more like a gate operation between the qubit and the cavity, where the cavity state can be amplified, squeezed, and released in a time-sequenced fashion. In addition, we also show that the non-demolition property of the off-resonant approximation breaks down much faster than its dispersive property, suggesting that many of the dispersive measurements to date have been implemented outside the QND regime.
Superconducting qubits are a leading candidate for quantum computing but display temporal fluctuations in their energy relaxation times T1. This introduces instabilities in multi-qubit device performance. Furthermore, autocorrelation in these time fluctuations introduces challenges for obtaining representative measures of T1 for process optimization and device screening. These T1 fluctuations are often attributed to time varying coupling of the qubit to defects, putative two level systems (TLSs). In this work, we develop a technique to probe the spectral and temporal dynamics of T1 in single junction transmons by repeated T1 measurements in the frequency vicinity of the bare qubit transition, via the AC-Stark effect. Across 10 qubits, we observe strong correlations between the mean T1 averaged over approximately nine months and a snapshot of an equally weighted T1 average over the Stark shifted frequency range. These observations are suggestive of an ergodic-like spectral diffusion of TLSs dominating T1, and offer a promising path to more rapid T1 characterization for device screening and process optimization.
Log in to be able to interact and post comments
comments
Fetching comments
Sign in to be able to follow your search criteria
