Prediction and real-time compensation of qubit decoherence via machine-learning
Control engineering techniques are emerging as a promising approach to realize the stabilisation of quantum systems, and a powerful complement to attempts to design-in passive robustness. However, applications to date have largely been limited by the challenge that projective measurement of quantum devices causes the collapse of quantum superposition states. As a result significant tradeoffs have been mandated in applying the concept of feedback, and experiments have relied on open-loop control, weak measurements, access to ancilla states, or largely sacrificing quantum coherence in the controlled system. In this work we use techniques from control theory and machine learning to enable the real-time feedback suppression of semiclassical decoherence in a qubit when access to measurements is limited. Using a time-series of measurements of a qubit's phase we are able to predict future stochastic evolution without requiring a deterministic model of qubit evolution. We demonstrate this capability by preemptively stabilising predicted qubit dephasing in two experiments. First, we realise periods of stabilised qubit operation during which projective measurements are not performed via a non-destructive time-division multiplexed approach. Second, we implement predictive feedback in a periodic loop where the presence of long free-evolution periods normally causes decorrelation between measurement outcomes and the qubit state at the time of control actuation. Both experiments demonstrate quantitative improvements in qubit phase stability relative to "traditional" measurement-based feedback approaches, including enhanced long-term qubit phase stabilisation. Our approach is extremely simple and applicable to any qubit with the ability to perform projective measurement, requiring no hardware modifications, alternate measurement strategies, or access to exotic ancilla states.
Optimisation of Quantum Control Feedback Techniques in an Ytterbium Ion Trap
Submitted Honours Thesis 2015.
Before quantum control techniques can be applied to any commercial engineering application, a deep understanding of the precision and accuracy of feedback methods in a quantum setting must be obtained. Similar to most engineered systems, ion control is hampered by a several sources and types of noise ranging from environmental to hardware specific noise which lead to deviations of an oscillator signal from a desired position over time. This thesis attempts to use knowledge about common sources of noise in an ion trapping implementation to assess the effectiveness and precision of feedback methods at suppression and stabilisation of the oscillator deviations. Traditional feedback methods are compared against a novel predictive protocol shown to work theoretically by some members in the quantum control community. Theoretically, the predictive technique uses offline knowledge of the noise frequency spectrum to linearly combine several previous measurements of the noise to perform a more accurate correction than other methods. This thesis explores experimentally the advantages the novel protocol has over traditional feedback methods in precision and accuracy as well as further optimisation techniques using experimental data. Ultimately, the assessment is made as to whether the techniques can be applied practically to enhance the users suppression of error in the ion trap system.