Theoretical tools used in processing continuous measurement records from real experiments to obtain quantum trajectories can easily lead to numerical errors due to a noninfinitesimal time resolution. In this work we propose a systematic assessment of the accuracy of a map. We perform error analyses for diffusive quantum trajectories, based on single-time-step Kraus operators proposed in the literature, and find the orders in time increments $\mathrm{\ensuremath{\Delta}}t$, to which such operators satisfy the conditions for valid average quantum evolution (completely positive, convex linear, and trace preserving), and the orders to which they match the Lindblad solutions. Given these error analyses, we propose a Kraus operator that satisfies the valid average quantum evolution conditions and agrees with the Lindblad master equation, to second order in $\mathrm{\ensuremath{\Delta}}t$, thus surpassing all other existing approaches. In order to test how well our proposed operator reproduces exact quantum trajectories, we analyze two examples of qubit measurement, where exact maps can be derived: a qubit subjected to a dispersive ($z$-basis) measurement and a fluorescence (dissipative) measurement. We show analytically that our proposed operator gives the smallest average trace distance to the exact quantum trajectories compared to existing approaches.
Theoretical tools used in processing continuous measurement records from real experiments to obtain quantum trajectories can easily lead to numerical errors due to a noninfinitesimal time resolution. In this work we propose a systematic assessment of the accuracy of a map. We perform error analyses for diffusive quantum trajectories, based on single-time-step Kraus operators proposed in the literature, and find the orders in time increments \ensuremathΔt, to which such operators satisfy the conditions for valid average quantum evolution (completely positive, convex linear, and trace preserving), and the orders to which they match the Lindblad solutions. Given these error analyses, we propose a Kraus operator that satisfies the valid average quantum evolution conditions and agrees with the Lindblad master equation, to second order in \ensuremathΔt, thus surpassing all other existing approaches. In order to test how well our proposed operator reproduces exact quantum trajectories, we analyze two examples of qubit measurement, where exact maps can be derived: a qubit subjected to a dispersive (z-basis) measurement and a fluorescence (dissipative) measurement. We show analytically that our proposed operator gives the smallest average trace distance to the exact quantum trajectories compared to existing approaches.
Quantum systems are the foundation of quantum technology development. To harness these systems, one must extract information from them and be able to control their dynamics. This thesis presents a body of work, containing both published and unpublished papers, that focuses on two key research areas: quantum continuous measurement and quantum bath sensing. While these topics may seem distinct, they are fundamentally connected through the framework of open quantum systems. As such, this thesis is structured into two parts. The first part of this thesis is dedicated to diffusive continuous quantum measurement, where the system of interest is continuously monitored through a coupled Markovian bath or meter. A key challenge in this context is the numerical errors introduced in quantum trajectory simulations caused by the finite time resolution ( 𝑡) in real experiments. To address this, we propose a systematic framework to assess theoretical tools used in processing measurement records. Focusing on a single Lindblad channel, we analyze the errors of existing measurement (or Kraus) operators in literature, in orders of 𝑡 under conditions that ensure valid average evolution (complete positivity, convex linearity, and trace preservation). We also analyze a stronger condition: finding the order at which these existing measurement operators match the Lindblad equation. Based on these analyzes, we propose a Kraus operator that satisfies the conditions for valid quantum evolution and matches the Lindblad master equation to second order in 𝑡, surpassing all existing approaches. Moreover, to evaluate how well our operator reproduces exact quantum trajectories, we analyze two qubit measurement examples where exact maps can be derived: 𝑧-basis measurement and fluorescence (dissipative) measurement. We show analytically that our proposed operator achieves the smallest average trace distance to the exact quantum trajectories compared to existing methods. Finally, we extend this Kraus operator to multiple Lindblad channels with a driving Hamiltonian. We test our analytical findings numerically using the two qubit examples, considering both single and multiple Lindblad channels with a driving Hamiltonian. The numerical results confirm our theoretical predictions. The second part of this thesis focuses on quantum bath sensing, where we use a quantum system (qubit) to extract information about its surrounding baths. Specifically, we develop protocols to detect nonclassical pure-dephasing noise in a single-qubit system coupled to temporally correlated (non-Markovian) baths. Inspired by the Ramsey protocol, we introduce a coherent intervention (𝜋 2 -pulse) within the protocol, which can reveal the quantum noise - noise correlations affecting the probe qubit that originate from quantum degrees of freedom. Our protocols are simple, experimentally viable, and rigorously derived for general bath models. Using an exactly solvable bosonic bath model, we analyse how key noise parameters respond to fixed control schemes, identifying the optimal strategy for detecting the quantum noise. From the perspective of bath dynamics, we further decompose the qubit-bath state, demonstrating that only Gaussian components of the bath contribute to quantum noise, highlighting the bath's role in system dynamics. Finally, we extend the modified Ramsey protocol by incorporating multi-axis 𝜋 2 pulse and demonstrate that it can emulate the modified Ramsey protocol with any completely positive trace-preserving intervention. This applies to both Gaussian and non-Gaussian baths under the pure-dephasing assumption.
To compare quantum estimation theory schemes we must acknowledge that, in some cases, the quantitative difference between them might be small and hence sensitive to numerical errors. Here, we are concerned with comparing estimation schemes for the quantum state under continuous measurement (quantum trajectories), namely quantum state filtering and, as introduced by us [Phys. Rev. Lett. 115, 180407 (2015)], quantum state smoothing. Unfortunately, the cumulative errors in the most typical simulations of quantum trajectories with time step $\mathrm{\ensuremath{\Delta}}t$ and total simulation time T can scale as $T\mathrm{\ensuremath{\Delta}}t$. Moreover, these errors may correspond to deviations from valid quantum evolution as described by a completely positive map. Here we introduce a higher-order method that reduces the cumulative errors in the complete positivity of the evolution to the order of $T{(\mathrm{\ensuremath{\Delta}}t)}^{2}$, whether for linear (unnormalized) or nonlinear (normalized) quantum trajectories. Our method also guarantees that the discrepancy in the average evolution between different detection methods (different ``unravelings,'' such as quantum jumps or quantum diffusion) is similarly small. This equivalence is essential for comparing quantum state filtering to quantum state smoothing, as the latter assumes that all irreversible evolution is unraveled, although the estimator only has direct knowledge of some records. In particular, here we compare the average difference between filtering and smoothing conditioned on an event of which the estimator lacks direct knowledge: a photon detection within a certain time window. We find that the smoothed state is actually less pure, both before and after the time of the jump. Similarly, the fidelity of the smoothed state with the ``true'' (maximal knowledge) state is also lower than that of the filtered state before the jump. However, after the jump, the fidelity of the smoothed state is higher.
Truly continuous quantum trajectories, obtained from homodyne or heterodyne readouts, can only ever be reconstructed approximately. The continuous measurement signal, needed for exact reconstruction, is averaged over bins of finite time Δt during any analog to digital conversion step. The best reconstruction possible, knowing only this discrete record, was introduced recently and dubbed the Robinet state. In this article, we show how the Robinet state can be computed with a numerical discretization scheme that is completely positive, accurate to arbitrarily high order in Δt, and that does not rely on any other external solver. Our derivation relies on a dilation of the stochastic master equation into a system + transmission line setup, constructed in such a way that measuring what we call the "zero mode" of the line yields the Robinet state. We test the method on a challenging example with random Hamiltonian and jump operator, and verify its accuracy up to order 10. Apart from its numerical interest, our approach provides a wealth of physical insights, extending in particular recent results on purity obtained by Wonglakhon, Chantasri, and Wiseman, that would be difficult to obtain in any other way.
Quantum trajectories are dynamical equations for quantum states conditioned on the results of a time-continuous measurement, such as a continuous-in-time current yt. Recently, there has been renewed interest in finite-interval dynamical maps for quantum trajectories, where the measurement record is available only over intervals of duration Δt. Guilmin \emph{et al.}\ (unpublished) derived such a map for the experimentally relevant case in which only a coarse-grained current (time-binned), denoted by It, is available. Remarkably, this coarse-grained record It still yields a conditioned state, termed the \emph{robinet} state, whose impurity scales as small as (Δt)3. We show, however, that the \emph{typical} trace distance between the robinet state and the fully conditioned state (obtained from the complete measurement record) scales as (Δt)3/2. We then introduce a higher-order finite-interval dynamical map, the ``nearly-exact'' map, which requires only one additional real statistic, φt, from each interval. We analytically show that the resulting conditioned state is much closer to the fully conditioned state, with a distance scaling as (Δt)2. We further numerically verify the distance scalings, measured relative to highly accurate approximations of the fully conditioned state, for the robinet map, our nearly-exact map, the lowest-order approximate Itô map, and two existing higher-order maps. Our results show that, for generic systems, whenever φt can be extracted experimentally alongside It, the nearly exact map provides the most accurate state estimate among the finite-interval maps considered.
Abstract For quantum diffusive measurements, the system’s dynamics can be described by the Itô stochastic master equation, which works well for an infinitesimal time resolution. However, in practical quantum experiments, one cannot make a time step to be infinitesimal, as it can introduce correlation of noise in time, making the Markovian assumption invalid. On the other hand, increasing a time step can cause errors from non-commuting operations describing the system’s dynamics. We therefore consider implementing the Euler-Milstein and relevant approaches, namely the Itô map, the high-order completely positive map, and the quantum Bayesian, to simulate quajntum trajectory for a quantum system under diffusive continuous measurements. In particular, we numerically simulate trajectories for a qubit measurement in z basis. We show the comparison of individual trajectories and their averaged trajectories among these approaches. We find that the high-order completely positive map approach yields the most accurate averaged quantum trajectory. Furthermore, we also investigate the trace distance from true stochastic quantum trajectories, comparing the four approaches using the numerical simulation. We show that, for a realistic time resolution (as in a superconducting qubit, experiment), the high-order map does give the most accurate estimate of the qubit trajectories.
Real-time reconstruction of conditional quantum states from continuous measurement records is a fundamental requirement for quantum feedback control, yet standard stochastic master equation (SME) solvers require exact model specification, known system parameters, and are sensitive to parameter mismatch. While neural sequence models can fit these stochastic dynamics, the unconstrained predictors can violate physicality such as positivity or trace constraints, leading to unstable rollouts and unphysical estimates. We propose a Kraus-structured output layer that converts the hidden representation of a generic sequence backbone into a completely positive trace preserving (CPTP) quantum operation, yielding physically valid state updates by construction. We instantiate this layer across diverse backbones, RNN, GRU, LSTM, TCN, ESN and Mamba; including Neural ODE as a comparative baseline, on stochastic trajectories characterized by parameter drift. Our evaluation reveals distinct trade-offs between gating mechanisms, linear recurrence, and global attention. Across all models, Kraus-LSTM achieves the strongest results, improving state estimation quality by 7% over its unconstrained counterpart while guaranteeing physically valid predictions in non-stationary regimes.