arXiv subjects

Mauro Paternostro

Publications and source records attributed to Mauro Paternostro.

At least 19 recordsLinked to original sources

arXiv2026

Triple descent in quantum linear regression

Quantum linear regression models, such as quantum extreme learning machines and quantum fidelity kernels, are trained on features that can only be estimated from finitely many measurement shots. Sampling noise is usually treated as a nuisance that more shots suppress; we show instead that it can fundamentally reshape the instabilities of the learning problem. As the number of measurement outcomes grows, the prediction error displays a triple-descent profile, with two peaks of fundamentally different origins. The first, classical, peak occurs when the measurement first becomes informationally complete, is a direct analogue of double descent in classical linear regression, and is caused by ill-conditioning of the measurement itself. This peak affects predictions only through their variance, and is thus suppressed by increasing the number of shots at inference. The second, quantum, peak occurs beyond informational completeness when the number of outcomes matches the number of training states: shot noise populates unphysical outcome combinations and the regression fits these pure-noise features with large weights, corrupting the learned model itself. More training or test shots cannot remove this quantum peak, which disappears only in the ideal limit of exact training features. Using random-matrix theory, we derive the shape of both peaks analytically, accounting also for ridge regularization. Our results show that treating quantum features as exactly known overestimates classical instabilities while missing genuinely quantum ones.

quant-ph↗

arXiv2026

Improving device-independent quantum key distribution protocols through multiple routed Bell tests

Device-independent quantum key distribution (DI-QKD) guarantees security with a minimal set of assumptions about the experimental setup. The challenge posed by its implementation could be tackled using routed Bell tests with entanglement swapping, or distant Bell state measurement (BSM) units. However, practical distances still require local tests with close-to-ideal violations or BSMs with high efficiency. We propose a DI-QKD protocol based on a multi-port BSM, and multiple sources and measurement devices, where in each round, routed tests are performed on randomly selected local devices. The violation of local Bell tests is checked even for the subset of rounds where successful BSM projections are achieved. By requiring that this violation remain consistent with the overall average, we specifically show improvements in the critical efficiencies of a three-port BSM, of up to $6\%$ in absolute terms for high visibilities. Without ideal violations, we also achieve arbitrarily small efficiencies with arbitrarily many BSM ports. Our approach enables long-distance DI-QKD, with access to highly efficient loophole-free routing setups, and multiple local tests (possibly imperfect) with very high local detection efficiencies. Remarkably, it avoids the use of quantum memories and the need for long-range Bell tests. Finally, we extend the concept of routing to dimension witnesses, where qubit-bounded sources send states to the BSM. This can be seen as a semi-device-independent extension of the aforementioned protocol.

quant-ph↗

arXiv2026

Who can sample forever? Shot noise effects in quantum linear regression

Quantum fidelity kernels and quantum extreme learning machines share the same basic architecture: a fixed quantum device generates features that are combined through a classically trained linear model. We refer to this broad non-variational setting as quantum linear regression. A defining aspect of these models is that the quantum features are not directly available, but must be estimated from a finite number of shots. We show that finite statistics makes quantum linear regression unavoidably ill-conditioned. Ideal feature vectors are confined to a subspace determined by the dimension of the Hilbert space, whereas finite-shot estimates generically acquire components outside this subspace. Training then attempts to fit directions that disappear in the infinite-statistics limit, making the learned model increasingly sensitive to sampling noise. We analyze this mechanism for quantum extreme learning machines trained on $n_{\rm tr}$ states with $N$ shots per state. Despite the ill-conditioning, predictions remain controlled when the training targets are exact. Finite sampling produces a systematic prediction error that decreases quadratically with $N$, while fluctuations induced by the random training measurements decrease inversely with the total training budget. Thus, unlike in standard linear regression with exactly known features, increasing the dataset size alone cannot remove the error caused by sampling noise. If the training targets are also noisy, due for instance to state-preparation errors, the otherwise hidden directions become active. The test error can then increase with $N$, a behavior that we ascribe to a form of quantum overfitting. Our results expose a fundamental tradeoff between dataset size and per-state statistics in non-variational quantum learning and provide a baseline for resource-aware training and regularization strategies.

quant-ph↗

arXiv2026

White paper: 1-10 Hz matter-wave interferometer to test the spin entanglement witness for quantum gravity

In this white paper, we highlight the importance of the ($1-10~{\rm Hz}$) frequency range for laboratory tests of the quantum nature of gravity using the quantum gravity-induced entanglement of masses (QGEM) protocol. QGEM requires matter-wave interferometers with masses ($m\sim10^{-15}-10^{-14}~{\rm kg}$), brought within separations ($d\sim30-50~μ{\rm m}$), while maintaining spatial superpositions of ($1-20~μ{\rm m}$) and coherence for ($τ\sim 0.1 - 1~{\rm s}$). These requirements make low-frequency environmental noise a central experimental challenge and place QGEM in a regime closely related to the low-frequency goals of the Einstein Telescope (ET) and the Cosmic Explorer (CE). In particular, QGEM is sensitive to relative acceleration noise (RAN) and to gravity-gradient noise (GGN) generated by seismic and other environmental mass-density fluctuations. For representative parameters $m=10^{-14}~{\rm kg}$, $Δx=10~μ{\rm m}$, and $τ=1~{\rm s}$, the differential acceleration-noise amplitude spectral density must be suppressed well below the $10^{-15}~{\rm m\,s^{-2}/\sqrt{Hz}}$ level to keep acceleration-induced dephasing below the relevant experimental scale. Achieving this level of low-frequency noise suppression is therefore a key requirement for QGEM and closely parallels the seismic and gravity-gradient noise challenges that ET and CE address.

quant-ph↗

arXiv2026

Machine Learning-Based Characterisation of the Non-Markovian Dynamics of a Nitrogen-Vacancy Centre

The interaction between a quantum system and its environment can be characterized by the spectral density function: knowing its structure is important for optimizing applications of quantum technologies such as quantum sensing protocols. In this work, we present the first experimental demonstration of a machine learning-based reconstruction of reaction-coordinate spectral density parameters from NV centre Rabi dynamics. Unlike the previous work, we recover all spectral density parameters rather than only the central frequency, and benchmark the performance of the neural network against the Cramér-Rao bound and maximum likelihood estimator. Our results demonstrate that the model predicted by the neural network can reliably reproduce the NV dynamics over the estimation window, and can produce estimates for some parameters with variances comparable to that of maximum likelihood.

quant-ph↗

arXiv2026

Energetics in daemonic work extraction protocols via non-ideal QND-energy measurement

We address the problem of extracting work from a quantum system assisted by a quantum non-demolition (QND) energy measurement. When a perfect QND measurement can be performed and an auxiliary zero-temperature bath is available, the full energy of the quantum state can in principle be extracted even without any prior information on the input state. Owing to the presence of a zero-temperature bath, this is achieved at no energetic cost for the measurement process itself. On the contrary, here we consider what happens when the same protocol is implemented in non-ideal scenarios, specifically when the auxiliary bath has a finite temperature. In this case, not only is it impossible to extract the entire energy from the system, but the measurement strategy also acquires a non-zero energetic cost, accounting for both the interaction between system and measurement apparatus, and the corresponding Landauer erasure cost. We quantitatively assess the performance of these work-extraction protocols, both in absolute terms and through the so-called daemonic net gain, which explicitly includes the energetic cost of the measurement. We rigorously prove that, when access to a thermal bath is allowed in the extraction protocol, the daemonic net gain is always non-positive for any temperature of the auxiliary bath. Conversely, when only unitary operations are considered, the daemonic net gain can attain positive values. We further discuss these different figures of merit by analyzing a paradigmatic example for a single qubit system.

quant-ph↗

arXiv2026

A general estimation framework for continuous-variable systems

We show that informational completeness, while sufficient to have a bijection between ideal measurement probabilities and quantum states, does not guarantee statistically stable reconstruction from finite measurement data. To address this problem, we develop a general estimation theory for continuous-variable systems in which stable reconstructibility is characterized by the POVM effects forming a measurement frame. Informational completeness is therefore necessary, but not sufficient, for stable reconstruction. Our framework is based on measurement frames in a $σ$-regularized operator geometry, where the reference state $σ$ encodes prior information about relevant features of the measured states. For any fixed measurement scheme, observables may be inaccessible, weakly reconstructible only through estimators with divergent variance, or stably reconstructible by finite-variance unbiased estimators. The relevant regime is determined by the range of the POVM synthesis operator. Our framework provides practical methods for constructing estimators and gives an operational interpretation of singular quasiprobability distributions, including the Glauber-Sudarshan $P$ representation: quasiprobabilities act as unbiased estimators for associated observables, and their singularities reflect a pathological feature of the corresponding measurement: its lack of loewr frame bound. We furthermore show how this formalism naturally provides operational regularization procedures tied to prior information. Overall, our framework provides a unified view of continuous-variable tomography, quasiprobability representations, and classical-shadow estimation.

quant-ph↗

arXiv2026

Broken Detailed Balance and Entropy Production in CPTP Quantum Brownian Motion

We rigorously analyze the non-equilibrium thermodynamic behavior of various formulations of quantum Brownian motion (QBM) using the framework of stochastic thermodynamics. While the widely used Caldeira-Leggett master equation exhibits desirable thermodynamic features, such as the fulfilment of a detailed balance, it fails to ensure complete positivity. In contrast, several completely positive and trace-preserving (CPTP) extensions turn out to be thermodynamically controversial. We show that such extensions introduce anomalous phase-space structures that violate detailed balance at the steady state, leading to non-vanishing entropy production and effective non-equilibrium current of unclear physical origins. Our results highlight a fundamental tension between quantum consistency and thermodynamic equilibration in open quantum systems.

quant-ph↗

arXiv2026

Stochastic Quantum Information Geometry and Speed Limits at the Trajectory Level

In quantum metrology, precision is typically characterized by an ensemble-averaged quantity, the quantum Fisher information (QFI), which averages over the fluctuations of individual measurement records. Here we introduce the conditional quantum Fisher information (CQFI), a trajectory-level version of the QFI that generalizes the classical stochastic Fisher information to the quantum domain. Defined through the symmetric logarithmic derivative and conditioned on a measurement outcome, the CQFI is a random variable whose average recovers the QFI. Using it, we derive a trajectory-level quantum speed limit, illustrated by the quantum-jump unraveling of a driven thermal qubit. Moreover, the CQFI decomposes into incoherent (population) and coherent (basis-rotation) contributions, together with an interference cross-term. This cross-term vanishes on average but can take negative values along single trajectories, providing a local witness of destructive interference between classical and quantum information channels.

quant-ph↗

arXiv2026

Entangled photons from quantum-dot-cavity systems under non-Markovian decoherence by pulsed excitation

Cascaded emission from the biexciton state of a quantum dot results in polarization entangled photon pairs. Cavity-enhancement of the direct two-photon emission channel bypasses the dominant source of decoherence in this system - that due to fine-structure splitting of the exciton levels. Here, we investigate the remaining non-Markovian phonon-induced decoherence using the numerically exact uniTEMPO algorithm. We compute the degree of entanglement of photon pairs generated by pulsed two-photon resonant excitation and find surprisingly good agreement between the numerically exact results and those calculated using the approximate polaron master equation method, permitting an efficient exploration of trends across system parameters.

quant-ph↗

arXiv2026

Non-equilibrium quantum thermodynamics of a memory-bearing open-system process

We show the emergence of memory effects in the dynamics of a driven two-level system interacting with a composite environment, and analyze their influence on work, heat and entropy production. We further investigate how the interplay between driving, dissipation and memory effects, stemming from the finiteness of the environment, shapes the thermodynamic response of the system, thus providing insight into quantum thermodynamics beyond the Markovian approximation.

quant-ph↗

arXiv2026

Non-equilibrium thermodynamics of collapse models in the strongly non-Gaussian regime

Standard objective collapse models offer a unified approach to the quantum measurement problem but predict an unphysical, indefinite increase in the energy of the system. The dissipative Diósi-Penrose (dDP) model resolves this heating issue by introducing a linear friction mechanism. However, this modification induces complex, non-Gaussian phase-space dynamics. We rigorously establish the thermodynamic consistency of this friction mechanism -- extended to the CSL model -- across both weakly and strongly non-Gaussian regimes. Using the Wigner phase-space formalism, we go significantly beyond the quadratic approximation and, to bypass the failure of perturbative methods under strong dissipation, introduce a novel exact pseudo-spectral simulation approach. Our analysis reveals that the system subjected to the dDP mechanism does not thermalize, but rather settles into a non-equilibrium steady-state (NESS) where the asymptotic non-Gaussianity scales as the third power of the dissipation parameter $β$. By evaluating the Wigner entropy production, we confirm the thermodynamic validity of the model and demonstrate that highly sensitive information-theoretic quantities require exact numerical methods to accurately capture the key non-Gaussian tails of the distribution.

quant-ph↗

arXiv2026

Detection of noise correlations in two qubit systems by Machine Learning

We introduce and validate a machine-learning assisted quantum sensing protocol to classify spatial and temporal correlations of classical noise affecting two ultrastrongly coupled qubits. We consider six distinct classes of Markovian and non-Markovian noise. Leveraging the sensitivity of a coherent population transfer protocol under three distinct driving conditions, the various forms of noise are discriminated by only measuring the final transfer efficiencies. Our approach achieves $\gtrsim 94\%$ accuracy in classification providing a near-perfect discrimination between Markovian and non-Markovian noise. The method requires minimal experimental resources, relying on a simple driving scheme providing three inputs to a shallow neural network with no need of measuring time-series data or real-time monitoring. The machine-learning data analysis acquires information from non-idealities of the coherent protocol highlighting how combining these techniques may significantly improve the characterization of quantum-hardware.

quant-ph↗

arXiv2026

Efficient classical training of model-free quantum photonic reservoir

Model-independent estimation of the properties of quantum states is a central challenge in quantum technologies, as experimental imperfections, drifts, and imprecise models of the actual quantum dynamics inevitably hinder accurate reconstructions. Here, we introduce a training strategy for photonic quantum extreme learning machines in which both the learning stage and the optimization of the measurement settings are performed entirely with classical light, while inference is carried out on genuinely quantum states. The protocol is based on the identity between the normalized output intensities following the evolution of coherent states through a linear optical reservoir, and the output statistics obtained with separable input quantum states. Building on this correspondence, we implemented a model-free, gradient-based optimization of the reservoir measurement projection directly on experimental data, without relying on a prior model of the device transformation. We experimentally show that the resulting classical-to-quantum transfer enables accurate reconstruction of single-qubit Pauli observables for previously unseen single-photon states, and extends to the estimation of a two-qubit entanglement witness for arbitrary bipartite states. Beyond demonstrating a qualitatively distinct form of out-of-distribution generalization across the classical-to-quantum boundary, our results identify a practical route to fast, adaptive, and resource-efficient training of photonic quantum learning devices.

quant-ph↗

arXiv2026

Robustness of the quantum Mpemba effect against state-preparation errors

The quantum Mpemba effect (QME) is a phenomenon observed in many-body systems where initial systems configurations farther from equilibrium can be observed to equilibrate faster than configurations that are closer to it. By considering noise induced error in the initial system state preparation, we analyse the robustness of various models exhibiting the QME. We demonstrate that exponentially accelerated thermalisation in open system dynamics modelled by a Gorini-Kossakowski-Sudarshan-Lindblad master equation is highly sensitive to noise induced deviations in the initial state, making this approach to accelerated thermalisation difficult to achieve. In contrast, we demonstrate that accelerated restoration of symmetry in $U(1)$ symmetric random unitary circuits via increased initial symmetry breaking is robust in the presence of state preparation error. When large errors are present in the state preparation, we show that this can in fact induce a higher rate of symmetry restoration and a stronger QME.

quant-ph↗

arXiv2026

Quantum Metrology of Newton's Constant with Levitated Mechanical Systems

Newton's constant is the least well-measured among the fundamental constants of Nature, and, indeed, its accurate measurement has long served an experimental challenge. Levitated mechanical systems are attracting growing attention for their promising applications in sensing and as an experimental platform for exploring the intersection between quantum physics and gravitation. Here we propose a mechanical interferometric scheme of interacting levitated oscillators for the accurate estimation of Newton's constant. Our scheme promises to beat the current standard by several orders of magnitude.

quant-ph↗

arXiv2026

Roadmap on Quantum Thermodynamics

The last two decades has seen quantum thermodynamics become a well established field of research in its own right. In that time, it has demonstrated a remarkably broad applicability, ranging from providing foundational advances in the understanding of how thermodynamic principles apply at the nano-scale and in the presence of quantum coherence, to providing a guiding framework for the development of efficient quantum devices. Exquisite levels of control have allowed state-of-the-art experimental platforms to explore energetics and thermodynamics at the smallest scales which has in turn helped to drive theoretical advances. This Roadmap provides an overview of the recent developments across many of the field's sub-disciplines, assessing the key challenges and future prospects, providing a guide for its near term progress.

quant-ph↗

arXiv2026

Revealing the non-classicality of a molecular nanomagnet

Molecular nanomagnets are compounds characterized by a high-spin magnetic core that is protected by organic ligands. They have recently gained attention as potential quantum information carriers in solid-state quantum computing platforms, simultaneously exhibiting classical macroscopic properties and quantum features in light of their complex nature and configuration. Addressing the condition when they manifest unquestionable quantum behavior is key to guarantee their effectiveness as resources for quantum information processing. We address the quantumness of molecular nanomagnets using a recently formulated criterion [cf. Krisnanda et al., Phys. Rev. Lett. 119, 120402 (2017)] demonstrating that these systems exhibit an intrinsic quantum nature, as evidenced by their ability to generate and enhance quantum correlations between two non-interacting probes. Our analysis, which is performed addressing various dynamical regimes, paves the way to the design of experimentally viable tests of non-classicality in multipartite registers consisting of ensembles of molecular nanomagnets.

quant-ph↗