arXiv Science⌕ Search

arXiv subjects

Cosmo Lupo

Publications and source records attributed to Cosmo Lupo.

At least 19 recordsLinked to original sources

Invariance can prevent overfitting and shortcut learning in Quantum Neural Networks

Many image classification tasks have labels that are unchanged by rotations and reflections. Equivariant quantum neural networks incorporate such symmetries by design, but the specific benefit of symmetry is difficult to isolate. Equivariant quantum circuits are usually compared with architectures that also differ in gates, structure or size. Here, we investigate the role of invariance in a controlled setting, classifying images whose labels are invariant under the dihedral group $D_4$. We compare four eight-qubit models with identical encoding, circuit structure and number of trainable parameters, and equivalent dynamical Lie algebras, and benchmark invariant implementations against non-invariant ones in terms of generalization error, robustness to shortcut learning and resilience to a realistic noise model. On four binary tasks, including MNIST and two scientific data sets of galaxies and red blood cells, the invariant models trained on a few tens of images show better generalization than non-invariant ones. Invariance also makes the models immune to orientation shortcuts: when trained with a class-dependent corner watermark, the non-invariant models learn the watermark and often drop to chance level when it is moved. We show that the gain in accuracy of invariant models persists also under a realistic noise model.

quant-ph↗

Preprocessing noise in finite-size quantum key distribution

It is known that trusted, preprocessing noise may boost quantum key distribution by expanding the range of values of tolerated noise. For BB84, adding trusted noise may allow the generation of secret keys even for qubit error rate (QBER) beyond the 11% threshold in the asymptotic regime. Here we study the effect of preprocessing noise in the finite-size regime where only a limited number of signals are exchanged between Alice and Bob. We compute the secret key rate from the full variational formulation of the sandwiched Renyi conditional entropy, optimized through nonsymmetric conic optimization. We find that trusted noise improves the key rate only for a finite interval of alpha, from the alpha approaching 1 limit up to alpha approximately 1.4. We then optimize the finite-size key rate jointly over both q, alpha, and the fraction of raw keys used for parameter estimation, to study how trusted preprocessing changes the maximum tolerable QBER as the block size increases. We show that the secret-key rate enhancement provided by optimized noisy preprocessing becomes more pronounced at higher QBER levels and can increase the maximum tolerable QBER over a broad range of finite block sizes.

quant-ph↗

Explicit construction of a 2-design of ${\rm U}(2)$ from the theory of angular momentum

The main aim of this work is to present an explicit construction of a 2-design of ${\rm U}(2)$, relying only on a tool that belongs to every physicists toolbox: the theory of angular momentum. Unitary designs are a rich and fundamental mathematical topic, with numerous fruitful applications in quantum information science and technology. In this work we take a peek under the hood. We begin with a minimal set of definitions and characterizations. Then we derive all 1-designs of ${\rm U}(2)$ of minimum size. Finally, we set out, step by step, a completion procedure extending such 1-designs to 2-designs. In particular, starting from the Pauli basis $\unicode{x2014}$ the prototypical unitary 1-design $\unicode{x2014}$ one $\unicode{x201C}$naturally$\unicode{x201D}$ obtains the 2-design originally employed by Bennett and coauthors in $\textit{Mixed State Entanglement and Quantum Error Correction}$. The present work also serves as a gentle and largely self-contained introduction to the subject.

quant-ph↗

Security of hybrid BB84 with heterodyne detection

Quantum key distribution (QKD) promises everlasting security based on the laws of physics. Most common protocols are grouped into two distinct categories based on the degrees of freedom used to carry information, which can be either discrete or continuous, each presenting unique advantages in either performance, feasibility for near-term implementation, and compatibility with existing telecommunications architectures. Recently, hybrid QKD protocols have been introduced to leverage advantages from both categories. In this work we provide a rigorous security proof for a protocol introduced by Qi in 2021, where information is encoded in discrete variables as in the widespread Bennett Brassard 1984 protocol but decoded continuously via heterodyne detection. Security proofs for hybrid protocols inherit the same challenges associated with continuous-variable protocols due to unbounded dimensions. Here we successfully address these challenges by exploiting symmetry. Our approach enables truncation of the Hilbert space with precise control of the approximation errors and lead to a tight, semi-analytical expression for the asymptotic key rate under collective attacks. As concrete examples, we apply our theory to compute the key rates under passive attacks, linear loss, and Gaussian noise.

quant-ph↗

Experimental subdiffraction source discrimination enabled by spatial demultiplexing and single-photon detectors

We experimentally demonstrate a universal, parameter-independent test for asymmetric source discrimination below the diffraction limit. The test exploits spatial-mode demultiplexing (SPADE) and single-photon detection to discriminate the presence of a faint source in the vicinity of a brighter one. We develop a complete theory of the test in the presence of arbitrary modal crosstalk and show that the crosstalk-affected test retains a favorable error exponent over diffraction-limited direct imaging, provided that the crosstalk remains below a threshold $C_{\mathrm{max}}\simeq 1/9$. This establishes a quantitative condition under which the subdiffraction advantage of SPADE survives realistic imperfections. We experimentally test the scheme for a range of source separations and intensity ratios, without requiring prior knowledge of these parameters. The measured false-negative error rates are in good agreement with the theoretical model and systematically outperform direct imaging. These results demonstrate that SPADE provides a robust and experimentally accessible advantage for asymmetric subdiffraction hypothesis testing in the photon-starved regime, even in the presence of realistic modal crosstalk.

quant-ph↗

Finite-size resource scaling for learning quantum phase transitions with fidelity-based support vector machines

Quantum kernels offer a valid procedure for learning quantum phase transitions on quantum processing devices, yet issues on the scalability of the learning strategy in connection with the symmetry of the critical model have not been clarified. We derive a link between model symmetry and fidelity-kernel resource scaling. We quantify the measurement resources required to estimate fidelity-based quantum kernels for many-body ground states while preserving the structure of the resulting Gram matrix under finite-shot sampling. Crucially, we show that increasing symmetry in the underlying spin model systematically amplifies these shot requirements. Moving from the $\mathbb{Z}_2$-symmetric Ising/XY regimes to the $U(1)$-symmetric XX (and XXZ) regimes leads to stronger kernel concentration and therefore substantially larger shot costs under the same bounds. We consider a tunable one-dimensional spin-$\tfrac{1}{2}$ Hamiltonian spanning the transverse-field Ising, XY, XX, and XXZ limits, and define the kernel as the ground-state fidelity. Kernel entries are estimated using a SWAP-test estimator with $S$ shots, and we adapt the ensemble spread and concentration-avoidance shot bounds to obtain practical shot requirements in terms of the interquartile range of kernel values and a representative kernel magnitude. For the free-fermion XY/XX family, we use the closed-form Bogoliubov-angle fidelity, while for the interacting XXZ chain we compute fidelities by exact diagonalization and benchmark shot-noise effects. Our symmetry-aware bounds provide a pragmatic procedure for physics-informed quantum machine learning.

quant-ph↗

Recovery-Free CHSH Nonlocality with Particle Loss

Can CHSH nonlocality survive particle loss without applying an explicit recovery operation? In principle, any deterministic recovery can be absorbed into the measurement. Operationally, we show that the answer depends on the allowed measurements on the lossy system. We distinguish flagged erasure, in which each lost particle leaves a detectable record, from unflagged deletion, in which no such record remains. For flagged erasure and survival probability $η>1/2$, using known quantum-capacity results, we show measurements that asymptotically approach the quantum maximum $2\sqrt2$. In contrast, for $η\le1/2$, CHSH violation is impossible for both loss models. Then, we construct explicit recovery-free protocols using permutation-invariant encodings built from $n$-qubit Dicke states $| D_N^n\rangle$ and measurements on the surviving particles. A one-excitation $(N=1)$ encoding violates CHSH for $η>1/\sqrt{2}$. Increasing the excitation number $N$ yields a family of protocols that violates CHSH for $η>η_G=(\sqrt{5}-1)/2$, with the asymptotic golden ratio approached as $N \to \infty$. Finally, we present a sparse-deletion binomial PI protocol that guarantees CHSH violation for up to $O(\sqrt n)$ deletion errors. Our results distinguish fundamental limits imposed by loss from those set by explicit measurements without recovery.

quant-ph↗

Distributed Phase Sensing with Multiphoton States in Optical Interferometry

We investigate interferometric phase-estimation using separable photon inputs that evolve into number-path entangled states through linear optical networks, followed by photon-number-resolving detection. A simple analytical expression for the classical Fisher information at zero phase is derived for arbitrary $N$-photon states distributed across 2$N$ optical modes, partitioned into phase-encoding and reference blocks. Among all possible photon distributions between these blocks, the balanced configuration maximizes the phase sensitivity for every photon number $N$ and uniquely exhibits a phase-independent response. The achievable sensitivity degrades monotonically with increasing asymmetry in the photon distribution. We further investigate the robustness of the protocol in the presence of realistic photon loss and extend the analysis to distributed architectures with multiple receivers. In the low photon-flux regime, vacuum fluctuations fundamentally limit local quadrature measurements, whereas nonlocal photon-number-resolving measurements exploit multiphoton interference to mitigate loss-induced sensitivity degradation. Together, these results establish a scalable framework for quantum-enhanced distributed multimode metrology.

quant-ph↗

Finite-size secret-key rates of discrete modulation continuous-variable quantum key distribution under Gaussian attacks

Quantum conditional entropies play a fundamental role in quantum information theory. In quantum key distribution, they are exploited to obtain reliable lower bounds on the secret-key rates in the finite-size regime, against collective attacks and coherent attacks under suitable assumptions. Here we consider continuous-variable communication protocols, where the sender Alice encodes information using a discrete modulation of phase-shifted coherent states, and the receiver Bob decodes by homodyne or heterodyne detection. We compute the Petz-Rényi and sandwiched Rényi conditional entropies associated with these setups, assuming either a passive eavesdropper or one that injects thermal photons into the channel, who gathers the quantum information leaked through a lossy communication line of known or bounded transmittance. Whereas our results do not directly provide reliable key-rate estimates, they do represent useful ball-park figures. We obtain analytical or semi-analytical expressions that do not require intensive numerical calculations. These expressions serve as bounds on the key rates that may be tight in certain scenarios. We compare different estimates, including known bounds that have already appeared in the literature and new bounds. The latter are found to be tighter for very short block sizes.

quant-ph↗

Finite-size security of QKD: comparison of three proof techniques

We compare three proof techniques for composable finite-size security of quantum key distribution under collective attacks, with emphasis on how the resulting secret-key rates behave at practically relevant block lengths. As a benchmark, we consider the BB84 protocol and evaluate finite-size key-rate estimates obtained from entropic uncertainty relations (EUR), from the asymptotic equipartition property (AEP), and from a direct finite-block analysis based on the conditional min-entropy, which we refer to as the finite-size min-entropy (FME) approach. For BB84 we show that the EUR-based bound provides the most favorable performance across the considered parameter range, while the AEP bound is asymptotically tight but can become overly pessimistic at moderate and small block sizes, where it may fail to certify a positive key. The FME approach remains effective in this small-block regime, yielding nonzero rates in situations where the AEP estimate vanishes, although it is not asymptotically optimal for BB84. These results motivate the use of FME-type analyses for continuous-variable protocols in settings where tight EUR-based bounds are unavailable, notably for coherent-state schemes where current finite-size analyses typically rely on AEP-style corrections.

quant-ph↗

High-Precision Measurement of Time Delay with Frequency-Resolved Hong-Ou-Mandel Interference of Weak Coherent States

We demonstrate a scheme for high-precision measurements of time delay based on frequency-resolved Hong-Ou-Mandel (HOM) interference. Our approach is applied to weak coherent states and exploits an array of single-photon avalanche diodes (SPADs). Unlike conventional HOM experiments, our setup enables high-precision measurements producing an uncertainty per coincidence of about $\sim 10$ ps even for photons separated by delays up to $\sim 4$ ps so much greater than their coherence time where ordinary non-resolved HOM fails. This result confirms our newly developed theoretical predictions that consider, differently from previous theoretical results, a finite frequency resolution in the detection. We compare the performance of this scheme against the conventional non-resolved case. Experimental data align well with the predictions of quantum estimation theory, demonstrating a significant reduction in the uncertainty. Due to the physics of the frequency-resolved HOM effect, the gain in precision is particularly high when the estimated time delay is much longer than the coherence time.

quant-ph↗

Momentum-resolved two photon interference of weak coherent states

We demonstrate an experimental scheme for high-precision position measurements based on transverse-momentum-resolved two-photon interferometry with independent photons and single photon avalanche diode (SPAD) arrays. Our scheme extends the operative range of Hong-Ou-Mandel interferometry beyond its intrinsic constraints due to photons indistinguishability, paving the way to applications in high-resolution imaging. We assess the experimental results against the ultimate precision bounds as determined by quantum estimation theory. Our experiment ultimately proves that transverse-momentum resolved measurements of fourth-order correlations in the fields can be employed to overcome spatial distinguishability between independent photons. The relevance of our results extends beyond sensing and imaging towards quantum information processing, as we show that partial photon distinguishability and entanglement impurity are not necessarily a nuisance in a technique that relies on two-photon interference.

quant-ph↗

Achieving quantum-limited sub-Rayleigh identification of incoherent sources with arbitrary intensities

The Rayleigh diffraction limit imposes a fundamental restriction on the resolution of direct imaging systems, hindering the identification of incoherent optical sources, such as celestial bodies in astronomy and fluorophores in bioimaging. Recent advances in quantum sensing have shown that this limit can be circumvented through spatial demultiplexing (SPADE) and photon detection, i.e. a semi-classical detection strategy. However, the general optimality for arbitrary intensity distributions and bright sources remains unproven. In this work, we develop a general model for incoherent light with arbitrary intensity undergoing diffraction. We employ this framework to compute the quantum Chernoff exponent for generic incoherent-source discrimination problems, focusing on the sub-diffraction regime. We show that, surprisingly, SPADE measurements saturate the quantum Chernoff bound only when certain compatibility conditions are met. These findings suggest that collective measurements may actually be needed to achieve the ultimate quantum Chernoff bound for the discrimination of specific incoherent sources. For the fully general case, our analysis can still be used to find the best SPADE configurations, generally achieved through a rotation of the SPADE interferometer that depends on the discrimination task. We also simulated the efficiency of a simplified Bayesian test that we developed for this identification task and show that the saturation of the Chernoff bound is already achieved for a finite number of repetitions $N\leqslant 5000$. Our results advance the theory of quantum-limited optical discrimination, with possible applications in diagnostics, automated image interpretation, and galaxy identification.

quant-ph↗

Dynamical cluster-based strategy for improving tensor network algorithms in quantum circuit simulations

We optimize matrix-product state-based algorithms for simulating quantum circuits with finite fidelity, specifically the time-evolving block decimation (TEBD) and the density-matrix renormalization group (DMRG) algorithms, by exploiting the irregular arrangement of entangling operations in circuits. We introduce a variation of the standard TEBD algorithm, we termed "cluster-TEBD", which dynamically arranges qubits into entanglement clusters, enabling the exact contraction of multiple circuit layers in a single time step. Moreover, we enhance the DMRG algorithm by introducing an adaptive protocol, which analyzes the entanglement distribution within each circuit section to be contracted, dynamically adjusting the qubit grouping at each iteration. We analyze the performances of these enhanced algorithms in simulating both stabilizer and nonstabilizer random-structured quantum circuits, with up to 1000 qubits and 100 layers of Clifford and non-Clifford gates, and in simulating Shor's quantum algorithm with up to hundreds of thousands of layers. Our findings show that, even with reasonable computational resources per task, cluster-based approaches can significantly speed up simulations of large-sized quantum circuits and improve the fidelity of the final states.

quant-ph↗

Optimal and robust error filtration for quantum information processing

Error filtration is a hardware scheme that mitigates noise by exploiting auxiliary qubits and entangling gates. Although both signal and ancillas are subject to local noise, constructive interference(and in some cases post-selection) allows us to reduce the noise level in the signal qubit. Here we determine the optimal entangling unitary gates that make the qubits interfere most effectively,starting from a set of universal gates and proceeding by optimizing suitable functionals by gradient-descent or stochastic approximation. We examine how our optimized scheme behaves under imperfect implementation, where ancillary qubits may be noisy or subject to cross-talk. Even with these imperfections, we find that adding more ancillary qubits helps in protecting quantum information . We benchmark our approach against figures of merit that correspond to different applications, including entanglement fidelity, quantum Fisher information (for applications in quantum sensing),and CHSH value (for cryptographic applications), with one, two, and three ancillary qubits. With one and two ancillas we also provide analytical explicit expressions from an ansatz for the optimal unitary. We also compare our method with the recently introduced Superposed Quantum Error Mitigation (SQEM) scheme based on superposition of causal orders, and show that, for a wide range of noise strengths, our approach may outperform SQEM in terms of effectiveness and robustness.

quant-ph↗

Quantum-Optimal Frequency Estimation of Stochastic AC Fields

Resolving frequencies in a time-dependent field is classically limited by the measurement bandwidth. Using tools from quantum metrology and quantum control may overcome this limit, yet the full advantage afforded by entanglement so far remains elusive. Here we map the problem of frequency measurement to that of estimating a global dephasing quantum channel. In this way, we determine the ultimate quantum limits of {frequency estimation in stochastic AC} sensing. We find exact {quantum Fisher information bounds} for estimating frequency and frequency differences of stochastic fields. In particular, given two close signals with frequency separation $ω_r$, we find that the quantum Fisher information (QFI) for the separation estimation is approximately $2/ω_r^2$, {i.e.}~\emph{inversely} proportional to the separation parameter. The bounds are achievable in certain regimes by superpositions of Dicke states. GHZ states are suboptimal but improve precision over unentangled states, achieving Heisenberg scaling in the low-bandwidth limit. This work establishes a robust framework for stochastic AC signal sensing that can be extended to arbitrary time-dependent and stochastic fields.

quant-ph↗

Percolation thresholds and connectivity in quantum networks

We study entanglement percolation in qubit-based planar quantum network models of arbitrary topology, where neighboring nodes are initially connected by pure states with quenched disorder in their entanglement. To address this, we develop a physics-informed heuristic algorithm designed to find a sequence of entanglement swapping and distillation operations to connect any pair of distant nodes. The algorithm combines locally optimal percolation strategies between nodes at a maximum distance of one swapping operation. If this fails to produce a maximally entangled state, it looks for alternative paths surrounding intermediate states within the process. We analytically find and numerically verify thresholds in quantum percolation, which depend on the initial network configuration and entanglement, and are associated with specific percolation strategies. We classify these strategies based on the connectivity, a quantity that relates the entanglement in the final state and the level of integrity of the network at the end of the process. We find distinct regimes of quantum percolation, which are clearly separated by the percolation thresholds of the employed strategies and vastly vary according to the network topology.

quant-ph↗

Compared analysis of DInSAR data from ascending and descending orbits of Sentinel-1: the Cazzaso case study

Differential SAR interferometry (DInSAR), by providing displacement time series over coherent objects on the Earth's surface (persistent scatterers), allows to analyze wide areas, identify ground displacements, and study their evolution at large times. In this work we implement an innovative approach that relies exclusively on line-of-sight displacement time series, applicable to cases of correlated persistent-scatterer displacements. We identify the locus of the final positions of the persistent scatterers and automatically calculate the lower bound of the magnitude of the potential three-dimensional displacements. We present the results obtained by using Sentinel-1 data for investigating the ground stability of the hilly village Cazzaso located in the Italian Alps (Friuli Venezia Giulia region) in an area affected by an active landslide. SAR datasets acquired by Sentinel-1 from both ascending and descending orbits were processed using the SPINUA algorithm. Displacement time series were analysed in order to solve phase unwrapping issues and displacement field calculation.

physics.geo-ph↗