Estimation of work at the quantum scale remains a central challenge in quantum thermodynamics. Canonical approaches have a fundamental drawback: they assume classical control of the quantum system, as implicit in a time-dependent Hamiltonian. Yet the energetic cost of implementing this control is omitted and can exceed the system's energy scale by orders of magnitude, calling into question the operational significance of work values. We address this by autonomizing the controlled energy transfer, embedding the driven dynamics into an energy-conserving evolution on a larger quantum system. Work is then unambiguously identified with the energy transferred between the two systems, singling out a unique observable on the driven system: the work operator. We show that the quantum work operator evades a no-go theorem by deriving a quantum fluctuation theorem that recovers the Jarzynski equality in classical scenarios. We incorporate imperfect control and quantify corrections to work statistics. Extending the framework to open quantum systems, we obtain an operatorial first law in which work, heat, and internal-energy changes are represented by distinct operators on the reduced system. Together, these results settle the long-standing debate over whether work is a quantum observable: it is, once the controlling system is included in the description rather than treated as external.

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 {}, Phys. Rev. Lett. {}, 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.

Bell non-locality is a fundamental feature of quantum mechanics related to strong correlations between spatially separated physical systems, which cannot be explained by local hidden variable models. Traditionally, the experimental verification of non-locality requires the repetition of many rounds followed by posterior statistical analysis. However, recent work has demonstrated that it is possible to certify non-locality with only a single experimental round, provided that carefully designed non-local games are used. Such games are operational reformulations of Bell inequalities, in which the quantum advantage over classical strategies manifests as a higher success rate. This project aims to find new games which nonlocality can be verified in one (or few) shots, requiring less resources than the already known to the literature.

The Lipkin-Meshkov-Glick (LMG) model is a prototypical model of quantum critical behavior in complex many-body systems. It originated as a model for numeric calculations of interacting fermions and has been experimentally realized as interacting qubits, bosons, and more. In an experimental set up with a large (but far the thermodynamic limit) spin in the LMG model, how could signs of a phase transition be measured?

We will broadly present recent experimental results of universal analog control of a many level transmon qudit and show how this allows us to realize the LMG model in a singly analog qudit. With this experimental platform, we develop a framework using nearly adiabatic ramps and observing Rabi oscillations which allows for the extraction of arbitrary energy gaps in the spin system and thus reconstructs the system’s full spectrum.

We will further present our experimentally motivated theoretical progress in identifying and measuring signatures of the critical properties of the LMG model in this finite size system. Including showing that our framework allows us to observe precursors of the excited state quantum phase transition (ESQPT) that occurs in the LMG mode. Lastly, we present further work towards more general detection of ESQPTs in near term experimental models and metrological significances of these ESQPTs.

This report presents a novel framework to perform Adaptive Mesh Refinement (AMR) in MHD simulations. Quantum-Hierarchical Adaptive Steering (Q-HAS for short) is a hybrid quantum-classical approach. Its main objective is instability detection by minimising a structured Hamiltonian describing plasma anomalies on a coarse-grained flux graph. Each qubit's polar angle theta encodes a classical multi-indicator score for the instability risk of an edge of the graph. A QAOA-based algorithm evaluates a cost function to locate the riskiest regions of the graph. The framework is implemented with a 4th-order finite-difference MHD solver, a Qiskit-based VQA pipeline, and Optuna Bayesian training. Evaluation on four scenarios (N = 256, 8 qubits) gives a 0.66% composite-loss advantage of Q-HAS over classical AMR, statistically significant in favour of Q-HAS only on Kelvin--Helmholtz (p = 0.026) and Harris Tearing (p = 0.001); and statistically significant in favour of classical AMR on Orszag--Tang (OT) and MHD Rotor. Topological attribution reveals +3.8pp and +5.3pp Q-HAS advantages in topology-rich sub-regions of Tearing and OT respectively. On aggregate per-cell decision accuracy, however, classical AMR is correct on 46.6% of cells vs 43.3% for Q-HAS, and achieves lower L2 error vs DNS on all four scenarios. Whether the localised advantage survives at higher Reynolds numbers remains open; the larger-grid (information-cone) route is tested classically and retired in the pre-registered evaluation below.

A pre-registered evaluation then asks whether any such criterion transfers to instability classes it has never seen. It does not. A gradient-boosted surrogate on the same 9 physical features falls from F1=0.986 within distribution to 0.189 under leave-one-scenario-out, well below the classical baseline (0.434) (which is itself a member of the family, and transfers). The hierarchy inverts: the more flexible the learned model, the worse it travels. Widening the neighbourhood improves the fit on known instabilities and actively degrades transfer; the temporal channel is significantly harmful under LOSO (Δ≈−0.23), because hardness persists and leaves anticipation nothing to anticipate. The solver is also not the bottleneck because annealing, QAOA and exact diagonalisation agree, and all lose to the classical score. Whether the localised advantage survives at higher Reynolds numbers remains open; the larger-grid (information-cone) route is tested classically and retired. The closed-loop level is pre-registered but not executed. All circuits run on the Qiskit-Aer simulator.

Work plays a central role in nonequilibrium thermodynamics, yet in quantum mechanics it is not an observable but a process-dependent quantity. This raises a fundamental question: can genuinely nonclassical correlations be revealed using work itself as the measured quantity [1]? Building on the quasiprobabilistic formulation of quantum work introduced in [2], we propose the first Bell test in which correlators are constructed directly from fluctuating work values. Our protocol employs the Margenau–Hill quasidistribution as an operational definition of work. In a bipartite scenario, each party performs local unitary drivings (the Bell inputs) and reconstructs local work statistics via weak measurements, yielding a joint quasiprobability distribution $p_{\text{weak}}(\omega_A, \omega_B)$. From it we define a CHSH-type Bell functional with correlators linear in the work values, $\langle W_A, W_B \rangle$. We show that the non-observable nature of work imposes a fundamental geometric constraint on the accessible correlators: although the underlying systems are qubits admitting maximally incompatible measurements, the effective work observables are fixed combinations of the initial and final energy gaps. As a result, a genuine work-based Bell protocol cannot saturate the Tsirelson bound $2\sqrt{2}$. We derive an explicit upper bound on the Bell functional as a function of the local energy scales, optimizing over the Bell settings for fixed Hamiltonians. This limitation is not a feature of weak measurements nor of a particular implementation, but a structural consequence of how work is defined in quantum thermodynamics: as an energy difference between noncommuting Hamiltonians. In this sense, our result complements earlier no-go theorems on quantum work: the same structural peculiarities that obstruct a classical work distribution also constrain the strength of achievable nonlocal correlations. Our findings uncover a fundamental interplay between quantum thermodynamics and nonlocality, establishing both the possibility and the intrinsic limits of work as a resource for Bell violations.

[1] Martí Perarnau-Llobet et al., No-Go Theorem for the Characterization of Work Fluctuations in Coherent Quantum Systems, Phys. Rev. Lett. 118, 070601 (2017).

[2] M. Lostaglio, Quantum Fluctuation Theorems, Contextuality, and Work Quasiprobabilities, Phys. Rev. Lett. 120, 040602 (2018).

Counting processes provide a fundamental description of stochastic events, ranging from photon detection to the ticks of a clock. A central question in quantum information and timekeeping is how accurately such events can be timed when only finite memory resources—specifically, a d-dimensional system—are available [1]. In this work, we investigate this problem within a general framework of discrete-time quantum counting processes generated by sequential quantum measurements. First, we establish a rigorous finite-memory variance bound obeyed by every classical d-state counting process, proving a conjecture previously established only for d = 2 [2]. We prove this by mapping the discrete process to a continuous-time absorption problem, demonstrating that the ultimate classical limit is governed by the Aldous-Shepp theorem and saturated by an Erlang-type ladder process [3]. Crucially, we show that quantum counting processes can violate this classical benchmark. Through analytical methods and numerical optimization, we identify optimal quantum processes that produce more regular event statistics than any classical process with the same memory size. We show that these optimal quantum counters rely on a coherent rank-one conditioned dynamics that fundamentally improves upon the classical Erlang cascade. Furthermore, we evaluate the continuous-time limit of these processes, unifying our discrete-time framework with the precision-resolution trade-offs of autonomous quantum clocks. Ultimately, our results establish finite-dimensional quantum measurements as a rigorous resource for temporal precision, bridging quantum measurement theory, stochastic processes, and quantum timekeeping.

[1] M. P. Woods et al., PRX Quantum 3, 010319 (2022).

[2] C. Budroni, G. Vitagliano, and M. P. Woods, Phys. Rev. Res. 3, 033051 (2021).

[3] D. Aldous and L. Shepp, Commun. Stat. Stochastic Models 3, 467 (1987).

Microwave-controlled qubits operate at millikelvin temperatures in cryogenic systems, where thermal noise propagating through control lines can induce unwanted qubit transitions and degrade performance. Attenuators suppress this noise and help stabilize the qubits, but they also dissipate microwave power as heat at cryogenic stages, increasing the cooling demand and the associated energy consumption. Here we determine the attenuation configuration that minimizes the cryogenic power cost under the constraint that the qubit error probability per control step remains below a target value. Using a Lagrange-multiplier approach, we derive analytic expressions for the optimal attenuations for an arbitrary set of cooling-stage temperatures and cooling efficiencies, and obtain a corresponding analytic expression for the minimum consumption power. We then illustrate the consequences of these results for realistic cryogenic architectures, including a typical system with six cooling stages.

This work presents a unified theoretical and experimental study of thermodynamic processes [1], focused on a three-qubit quantum refrigerator. We introduce a protocol that leverages ergotropy [2], the maximal work extractable via cyclic (unitary) operations, to maximize energy extraction from a cold subsystem while minimizing work input and implementation complexity [3]. Theoretical analysis reveals that cooling is feasible only within a specific temperature regime governed by the ordering of thermal populations relative to energy eigenvalues.

We implement this protocol on IBM’s superconducting quantum processors [4], tackling challenges in thermal state preparation, circuit synthesis, and noise mitigation through circuit complexity reduction, particularly by minimizing the number of two-qubit entangling gates. Measurements demonstrate a temperature-dependent cooling efficiency that peaks when hot and cold qubit temperatures closely match, validating the protocol’s viability on NISQ hardware.

A critical examination of correlations in multipartite systems shows their dual role [5]: depending on the thermodynamic quantity of interest, they can either support or hinder energy transfer. Yet when correlations become too pronounced, they drive up dissipation and ultimately impair performance.

This work bridges fundamental quantum thermodynamics and near-term quantum computing, delivering optimized unitary designs, real hardware validation, and a nuanced understanding of correlations’ impact on quantum thermal machines.

Quantum information is the branch of physics which aims to exploit quantum mechanics in order to achieve a capability of processing information which goes beyond what can be obtained with classical approaches. The experimental implementations of these processes, which are currently subject to a profound research, rely on systems based on different technologies, as single photons, superconducting devices, trapped ions etc. In recent years the development of advanced lithographic techniques allowed the integration of complex quantum functions on single integrated photonic circuits and, following this research topic, our work has been focused in the design and test of a quantum controlled not gate (CNOT), realized with 220nm silicon waveguides on a silicon oxide substrate, optimized for photons with a wavelength of 1550nm. The aim of this research is to highlight the capabilities of silicon on insulator platforms as an highly effective technology for experimental quantum information, exploring the possibility to reach high fidelity manipulation of quantum states via the exploitation of single photon properties.

Arguably, the clearest manifestation of non-classicality in quantum systems is the phenomenon of Bell-nonlocality. In the device-independent approach, this phenomenon corresponds to the inaccuracy of classical models in describing correlations between measurement results in space-like separated experiments. While quantum theory predicts a well-defined limit on the strength of such correlations, this limitation lacks a clear physical explanation. This motivates the question of whether the quantum boundary on nonlocality can be understood independently of the abstract formalism of the theory, in terms of more transparent constraints. In this work, we investigate this problem from an information-theoretic perspective, using entropic inequalities to bound the set of nonlocal correlations. Going beyond the standard information-causality framework — which limits the amount of accessible information in a specific communication task, assisted by nonlocal resources — we identify different classes of communication tasks that also witness stronger-than-quantum correlations in general communication settings. Moreover, using information inequalities and tools from causality theory, we derive entropic constraints from relevant causal structures that impose stronger bounds than previous approaches. Importantly, our results are independent of specific encoding and decoding strategies.

The Quantum Approximate Optimization Algorithm QAOA [1] is a hybrid variational quantum algorithm designed to produce approximate solutions for combinatorial problems that are difficult to solve using classical algorithms. [2] introduces multi-angle QAOA (MA-QAOA), where each rotational gate is given a separate classical parameter, thus reducing the quantum circuit depth. [3] shows that if an automorphism exists in the graph of the problem, then the number of classical parameters in MA-QAOA can be reduced further. In this study, we find that one can further reduce the number of classical parameters.

[1] Edward Farhi, Jeffrey Goldstone, and Sam Gutmann. A Quantum Approximate Optimization Algorithm, 2014.

[2] Rebekah Herrman, Phillip C. Lotshaw, James Ostrowski, Travis S. Humble, and George Siopsis. Multi-Angle Quantum Approximate Optimization Algorithm. Scientific Reports, 12(1):6781, 2022.

[3] Kaiyan Shi, Rebekah Herrman, Ruslan Shaydulin, Shouvanik Chakrabarti, Marco Pistoia, and Jeffrey Larson. Multi-Angle QAOA Does Not Always Need All Its Angles. In 2022 IEEE/ACM 7th Symposium on Edge Computing (SEC), pages 414–419, 2022

Recent theoretical and experimental progress suggest that quantum metrology techniques, exploiting nonclassical resources, may provide viable routes to test the interplay between gravity and matter. Taking issue from such advances, we investigate the use of Quantum Hypothesis Testing (QHT) and Quantum Fisher Information (QFI) to distinguish between two formulations of the gravitational interaction — classical Newtonian law of gravitation and the Modified Newtonian dynamics (MOND) framework — in an optomechanical platform where two massive mechanical oscillators interact. We highlight both the feasibility and limitations of employing QHT in this context, outlining conditions under which MOND-induced effects might be distinguished from Newtonian predictions. Our investigation suggests that quantum information-inspired techniques could play a significant role in future experiments designed to test alternative models of gravity beyond the classical paradigm.

The emergence of thermalization and statistical-mechanical predictions in closed many-body quantum systems has been largely studied through the ansatz of Eigenstate Thermalization Hypothesis (ETH). This framework links the generic behavior of chaotic systems to predictions of Random Matrix Theory, providing insight into the average and fluctuations of time-evolved observables. Motivated by the Random Matrix behavior of chaotic systems beyond low-order correlation functions, recent work has explored toy models such as Haar-random circuits, the characterization of entanglement in eigenstates or out-of-equilibrium states, and operator growth. A key phenomenon in this context is the Heisenberg evolution of the entanglement of an initially localized observable. This quantity, known as Local Operator Entanglement (LOE), presents a linear evolution with time for chaotic dynamics, being expected to satisfy volume laws at late times in finite systems. This behavior has manifested in numerical simulations of particular chaotic Hamiltonians and analytically under the description of late-time chaotic dynamics with pseudorandom quantum circuits. In this work, we provide further analytical and numerical foundations for the role of Random Matrix Theory in the emergence of volume laws in continuous time evolution of the LOE. Using the typicality of Hamiltonian eigenstates within small energy windows—where they can be treated as uniformly distributed—as our working hypothesis, we define the microcanonical LOE by restricting the dynamics to eigenstates within a chosen window. Starting from a basis of the corresponding Krylov subspace, we obtain an analytical formula for the average late-time value of the LOE, expressed in terms of the total Hilbert space dimension and the dimension of the subsystem over which the entanglement entropy is computed. We show that for small subsystem dimensions the analytical values and Exact Diagonalization simulations almost overlap and that the error does not strongly depend on the window size. Our work highlights the role of the Random Matrix Theory in the operator entanglement volume laws, characterizing its universal behavior in chaotic systems.

We study quantum state and entanglement transfer in quantum channels operated sequentially without resetting, leading to memory effects. While quantum state transfer (QST) is typically analyzed under the assumption of memoryless channels, a more realistic implementation involves continuous use, where residual correlations between transmissions accumulate. For U(1)-symmetric channels, we derive a general analytical expression for the average fidelity at the nth use, enabling a systematic characterization of performance degradation over repeated transmissions. Applying our framework to a perfect state transfer channel with imperfect readout timing, we show that even small timing errors induce memory effects that significantly reduce fidelity in subsequent uses. We further show that these effects strongly impact entanglement transfer, leading to a rapid deterioration of distributed quantum correlations. We also present bounds on the channel's quantum capacity, showing that this too degrades with each use. These findings highlight the importance of accounting for memory effects in realistic quantum communication protocols.

Open quantum system dynamics is efficiently described by the quantum master equation formalism. Therein, quantum master equations in Lindblad form constitute an important subclass describing Markovian dynamics. When an open quantum system is in an out-of-equilibrium state, an exchange of particles between the open system and reservoirs takes place yielding to a non-zero average net current and associated current fluctuations, which can be characterised with the quantum jump formalism for quantum master equations expressed in Lindblad form. However, a large class of quantum master equations cannot be described by Lindblad dynamics. Here we assess the validity and the effectiveness of the quantum jump formalism when the dissipators in the quantum master equation describe a non-additive, open quantum system dynamics. We find that an additive unravelling of the non-additive quantum master equation does not generate a completely-positive dynamics in the scenario of perfect jump detection. Nevertheless, allowing for an imperfect jump detection scenario, we find that an additive unravelling is possible that reproduces the current and the fluctuations obtained via the Landauer-Büttiker formalism.

We consider a multiplexed quantum repeater that distributes entanglement between two end nodes, where multiplexing is achieved through optical integration of many quantum chips. Each chip hosts an optically addressable communication qubit and a separate memory qubit. The communication qubit serves as an entanglement generation interface between different quantum chips, and the memory qubit can be used to store entanglement. The quantum chips on the repeater node are interconnected using a reconfigurable router. By reconfiguring the router, it is possible to dynamically assign quantum chips for entanglement generation with either of the end nodes. Previous work has shown that in the limit of many quantum chips, a dynamic multiplexing policy, in which chips are assigned to end nodes based on the current entanglement in the chain, asymptotically yields the same rate of end-to-end entanglement generation as a policy in which the assignment of quantum chips to the end nodes is fixed. However, the fidelity of the end-to-end entanglement was not considered. We show that a dynamic multiplexing policy can lead to a significant improvement in fidelity over a fixed policy, if on average less than one link is generated per end-to-end communication cycle. This makes dynamic multiplexing with a many-quantum-chip repeater especially relevant for the development of near-term quantum networks.

Quantum cooling counteracts the spontaneous thermalization of high-fidelity qubits by driving their state population back towards an ideal pure ground state. To cool a single-qubit target, one typically couples it to an n-qubit machine via joint unitary operations. How does the machine’s structure affect our cooling ability and protocol complexity? First, we derive inequalities based on the machine’s energy-level structure, establishing a cooling bound analogous to Carnot’s theorem. Second, we map the protocol to a MWPM problem on an (n+1)-D hypercube. This allows us to optimize cooling circuit complexity and provides a novel framework for quantum algorithmic cooling.

Combinatorial optimization problems can be naturally formulated as Quadratic Unconstrained Binary Optimization (QUBO) models and subsequently mapped onto Ising Hamiltonians, making them suitable for Quantum Annealing (QA) techniques. However, the practical implementation of these approaches is hindered by the limited connectivity of current quantum hardware and by the difficulty of accurately simulating the annealing dynamics for large-scale instances. In this work, a Direct Current Optimal Power Flow (DC-OPF) problem is formulated as a QUBO model and studied from both the embedding and annealing perspectives. Two embedding strategies, namely minorminer and the Triangular Architecture, are analyzed in terms of their physical qubit requirements. In addition, two simulation approaches, QiliSim and a variational method, are employed to investigate the annealing dynamics. To improve the annealing performance, catalyst terms previously developed for systems without local fields are extended to account for the local field contributions naturally present in the DC-OPF formulation. These additional interactions, inspired by diagonal augmentations of the Quantum Approximate Optimization Algorithm (QAOA), act as catalysts by enhancing the ground state fidelity with respect to standard quantum annealing. Different catalyst configurations are studied and their effect on the probability of obtaining low energy states is analyzed. Furthermore, qubit reduction strategies based on thresholding and spectral decomposition are investigated in order to decrease the physical resources required for implementation. The results show that the inclusion of local field contributions in the catalyst term, leads to a significant improvement in the annealing performance and that suitable thresholding techniques allow the number of physical qubits to be reduced by approximately a factor of two while preserving the quality of the obtained solutions. These findings provide further support for the catalyst framework and contribute to bringing realistic optimization problems closer to current quantum hardware.

Superconducting circuits containing Josephson junctions are important for both fundamental research and applications ranging from precision metrology to quantum technologies. Even relatively simple circuits exhibit rich nonlinear dynamics, including phase slips and switching between different dynamical regimes [1]. Such systems are often well described by the resistively and capacitively shunted junction (RCSJ) model and its extensions. However, reliable models of more complex and noisy circuits are often difficult to formulate. In this work, we employ neural stochastic differential equations [2], a machine-learning approach that combines known physical models with neural networks to describe such systems. The neural network learns the missing contributions to the dynamics, including stochastic effects, while preserving consistency with the underlying physics. The approach is sufficiently flexible to describe different circuit architectures and may provide a general framework for modeling complex superconducting circuits. As a proof of concept, we apply the method to RCSJ models with different Josephson potentials [3,4] and a more complex circuit.

References:

[1] M. Žonda, W. Belzig, T. Novotný, Phys. Rev. B 91 (13), 134305 (2015)

[2] X. Li, T.-K. L. Wong, R. T. Q. Chen, D. Duvenaud, Proc. Mach. Learn. Res. 108, 3870-3882 (2020)

[3] M. Žonda, W. Belzig, E. Goldobin, T. Novotný, Phys. Rev. B 110 (5), 054306 (2024)

[4] F. Dominguez, F. Hassler, G. Platero, Phys. Rev. B 86, 140503(R) (2012)

We formulate a canonical quantization of equilibrium thermodynamics by applying Dirac's theory of constrained systems. Thermodynamic variables are treated as conjugate pairs of coordinates and momenta, allowing extensive and intensive quantities to be promoted to operators in a Hilbert space. The formalism is applied to the ideal gas, the van der Waals gas, and the photon gas, illustrating both first- and second-class quantization procedures. For the ideal gas, a Schrödinger-like equation emerges in which entropy plays the role of time, and the wave function acquires a phase determined by the internal energy. A pseudo-Hermitian framework restores Hermiticity of the temperature operator and establishes the equivalence among constraint realizations. The approach naturally leads to thermodynamic uncertainty relations and suggests extensions to quantum and topological phase transitions, as well as black-hole and nonequilibrium thermodynamics.

Transport-induced self-sustained oscillations in electromechanical systems convert a static electrochemical bias into robust, autonomous oscillatory motion in the absence of any external periodic drive. However, an exact description of such self-oscillations remains challenging in nanoscale electromechanical devices featuring a simultaneously large bosonic Hilbert space, strong interactions, and structured fermionic leads. We formulate a tensor-network framework that combines a binary representation of the vibrational mode with mesoscopic reservoir embeddings that enable controlled access to the self-oscillatory steady states and relevant transport observables without explicit real-time propagation. We demonstrate the emergence of mechanical self-oscillations across a broad set of operating conditions, in which strong electromechanical backaction, nonadiabatic oscillator dynamics, and energy-dependent electronic tunneling processes compete. Furthermore, we observe that for both slow and fast vibrating mechanical modes, suppressed vibrational occupation fluctuations in the self-oscillation window along the electromechanical coupling strength sweep are preceded by a peak in the occupation fluctuations. Collectively, we explore how both intrinsic system properties and environmental parameters govern such autonomous oscillations over a broad range of operating conditions. The broad applicability of our framework will facilitate further extension of the method to more complex and experimentally relevant scenarios to study the thermodynamics of quantum autonomous devices with potential applications in quantum sensing, timekeeping, information processing, and energy conversion.

We investigate the open-system dynamics of a micromaser quantum battery in the ultrastrong-coupling (USC) regime. The battery consists of a quantized harmonic mode sequentially interacting, via the Rabi Hamiltonian, with a stream of qubits acting as chargers. USC enhances the charging speed but also induces unbounded energy growth and highly mixed cavity states. Dissipation suppresses this behavior, driving the system to a steady state with finite energy and ergotropy. Using optimal control theory, we show that the interplay between USC and dissipation enhances both charging performance and long-term stability against losses.

Quantum computing has emerged as a promising paradigm for addressing combinatorial optimization problems that are challenging for classical algorithms. A common approach for executing such problems on quantum processing units (QPUs) is to reformulate them as Quadratic Unconstrained Binary Optimization (QUBO) models. While objective functions can often be represented efficiently in QUBO form, the incorporation of constraints remains a significant challenge. In particular, inequality constraints typically require the introduction of auxiliary binary variables and slack-variable encodings, increasing the overall number of qubits and enlarging the search space. These overheads become more challenging on current Noisy Intermediate-Scale Quantum (NISQ) devices, where qubit resources are limited.

In this work, we investigate a novel QUBO encoding strategy for inequality constraints based on a class of tunable exponential penalty functions. The proposed approach aims to reduce the auxiliary-variable overhead associated with traditional encodings while maintaining the structure of the original optimization problem. We hypothesize that approximating the exponential penalty through its Taylor expansion up to the second-order term preserves the quadratic structure required by QUBO formulations, enabling implementation without increasing qubit requirements.

Furthermore, we analyze the effect of penalty magnitudes and tuning parameters on constraint satisfaction and solution quality. Through empirical evaluation on representative combinatorial optimization instances, we compare the behavior of the exponential penalty formulation with standard polynomial(quadratic) penalty approaches. Our results demonstrate that tuned exponential penalties can effectively encode inequality constraints while achieving competitive solution quality with reduced qubit overhead, making them a promising alternative for resource-constrained quantum optimization on NISQ hardware.

The Selected Quantum Doubles (SQD) methods are among the most reliable approaches for quantum computation, capable of achieving high-accuracy simulations of electronic structure problems. Despite their strengths, SQD methods encounter challenges related to the selection of the initial state and the efficient management of extensive configuration spaces. This is particularly due to the presence of numerous configurations with low significance, which can limit scalability and accuracy. To address these issues, we have developed a perturbative quantum method capable of generating an initial state that is both accurate and flexible, alongside a procedure to manage the scalability of the method. Regarding the first objective, a parameterized Unitary Coupled Cluster Doubles (UCCD) approach has been developed, which generates a wavefunction comparable in quality to the Coupled Cluster Doubles (CCD) reference state. At this juncture, it is possible to augment this parameterized UCCD state to incorporate the principal single excited configurations. This augmentation step is essential to ensure that the initial state possesses sufficient flexibility to account for configurations that are combinations of single and double excited configurations. The second objective is the enhancement of the methodology's scalability. This objective is achieved by first identifying, through a perturbative approach, the most significant higher-order configurations. This allows for a smaller configurational CI Hamiltonian. Conversely, the second set of configurations will not be included in the Hamiltonian CI but will be evaluated solely within a perturbative framework using second-order perturbation theory (PT2). This combined approach leverages the strengths of both variational and perturbative methods, enabling the simulation to closely approximate Full Configuration Interaction (FCI) results while operating within a significantly reduced configurational space. This methodology resembles the most common Selected Configurations Interaction plus Perturbation methodology, while distinguishing itself through an improved selection process, which will be further refined in the future. The method will be tested on a highly correlated molecular system (linear chain of 8 Hydrogen) in minimal basis set. The molecule has been optimized in PySCF with MP2 method and def2-svp basis set.

Schrödinger bridge problems seek the most likely stochastic evolution between prescribed initial and final probability distributions under given dynamics. They form a class of stochastic optimal control problems and have recently attracted interest in generative modelling.

In high dimensions, dense discretizations quickly become infeasible due to exponential scaling. Tensor network methods, originally developed in numerical quantum mechanics and condensed matter physics to represent many-body wavefunctions, provide structured low-rank representations of high-dimensional objects. In our work, we use the quantics tensor train (QTT) representation to compress the probability densities, Schrödinger factors, and differential operators appearing in the bridge problem.

We develop a QTT-based solver that combines factorized iterations with tensorized heat propagation and low-rank function reconstruction. The method exploits low-rank structure on exponentially large grids, enabling computations that would be infeasible by conventional grid-based approaches.

Structured baths can reshape transport fluctuations in mesoscopic quantum devices, yet a predictive criterion for when this enhances precision has been lacking. We propose a route towards such precision advantages by utilizing bath memory in coherent fermionic transport through a noninteracting quantum-dot chain. Using the Landauer-Büttiker formalism, we derive a dual impedance-matching condition that synchronizes the conductor mode splitting, boundary dissipation, and bath bandwidth, and sustains constructive multimode interference across the transmission window. The analytical predictions for the optimal bath bandwidths show excellent agreement with exact nonequilibrium Green's function calculations of the transport for Lorentzian, Gaussian, and Newns spectral densities. The prescription yields an optimal bath bandwidth at which the current Fano factor is minimized and the thermodynamic and kinetic precision coefficients are simultaneously enhanced beyond their Markovian limits. The alignment of the optimal precision regime with the experimentally accessible current Fano factor minimum thus provides a practical strategy for designing precision-enhanced transport in mesoscopic platforms such as semiconductor quantum-dot arrays and ultracold fermionic channels.

Nonstabilizerness is a fundamental resource for quantum computation, enabling quantum algorithms to surpass classical capabilities. Despite its importance, characterizing this resource remains challenging due to the intricate geometry of stabilizer polytopes and the difficulty of simulating nonstabilizer states. In this work, through device-independent considerations, we reveal an unexpected connection between nonstabilizerness and Bell inequalities. Although certain stabilizer states can already achieve maximal violations of specific Bell inequalities, we demonstrate that appropriately constructed Bell inequalities can nevertheless serve as witnesses of nonstabilizerness, revealing when a state lies beyond the stabilizer set. This result bridges two key quantum resources, uncovering a novel relationship between the device-independent framework and resource-theoretic properties of quantum computation.

In finance, assessing the creditworthiness of loan applicants requires lenders to cluster borrowers using rating scales. Financial institutions must define the scales of applicants in compliance with strict institutional constraints, resulting in solving a complex combinatorial constrained optimization problem. This poster outlines how we solved this problem using a Quadratic Unconstrained Binary Optimization (QUBO) model, a formulation suitable for quantum hardware. We validated this approach by testing the proposed formulation with classical heuristics. We then benchmarked the results against a brute-force method to demonstrate consistent solution quality and highlight the framework's suitability for more complex scenarios.

Entanglement is routinely generated in photonic systems, forming the basis of most quantum technological applications. Still, it has not yet been observed in free electron optics, a gap we fill in this work. Nearly a century ago, the first transmission electron microscope was built, representing a leap forward in controlling free electrons and revolutionising the field of microscopy by enabling atomic-resolution imaging and spectroscopy for advanced materials characterisation. Here, we demonstrate the first experimental evidence of entanglement in electron-photon pairs in a transmission electron microscope. I will go into the witnesses behind those experiments, their background and limits, and explain how we showed the hybrid entanglement. Specifically, we apply a witness from photonic quantum optics to the pair state, utilizing a quantitative bound derived from the Heisenberg uncertainty relation between the transversal position and momentum. We retrieve the strength of the correlations between these two variables through coincidence-based imaging. They reveal a strong violation of the bound for classical states: ∆x−2 ∆k+2 ≤ 0.321 ± 0.027 < 1, identifying the entanglement. Hence, our work confirms quantum correlations in free electron-photon systems, bridging the fields of electron microscopy and photonic quantum optics. This poster is based on https://arxiv.org/abs/2504.13163.

QMetro++ is a Python package that provides a set of tools for identifying optimal estimation protocols that maximize quantum Fisher information (QFI). Optimization can be performed for arbitrary configurations of input states, parameter-encoding channels, noise correlations, control operations, and measurements. The use of tensor networks and an iterative see-saw algorithm allows for an efficient optimization even in the regime of a large number of channel uses (𝑁 ≈ 100). Additionally, the package includes implementations of the recently developed methods for computing fundamental upper bounds on QFI, which serve as benchmarks for assessing the optimality of numerical optimization results. All functionalities are wrapped up in a user-friendly interface which enables the definition of strategies at various levels of detail.

Surface roughness is an important quantity to many engineering and precision manufacturing disciplines. In this paper we investigate the problem of estimating the root-mean-square roughness of a sample by passive linear optics. By adopting quantum parameter estimation methods, we determine the ultimate precision limits for estimating spatial moments of a general three-dimensional distribution of incoherent point sources in the sub-diffraction regime. Specializing this result to the axial profile, we show that the information on the first moment (mean height) and standard deviation (roughness) is bounded by a constant. While classical imaging techniques fail to achieve this bound, a quantum inspired imaging technique based on spatial mode demultiplexing is proven to be optimal for estimating the axial standard deviation. This provides a powerful and experimentally accessible route to measuring roughness of nearly smooth surface patches beyond the diffraction limit.

Quantum Imaginary-Time Evolution (QITE) is a promising approach for preparing ground states and studying quantum many-body systems, but its cost grows rapidly with the size of the local operator pool. We introduce Sparse-QITE, a regularized formulation of QITE that promotes compact Pauli generators at each imaginary-time step. In transverse-field Ising model benchmarks, Sparse-QITE preserves QITE-level accuracy while using a much smaller active set of Pauli strings. We also study the evolution of the active support and show that, in the regimes considered, the number of relevant operators can saturate at early imaginary time. This provides information about the relevant active support and suggests that, after discovery, only this active set may need to be measured. Our results point toward reducing both measurement and circuit-depth costs in practical QITE implementations.

Given that quantum computers are naturally suited to simulate the behavior of quantum many-body systems, an immediate question arises: can one formulate \emph{physically motivated} quantum machine learning (QML) tasks that exhibit learning separations? We address this problem by studying the learnability of quantum many-body dynamics from the perspective of probably approximately correct (PAC)-learning. Concretely, we devise a supervised learning problem where the training set consists of specifications of randomized stabilizer probe states, evolution times sampled uniformly at random from a polynomially large time interval $\left[ 0, T \right]$, coupled with expectation values of certain observables evaluated on the resulting time-evolved state under an unknown Hamiltonian. For this learning task, we provide an efficient quantum procedure whose training phase learns the underlying Hamiltonian from short-time training samples, and whose deployment phase combines Hamiltonian simulation with the classical shadows protocol to perform inference on a newly given data point. By contrast, the existence of $\mathcal{O} \left( \mathsf{poly} \left( n \right) \right)$-time instances ensures classical hardness: by embedding a $\mathsf{BQP}$-complete computation into the polynomially long time-dynamics of a low-intersection variant of the Feynman-Kitaev clock Hamiltonian construction, we show that, for a certain family of input distributions, no randomized classical polynomial-time algorithm can fulfill our learning condition, unless $\mathsf{BQP} \subseteq \mathsf{P/poly}$. Furthermore, we show that the classically hard instance maintains quantum learnability. We also give an interpretation of our results in learning-assisted certified quantum simulation. Taken together, our results demonstrate a rigorous learning separation for a natural ML task based on Hamiltonian evolution, while building connections between quantum learning theory, quantum simulation, and QML.

Variational Quantum Eigensolvers (VQEs) depend critically on the geometry of the underlying optimization procedure. Quantum Natural Gradient (QNG) exploits the Fubini–Study geometry of quantum states to accelerate convergence, yet its metric generally does not respect particle-number and spin symmetries. Consequently, geometrically efficient update directions can drive the variational state into unphysical symmetry sectors.

We investigate symmetry leakage as a problem of geometry-constrained quantum optimization and introduce forbidden transition dipole moments (TDMs) as sensitive diagnostics of representation mixing. Unlike conventional expectation-value-based measures, forbidden TDMs are linearly sensitive to the leakage amplitude and reveal residual symmetry contamination that otherwise remains hidden. Using H2 as a benchmark, we systematically compare gradient-level filtering, geometry-aware metric regularization, and adaptive objective shaping within the QNG framework. Our results show that preserving physical fidelity requires not only efficient optimization geometry but also alignment between the optimizer's notion of distance and the representation structure of the quantum system.

The identification of Minimal Cut Sets (MCS) in Fault Trees (FT) is a fundamental task in Probabilistic Safety Assessment (PSA), yet it remains computationally demanding for large-scale systems. While several gate-based quantum approaches have been investigated for this problem, the potential of Quantum Annealing (QA) has not yet been explored in this industrial context. In this work, we study the application of QA to MCS identification by representing FTs as Boolean functions and implementing three existing SAT-to-Quadratic Unconstrained Binary Optimization (QUBO) encoding methods. These encodings are evaluated on industrial FT instances, using a simulated annealer and two D-Wave quantum annealing architectures with different topologies. The results indicate that the encoding based on reusable auxiliary variables achieves the best trade-off between qubit usage and embedding scalability. Furthermore, the Zephyr architecture shows superior embeddability compared with Pegasus, leading to fewer chain breaks and improved performance on larger problem instances. Overall, QA demonstrates an ability to sample diverse valid solutions, making it a promising approach for exploring the solution space associated with MCS problems. Nevertheless, its practical effectiveness is currently constrained by hardware limitations and the finite number of annealing samples, which hinder the exhaustive identification of all solutions for larger instances.

The study of non-classicality is essential to understand the quantum-to-classical transition in physical systems. Recently, a witness of non-classicality has been proposed, linking the ability of a system (``the mediator") to create quantum correlations between two quantum probes with its non-classicality, intended as the existence of at least two non-commuting variables.

Here, we propose a new inequality that quantitatively links the increase in quantum correlations between the probes to a function of the non-commutativity of the mediator's observables. We test the inequality for various degrees of non-classicality of the mediator, from fully quantum to fully classical.

This quantum-to-classical transition is simulated via a phase-flip channel applied to the mediator, inducing an effective reduction of the non-commutativity of its variables. Our results provide a general framework for witnessing non-classicality, assessing the non-classicality of a system via its intrinsic properties, independently of the specific chosen interaction dynamics.

Encoding combinatorial optimization problems into physically meaningful Hamiltonians with tractable energy landscapes forms the foundation of quantum optimization. Numerous works have studied such efficient encodings for the class of Quadratic Unconstrained Binary Optimization (QUBO) problems. However, many real-world tasks are constrained, and handling equality and, in particular, inequality constraints on quantum computers remains a major challenge. We show that including inequality constraints is equivalent to solving a multi-objective optimization. This insight motivates the Multi-Objective Quantum Approximation (MOQA) framework, which approximates the maximum via smaller 𝑝-norms and comes with rigorous performance guarantees. MOQA operates directly at the Hamiltonian level and is compatible with, but not restricted to, ground-state solvers such as quantum adiabatic annealing, the Quantum Approximate Optimization Algorithm (QAOA), or imaginary-time evolution. Moreover, it is not limited to quadratic functions.

We introduce a loss-tolerant measurement-device-independent quantum position-verification (MDI-QPV) protocol secure against adversaries restricted to local operations and simultaneous quantum communication (LOSQC). The protocol has two verifiers encode qubit states using a shared random basis, while the prover returns parity information from a Bell-state measurement or declares a loss. An honest prover succeeds perfectly, with losses naturally arising from photon-number detection and postselection. For the loss-tolerant setting, we give two security proofs. A geometric uncertainty relation yields the optimal LOSQC cheating probability of 3/4, while a complementary entropic-uncertainty-based reduction to BB84-QPV provides a more flexible, though non-tight, argument. We extend the latter approach to arbitrary loss by exploiting a key MDI feature: conditioning on successful Bell-state measurements produces a basis-independent effective state. We also outline extensions to decoy-state implementations, finite-size security, and high-dimensional encodings.

In arXiv:2605.27369 we introduce a new class of quantum channels, the Generalized Multilevel Amplitude Damping (GMAD) channels, to model noise and decoherence effects in a qudit coupled to a thermal environment. Our work extends the Multilevel Amplitude Damping channels to the finite-temperature regime, as well as the GADC beyond the d = 2 case. The degradation of energetic resources under GMADs was investigated by evaluating work functionals and ergotropic capacitances, with particular attention to the coherent and incoherent contributions to ergotropy, for which we introduced new quantifiers. Our analysis revealed several counterintuitive phenomena: the ergotropic capacitance of a GMAD channel is not monotonic in the temperature of the environment; moreover, iterating the map can lead to crossings between ergotropic functionals at different temperatures, indicating the presence of a Markovian Mpemba effect.

Quantum cellular automata (QCA) are locality-preserving automorphisms of quantum spin systems. Haah classified strictly local QCA up to blending equivalence using Brauer groups of invertible subalgebras. This project aims to extend this classification to approximately local QCA, where locality holds only up to a decaying error. Using a near-inclusion theorem, we construct a rounding procedure and define an approximate boundary algebra. We then introduce approximately invertible subalgebras as the natural analogue of invertible subalgebras in this setting. The next step is to study the resulting Brauer group and its relation to the blending group of approximately local QCA.

Quantum thermodynamics has recently emerged as a rich field of both fundamental interest and practical utility [1]. Exactly how classical thermodynamics and irreversibility emerge at large scales from unitary quantum mechanics is the subject of current research. At the microscopic level, coherent interactions can be harnessed to build devices such as quantum heat engines or refrigerators that outperform their classical counterparts.

Trapped ion systems are an excellent platform for quantum thermodynamics; they enjoy long coherence times and demonstrate state preparation, measurement, single- and two-qubit gates at high fidelity [2]. Laser pulses can implement Hamiltonian quenches that manipulate the ions' energy landscape, thereby performing microscopic work instantaneously relative to the thermalization time [3]. Furthermore, ground state laser cooling enables coherent control of the ions' harmonic vibrational modes, which provide an additional degree of freedom and all-to-all connectivity between ions in a chain [1].

We present an individually-addressed chain of Ba ions as an especially effective system for thermodynamic tasks. Its hyperfine structure () and long-lived metastable manifold permit numerous encoding options for clock qubits across 32 ground and metastable levels. This readily enables mid-circuit measurement, which is crucial for re-thermalizing specific parts of the system and implementing two-point measurement protocols [3]. In addition, we have demonstrated coherent manipulation of higher-dimensional qudit encodings within a single ion [4]. As a result, this platform can realize multi-level thermal machines and measure work and heat statistics in regimes inaccessible to classical systems.

[1] S.Campbell et al., Quantum Science and Technology 11 012501 (2026).

[2] M.Foss-Feig et al., Annual Review Condensed Matter Physics 16:145-72 (2025).

[3] O.Onishchenko et al., Nature Communications 15 6794 (2024).

[4] A. Vazquez-Brennan, PhD thesis. Manuscript in preparation (2026).

Nonstabilizerness, commonly called quantum magic, is a necessary resource for universal quantum computation, yet its quantification and classification remain computationally challenging. In this work, we introduce the stabilizer partition function (SPF), viewing the Pauli spectrum as the spectrum of a many-body system and thereby encoding all moments of the Pauli distribution within an efficiently computable framework. The SPF enables the efficient characterization of quantum states across Hilbert space, including highly magical states such as Haar-random and pseudomagical states. We further use the SPF to define stabilizer work, a computable, faithful magic monotone, and apply it to molecules and many-body systems, establishing a thermodynamics-inspired framework for quantum magic.

Understanding and characterizing large-scale analog quantum systems remains a central challenge in quantum many-body physics. Recent studies have shown that randomized measurements assisted by ancillary lattice systems offer a promising route to overcome the few-basis measurement limitation. However, such approaches face two major bottlenecks: the required ancillary system size scales with the target system size, and long timescales are required to achieve ergodicity.

Here, we propose a periodically driven hybrid (lattice-continuum) quantum system as a platform for efficient quantum state characterization by leveraging continuum degrees of freedom to access exponentially large state spaces even in the few-body regime and periodic driving to enhance ergodicity. To evaluate ergodic dynamics in driven continuum systems, we study chaos and thermalization in the Lieb-Liniger model under harmonic-trap modulation in a few-body regime, focusing on energy absorption and delocalization. We find that finite interactions break the regularity of the harmonic system, thereby suppressing energy absorption in the near-resonant regime and enhancing delocalization in the energy eigenbasis, as characterized by the inverse participation ratio. Moreover, we identify a slow-heating regime indicative of prethermalization that emerges only at finite interaction strength. Our results provide insight into nonequilibrium dynamics in interacting quantum systems and establish driven continuum systems as promising ancillary resources for quantum-state characterization.

Bell tests based on homodyne detection are strongly constrained in continuous-variable systems. Can Gottesman–Kitaev–Preskill (GKP) encoding turn homodyne detection into a practical tool for revealing Bell nonlocality? We consider a physically motivated model in which each party performs homodyne detection and digitizes the continuous outcome via a fixed periodic binning, corresponding to logical Pauli measurements. Within this framework, we derive a bipartite no-go: CHSH cannot be violated for Bell-pair states. Moving beyond two parties, we show that finitely squeezed GKP-encoded GHZ and W states nevertheless exhibit strong multipartite nonlocality, violating multipartite Bell inequalities with homodyne-only readout. We quantify the required squeezing thresholds and robustness to loss, providing a route toward homodyne-based Bell tests in continuous-variable systems.

The Mpemba effect has attracted growing interest in both classical and quantum nonequilibrium systems. In this poster, we study coherence-phase control of quantum Mpemba relaxation in a one-dimensional open quantum chain with coherent hopping and symmetric incoherent hopping processes. Under open boundary conditions, the asymmetric dissipation gives rise to Liouvillian skin modes, so that relaxation is strongly affected by the biorthogonal overlap between the initial state and the slow Liouvillian modes. We introduce a boundary-localized coherent initial state with a tunable relative phase and show that this phase changes the projection onto the slowest Liouvillian skin mode. For suitable phases, destructive interference suppresses the slow-mode weight, thereby modifying the relaxation trajectory and shifting the Mpemba crossing time. Our results identify the coherence phase as an effective control parameter for tuning quantum Mpemba relaxation in Liouvillian skin systems through initial-state preparation. Future work will explore how interactions modify this coherence-controlled relaxation mechanism.