https://arxiv.org/api/s+XtiegnP/PHnYhV32gcKOkI+XA 2026-04-10T22:05:08Z 27985 345 15 http://arxiv.org/abs/2508.19537v3 2026-03-17T20:59:28Z

Humans can often predict physical outcomes after only a few observations, a capability known as physical intuition. The mechanisms underlying this efficient learning remain elusive. Here, we introduce a variational learning framework in which small neural networks learn to discover optimal physical states from merely two or three similar examples. Demonstrating across classical and quantum systems including strongly correlated molecules, we show that networks trained this way generalize accurately across wide observation ranges, far beyond the training data. This generalization is explained by a unified theory: it arises when the network approximates a solution manifold where the Euler-Lagrange operator is stationary with respect to observation features. The theory predicts the existence of a critical network size below which robust generalization fails to emerge. Our work establishes variational learning as a principled route to acquiring artificial physical intuition and offers a theoretical perspective for understanding similar capabilities in biological intelligence.

2025-08-27T03:26:57Z Jingruo Peng Shuze Zhu http://arxiv.org/abs/2603.17032v1 2026-03-17T18:13:27Z

We present an efficient Monte Carlo algorithm for the simulation of the two-dimensional Random Field Ising Model (RFIM). The method combines the event-driven, rejection-free character of the Bortz Kalos-Lebowitz (BKL) algorithm with Glauber transition probabilities, introducing hierarchical probabilistic counters to perform spin selection in O(log N) operations. This enables efficient sampling of the system's dynamics, especially in the low-temperature and low-disorder regime, where traditional Metropolis updates suffer from critical slowing down. Furthermore, this approach allows a proper dynamical simulation of the Ising system's behavior even in the presence of a Random Field (RF), unlike the BKL method. RFIM simulations with Gaussian field distributions reproduce the expected reduction of the pseudo-critical temperature with increasing disorder. Benchmarking shows speedups exceeding two orders of magnitude compared to the Metropolis algorithm in the low-temperature regime. The proposed method provides an efficient and dynamically faithful tool for studying both equilibrium and non-equilibrium phenomena in disordered spin systems.

2026-03-17T18:13:27Z 16 pages, 11 figures Luca Cattaneo Federico Ettori Giovanni Cerri Paolo Biscari Ezio Puppin http://arxiv.org/abs/2603.16830v1 2026-03-17T17:36:38Z

We introduce a postprocessing procedure that recovers sub-cell wave geometry from a standard one-dimensional Euler shock-capturing computation using differentiated Riemann variables (DRVs) -- characteristic derivatives that separate the three wave families into distinct localized spikes. Filtered DRV surrogates detect the waves, plateau sampling extracts the local states, and a pressure-wave-function Newton closure completes the geometry. The entire pipeline adds less than $0.25\%$ to the cost of a baseline WENO--5/HLLC solve. For Sod, a severe-expansion problem, and the LeBlanc shock tube, wave locations are recovered to within roundoff or $O(10^{-4})$ and the contact is sharpened to one cell width; a pattern-agnostic extension handles all four Riemann configurations with errors at the $10^{-6}$--$10^{-8}$ level. Direct comparison with MUSCL--THINC--BVD and WENO-Z--THINC--BVD shows that neither reproduces the combination of sharp contacts, small contact-window internal-energy error, and elimination of the LeBlanc positive overshoot achieved by the DRV reconstruction.

2026-03-17T17:36:38Z 28 pages, 6 figures Steve Shkoller http://arxiv.org/abs/2603.16623v1 2026-03-17T14:58:38Z

The accurate modeling of plasma-based accelerators relies on costly numerical simulations due to the complexity of laser-plasma and beam-plasma interactions. Several strategies can highly reduce the computational cost compared to 3D first-principles particle-in-cell simulations, such as exploiting the near axial symmetry and quasistatic nature of plasma wakefields in many practical cases. Here, we propose a quasistatic algorithm that enables the modeling of axially symmetric plasma wakes without the need of a numerical grid. The gridless approach allows extremely fine features to be resolved without a dramatic increase in computational cost. This is critical, e.g., for the design of future plasma-based colliders with nanometer emittance beams. The proposed model has been implemented in the Wake-T code, where it is coupled to a laser envelope solver and a particle beam pusher to enable the efficient simulation of laser- and beam-driven plasma accelerators.

2026-03-17T14:58:38Z 13 pages, 5 figures Ángel Ferran Pousa Wilbert M. den Hertog Severin Diederichs Al berto Martinez de la Ossa Jorge L. Ordóñez Carrasco Alexander Sinn Maxence Thévenet http://arxiv.org/abs/2603.16524v1 2026-03-17T13:46:29Z

This study presents a novel algorithm based on graph theory for the precise segmentation and measurement of detonation cells from 3D pressure traces, termed detonation lattices, addressing the limitations of manual and primitive 2D edge detection methods prevalent in the field. Using a segmentation model, the proposed training-free algorithm is designed to accurately extract cellular patterns, a longstanding challenge in detonations research. First, the efficacy of segmentation on generated data is shown with a prediction error 2%. Next, 3D simulation data is used to establish performance of the graph-based workflow. The results of statistics and joint probability densities show oblong cells aligned with the wave propagation axis with 17% deviation, whereas larger dispersion in volume reflects cubic amplification of linear variability. Although the framework is robust, it remains challenging to reliably segment and quantify highly complex cellular patterns. However, the graph-based formulation generalizes across diverse cellular geometries, positioning it as a practical tool for detonation analysis and a strong foundation for future extensions in triple-point collision studies.

2026-03-17T13:46:29Z Vansh Sharma Venkat Raman http://arxiv.org/abs/2503.22596v3 2026-03-17T12:28:56Z

Polynomial machine learning potentials (MLPs) based on polynomial rotational invariants have been systematically developed for various systems and applied to efficiently predict crystal structures. In this study, we propose a robust methodology founded on polynomial MLPs to comprehensively enumerate crystal structures under high-pressure conditions and to evaluate their phase stability at finite temperatures. The proposed approach involves constructing polynomial MLPs with high predictive accuracy across a broad range of pressures, conducting reliable global structure searches, and performing exhaustive self-consistent phonon calculations. We demonstrate the effectiveness of this approach by examining elemental silicon at pressures up to 100 GPa and temperatures up to 1000 K, revealing stable phases across these conditions. The framework established in this study offers a powerful strategy for predicting crystal structures and phase stability under high-pressure and finite-temperature conditions.

2025-03-28T16:42:40Z REVTeX 4-2; 24 pages, 18 figures, and 2 tables in the main text; 4 pages and 5 figures in the supplemental material Hayato Wakai Atsuto Seko Isao Tanaka http://arxiv.org/abs/2603.16347v1 2026-03-17T10:31:27Z

Cu-diamond composites are recognized as promising high-thermal-conductivity candidates for electronic cooling, offering tunable properties and competitive cost. However, their performance is significantly limited by the poor Cu/diamond interfacial thermal conductance (ITC). Here, we propose a nitrogen-termination strategy to tune the ITC of Cu/diamond interfaces and unravel atomistic mechanisms by which nitride interlayers tailor phonon transport. Based on the MACE machine-learning interatomic potential (MLIP) framework, we fine-tune the pre-trained MACE-MPA-0 foundation model by incorporating customized C-N-Cu training datasets. Through MLIP-driven lattice dynamics simulations, we demonstrate that an atomically flat N-termination on diamond enhances the ITC by 21% compared to the bare Cu/diamond interface. Mode-resolved phonon spectroscopy reveals that the LA phonons with frequency above 4 THz and wavevectors near Γ-X and Γ-U directions are selectively modulated by N-termination engineering. Analyses of local vibrational states and interfacial bonding further indicate that the N-termination on diamond tunes the interfacial heat conduction via surficial mass modification and bonding regulation, as evidenced by variations in LDOS overlap and COHP spectra. These findings open venues for tuning heat transfer across Cu/diamond interfaces via non-metallic modification, which avoids the graphitization issues associated with metallic coatings, and provide novel guidelines for upgrading the phonon-mediated heat transfer in Cu-diamond composites.

2026-03-17T10:31:27Z Guang Yang Xinling Tang Zhongkang Lin Yulin Gu Wei Hao Yujie Du Xiaoguang Wei http://arxiv.org/abs/2603.16287v1 2026-03-17T09:20:54Z

Pressure-sensitive adhesives (PSAs) are soft polymeric materials that exhibit complex rheological and mechanical behavior gov- erned by the interplay between polymer architecture, crosslink density, and entanglement constraints. Predicting their rheological properties from underlying microstructure remains a central challenge in adhesive design. In this work, we adopt a multiscale com- putational framework based on the Lagrangian Heterogeneous Multiscale Method (LHMM), coupling a macroscopic continuum description with a mesoscale polymer network model featuring breakable bonds embedded in a viscous medium. The approach enables consistent information transfer across scales and captures both elastic network response and viscous dissipation. The framework is calibrated using experimental rheological data and tensile measurements for four PSA formulations with varying gel fractions and crosslink densities. The simulations reproduce key experimental trends in storage modulus (G'), loss modulus (G"), and tensile stress-strain behavior under planar extension, while differentiating the distinct mechanical signatures of each formula- tion. The results elucidate how crosslink density and effective network connectivity control stiffness, stress localization, and failure characteristics. Overall, the proposed multiscale methodology provides a predictive platform for linking microstructural design pa- rameters to macroscopic mechanical properties and offers a rational basis for the formulation and optimization of next-generation PSAs.

2026-03-17T09:20:54Z Nicolas Moreno Elnaz Zohravi Shaghayegh Hamzehlou Edgar Patino-Narino Malavika Raj Mercedes Fernandez Nicholas Ballard Jose M. Asua Marco Ellero http://arxiv.org/abs/2511.11160v3 2026-03-17T09:00:22Z

Electron accumulation in interatomic regions is a fundamental quantum phenomenon dictating chemical bonding and material properties, yet its origin remains elusive across disciplines. Here, we report a quantum accumulation effect -- potential-barrier affinity (PBA) -- revealed by solving the Schrödinger equation for a crystalline potential. PBA effect drives significant interatomic electron accumulation when electron energy exceeds the barrier maximum. This effect essentially enhances interatomic electron density, governing microstructures and properties of condensed matter. Our theory overturns the traditional wisdom that the interstitial electron localization in electride requires potential-well constraints or hybrid orbitals, and it serves as the fundamental mechanism underlying the formation of conventional solid bonding. This work delivers a paradigm shift in understanding electron distribution and establishes a theoretical foundation for the microscopic design of material properties.

2025-11-14T10:43:27Z 13 pages, 9 figures Qiang Xu Zhao Liu Yanming Ma http://arxiv.org/abs/2509.25386v2 2026-03-17T08:34:43Z

In network-based SIS models of infectious disease transmission, infection can only occur between directly connected individuals. This constraint naturally gives rise to spatial correlations between the states of neighboring nodes, as the infection status of connected individuals becomes interdependent. Although mean-field approximations and the standard pairwise model are commonly used to simplify disease forecasting on networks, they inadequately capture spatial correlations; mean-field frameworks assume that populations are well-mixed, while the pairwise model neglects correlations beyond nearest-neighbor connections, which leads to inaccurate predictions of infection numbers over time. As such, the development of approximations that account for higher order spatially correlated infections is of great interest, as they offer a compromise between accurate disease forecasting and analytic tractability. Here, we use existing corrections to mean-field theory on the regular lattice to construct a more general framework for equivalent corrections on random regular graph topologies. We derive and simulate a hierarchical system of ordinary differential equations for the time evolution of the spatial correlation function at various geodesic distances on random networks. Solving these equations allows us to predict the time-dependent global infection density, which agrees well with numerical simulations. Our results substantially improve on existing corrections to mean-field theory for infectious individuals in SIS processes and provide an in-depth characterization of how structural randomness in networks affects the dynamical trajectories of infectious diseases on networks.

2025-09-29T18:42:09Z 10 pages, 6 figures Alexander Leibenzon Samuel W. S. Johnson Ruth E. Baker Michael Assaf http://arxiv.org/abs/2504.21666v2 2026-03-17T05:43:02Z

Estimating partition functions of Ising spin glasses is a cornerstone of statistical physics and computational science, yet it remains classically challenging due to its $\#$P-hard complexity. While Jarzynski's equality offers a theoretical pathway, its practical application is crippled at low temperatures by rare, divergent statistical fluctuations. Here, we introduce a quantum protocol that overcomes this fundamental limitation by synergizing reverse quantum annealing with optimized nonequilibrium initial distributions. Our method dramatically suppresses the estimator variance, achieving saturation in the low-temperature regime where existing methods fail. Numerical benchmarks on the Sherrington-Kirkpatrick spin glass and the 3-SAT problem demonstrate that our protocol reduces computational scaling exponents by over an order of magnitude (e.g., from $\sim 8.5$ to $\sim 0.5$), despite retaining exponential system-size dependence. Crucially, our protocol circumvents stringent adiabatic constraints, making it feasible for near-term quantum devices like superconducting qubits, trapped ions, and Rydberg atom arrays. This work provides a methodological framework for quantum-enhanced estimation in spin glass thermodynamics and beyond by harnessing non-adiabatic quantum dynamics to address a classically difficult problem.

2025-04-30T14:09:40Z 7+11 pages, 3+3 figures Phys. Rev. Lett. 136, 100601 (2026) Haowei Li Zhiyuan Yao Xingze Qiu 10.1103/8gmb-p619 http://arxiv.org/abs/2603.15584v2 2026-03-17T04:43:35Z

Physics-informed neural networks (PINNs) and neural operators (NOs) for solving the problem of diffraction of Extreme Ultraviolet (EUV) electromagnetic waves from contemporary lithography masks are presented. A novel hybrid Waveguide Neural Operator (WGNO) is introduced, based on a waveguide method with its most computationally expensive components replaced by a neural network. To evaluate performance, the accuracy and inference time of PINNs and NOs are compared against modern numerical solvers for a series of problems with known exact solutions. The emphasis is placed on investigation of solution accuracy by considered artificial neural systems for 13.5 nm and 11.2 nm wavelengths. Numerical experiments on realistic 2D and 3D masks demonstrate that PINNs and neural operators achieve competitive accuracy and significantly reduced prediction times, with the proposed WGNO architecture reaching state-of-the-art performance. The presented neural operator has pronounced generalizing properties, meaning that for unseen problem parameters it delivers a solution accuracy close to that for parameters seen in the training dataset. These results provide a highly efficient solution for accelerating the design and optimization workflows of next-generation lithography masks.

2026-03-16T17:46:15Z arXiv admin note: substantial text overlap with arXiv:2507.04153 Vasiliy A. Es'kin Egor V. Ivanov http://arxiv.org/abs/2603.16049v1 2026-03-17T01:24:44Z

Qudits, the multi-level generalization of qubits, provide a natural extension of the binary paradigm in quantum computation and offer new opportunities to enhance algorithmic performance. Beyond their direct applicability to the simulation of multi-level quantum systems, higher-dimensional ancillae can improve sampling efficiency in quantum algorithms by enabling the simultaneous implementation of multiple control operations, thereby reducing circuit complexity. In this work, we pursue three main objectives. First, we present a formulation of the Rodeo algorithm employing a general $d$-level ancilla qudit. Second, we introduce the concept of the \emph{Rodeo kernel}, defined as a two-frequency interferometer, which acts as a spectral filter in the energy domain. Finally, we propose a microcanonical protocol for the Rodeo algorithm. This protocol enables the estimation of entropic quantities through a single energy sweep and admits a natural interpretation as a Gaussian convolution of the density of states. To support the theoretical analysis, we perform numerical evaluations of the corresponding quantum circuit using ancilla qudits of dimensions three, four, and five. The simulations are performed for the one-dimensional Ising model, considering both spin-$\frac{1}{2}$ and spin-$1$ particles. The ancilla qutrit implementation exhibits an $18\%$ reduction in fluctuations compared to the qubit implementation. Our results show that the qudits provide a framework for spectral analysis and thermodynamic characterization of multi-level quantum systems.

2026-03-17T01:24:44Z 19 pages, 6 figures Julio Cesar Siqueira Rocha Rodrigo Alves Dias http://arxiv.org/abs/2602.22372v2 2026-03-16T23:48:00Z

Quantics Tensor Train (QTT) operations such as matrix product operator contractions are prohibitively expensive for large bond dimensions. We propose an adaptive patching scheme that exploits block-sparse QTT structures to reduce costs through divide-and-conquer, adaptively partitioning tensors into smaller patches with reduced bond dimensions. We demonstrate substantial improvements for sharply localized functions and show efficient computation of bubble diagrams and Bethe-Salpeter equations, opening the door to practical large-scale QTT-based computations previously beyond reach.

2026-02-25T20:08:43Z 38 pages, 21 figures, codes at https://tensor4all.org Gianluca Grosso Marc K. Ritter Stefan Rohshap Samuel Badr Anna Kauch Markus Wallerberger Jan von Delft Hiroshi Shinaoka http://arxiv.org/abs/2603.15992v1 2026-03-16T23:04:17Z

We present a physics-constrained, data-driven subgrid-scale (SGS) stress closure for large-eddy simulation (LES) in the lattice Boltzmann method (LBM). Trained on filtered-downsampled (FD) data from LBM direct numerical simulation (DNS) of forced homogeneous isotropic turbulence (FHIT) spanning multiple filter widths, a compact neural network maps nine macroscopic derivative inputs - six strain-rate and three vorticity components - to the six independent components of the SGS stress tensor; a deviatoric projection is applied post-inference to obtain the traceless stress used in the solver. Training combines a stress data loss with physics terms for SGS energy-transfer (Pi) matching, rotational equivariance under cube rotations, and compatibility of the implied SGS forcing with the divergence-based coupling. The predicted stress is coupled to the solver through a split strategy: a dissipative, strain-aligned contribution is represented through an effective-viscosity projection, while the remaining anisotropic residual is applied through a forcing term. This construction is intended to retain both backscatter (via the effective viscosity) and non-dissipative anisotropic effects (via the residual forcing), while remaining compatible with LBM deployment. In the cases considered here, a priori results show good agreement with FD references across stress components and SGS-transfer statistics, and a posteriori rollouts improve several energetic and statistical measures relative to static and dynamic Smagorinsky baselines. A preliminary transfer test in turbulent channel flow is also reported without retraining. Finally, we demonstrate production deployment via ONNX Runtime, with throughput comparable to a dynamic Smagorinsky baseline in the tested configuration.

2026-03-16T23:04:17Z 32 pages, 16 figures Muhammad Idrees Khan University of Rome Tor Vergata, Rome, Italy Sauro Succi Italian Institute of Technology, Rome, Italy Harvard University, Cambridge, USA Hua-Dong Yao Chalmers University of Technology, Gothenburg, Sweden Giacomo Falcucci University of Rome Tor Vergata, Rome, Italy Harvard University, Cambridge, USA