Papers

A fundamental mechanism underlying collective phenomena in social, technological, and economic systems is decentralized decision-making, in which the behavior of individual agents follows from local interactions in the absence of central coordination. Consensus formation is a central feature of such systems, with relevance to blockchain networks, distributed artificial intelligence, autonomous multi-agent systems, distributed control, and collective decision networks. We investigate consensus formation using a stochastic consensus model on random networks and examine how initial conditions, network connectivity, and system size shape the emergence of unanimous states. We show that the stochastic dynamics strongly amplify small initial majorities, progressively suppress the competing state, and drive the system toward a predictable collective outcome. Network connectivity primarily controls the efficiency of this process: increasing connectivity accelerates the propagation of local agreement and reduces the likelihood that fluctuations reverse the initially dominant state, although these gains gradually saturate in highly connected networks. System size produces a complementary effect. Larger networks require more individual updates to reach unanimity, but they are also increasingly reliable in selecting the state favored by the initial majority. Finite-size analysis shows that the range of initial conditions associated with uncertain outcomes becomes progressively narrower as the network grows, decreasing approximately with the inverse square root of the system size. These results reveal a collective amplification mechanism by which weak initial asymmetries become increasingly decisive in large decentralized networks, and they provide a simple framework for understanding the efficiency, predictability, and reliability of consensus formation in distributed decision systems.

Orbital-free density functional theory seeks to compute the energy of an electronic system directly from its electron density, avoiding one-electron wave functions and thereby offering a route to scalable electronic structure calculations. Machine-learned orbital-free density functionals have recently achieved promising results on small organic molecules, predicting energies with sub-millihartree accuracy. However, their convergence in density optimization remains sensitive to hyperparameter tuning and architectural choices. Here, we extend the recently introduced (weak) surrogate functional framework - designed to predict ground-state electron densities only - to also yield their energy, resulting in "strong" surrogate functionals. We find that these learned functionals enable stable convergence across all tested neural network backbones, reducing electron density errors relative to the Kohn-Sham reference by an order of magnitude compared to previous OF-DFT methods. More importantly, the predicted energies are competitive with state-of-the-art machine-learned interatomic potentials (MLIPs) trained only on energies, while exhibiting superior generalization to larger, unseen systems.

PbSe is a narrow-gap IV-VI semiconductor, whose multivalley valence bands, with maxima at the L, $\Sigma$, and $\Delta$ points, underpin its intermediate-temperature thermoelectric properties. We combine soft-X-ray angle-resolved photoemission spectroscopy (SX-ARPES) with first principles simulations to study the valence band structure of bulk PbSe. High resolution measurements are conducted at photon energies of 400-900 eV to map the valence manifold along X$\Gamma$X, WXW, and K$\Gamma$K, and iso-energy surfaces are collected in the $k_z=0$ plane. Comparison to ARPES enables a rigorous assessment of the performance of density functional theory (DFT), using semi-local and hybrid functionals, as well as many-body perturbation theory within the quasiparticle self-consistent $GW$ approximation. We find that the Heyd-Scuseria-Ernzerhof (HSE) hybrid functional and QP$GW$ reproduce the measured band dispersions to within 0.1-0.2 eV over the entire valence band. In contrast, the semi-local Perdew-Burke-Ernzerhof (PBE) functional compresses the band width and deviates from experiment by up to 0.6 eV. We further show that an accurate band structure and band gap are vital to obtaining a correct description of the dependence of the Seebeck coefficient of p-type PbSe on the hole concentration (Pisarenko relation). This has implications for computational efforts to discover thermoelectric materials.

Low critical charge-current density and low energy dissipation are highly desired in magnetic random-access memories, requiring spin sources to exhibit both high charge-to-spin conversion efficiency (CSE) and high charge conductivity. Altermagnets with vanishing net magnetic moment and spin-splitting bands provide promising spin-source candidates for spin-splitting-torque magnetic random-access memories. However, achieving both high CSE and charge conductivity remains challenging in altermagnets. In this work, we introduce Dirac cones into two-dimensional d-wave altermagnets, where their intrinsically high carrier mobility enables tunable charge and spin conductivities with high CSE. Dirac-cone anisotropy provides an effective means of enhancing both CSE and charge conductivity, with cone tilting serving as an additional degree of tunability. Guided by this design principle, we identify a maximum CSE of 92% in Cr2SeTeS. When the Fermi level moves slightly away from the Dirac point, high CSE, high charge conductivity, and the resulting high spin conductivity can be simultaneously achieved. Our study advances the understanding of time-reversal-odd spin transport via Dirac-cone engineering and provides a practical route toward developing spin-source materials that combine high charge conductivity with highly efficient charge-to-spin conversion.

We present the first complete machine learning method for accelerating plane-wave density functional theory (DFT) in materials under the projector augmented wave (PAW) formalism. We formalize seven criteria that a Complete Neural Electronic Initializer must satisfy for practical end-to-end PAW DFT acceleration. Applying these to prior work reveals two structure-dependent components, augmentation occupancies and spin initialization, whose absence prevents existing acceleration methods from providing complete reference-free initialization. We show that omitting these components can eliminate or reverse the acceleration obtained via models that only predict the smooth valence density. We satisfy the missing requirements by introducing AugNet, a general equivariant model for PAW augmentation occupancies, and the first general spin density model for materials, which predicts the smooth spin-difference density and spin-difference PAW augmentation occupancies using predicted magnetic moments to constrain the global magnetic state. Combined with existing valence density models, our full method satisfies all seven criteria and forms a fully reference-free electronic initializer for materials DFT, requiring no electronic quantities from a converged target calculation. We show that perfect initialization could cut PAW DFT wall time by 40-52%, and our method recovers up to 62% of this saving, reducing end-to-end DFT wall time by up to ~25% on unseen structures while preserving converged energies.

The collinear-antiferromagnetic spin-fluctuation paradigm has long guided unconventional superconductivity research, yet fails to reconcile the noncollinear spin phenomena observed across cuprates, iron-based superconductors, and nickelates. Using extensive first-principles calculations and unbiased large-scale DMRG simulations, we show that Dzyaloshinski-Moriya interaction (DMI)-arising from local inversion-symmetry breaking-is a common ingredient across these families. This DMI unifies hallmark observations in parent compounds-incommensurate orders, spin-wave gaps, and noncollinear textures. Under hole doping, strong DMI drives spin vortices to merge with pi-shifted hole stripes, forming hybrid vortex-hole stripe phases. These phases stabilize charge order while supporting, not suppressing, superconductivity. By contrast, under electron doping, these vortices pin holes and suppress long-range superconductivity. Our results establish DMI as a unifying link between noncollinear magnetism and superconductivity, identifying hole-strip-vortex coupling as a microscopic pairing engine. Given that DMI is common across major superconductor families, these findings challenge the prevailing pairing mechanism and offer an experimentally testable roadmap for materials optimization.

The response of magnetic nanoparticles (MNPs) to applied magnetic fields underpins a broad range of biomedical and technological applications. In this Perspective, we review the principal modeling approaches for describing MNP dynamics across different physical regimes, ranging from coarse-grained macrospin descriptions to spatially resolved micromagnetic simulations. Selecting an appropriate model depends on the relevant energy scales and timescales, including those associated with magnetic anisotropy. We compare the assumptions, computational requirements, and regimes of applicability of the fixed-point-dipole, effective-field, thermal Stoner-Wohlfarth, diffusion-jump, coupled Landau-Lifshitz-Gilbert, egg, and micromagnetic models. Particular attention is given to coupling magnetization dynamics with translational and rotational particle motion, hydrodynamic interactions, and long-range dipolar interactions. By relating the relevant physical regimes to the resolution and computational cost of each approach, we provide practical guidance for model selection and outline challenges for predictive multiscale simulations of interacting MNP systems.

Cryogenic computing technologies are maturing rapidly, but heat removal remains a key challenge when increasing the device density and operating power. Two-phase evaporative cooling is a promising approach to solve this issue, because it can dissipate high heat fluxes while maintaining small temperature rises. Here, we numerically investigate liquid-$\mathrm{N_2}$-filled silicon micropillar wicks as a capillary-fed thin-film evaporation concept for cryogenic electronics. The model combines Young-Laplace meniscus calculations, Hertz-Knudsen-Schrage evaporation, unit-cell heat-transfer and liquid-flow simulations, and an array-level thermal and capillary-flow model. For a representative geometry with a pillar diameter of $10\,\mu\mathrm{m}$, pitch of $24\,\mu\mathrm{m}$, and pillar height of $75\,\mu\mathrm{m}$ at an applied heat flux of $20\,\mathrm{W\,cm^{-2}}$, the predicted chip-temperature rise is approximately $2.7\,\mathrm{K}$. This is substantially below the estimated temperature rises for representative conduction cooling through an indium-interlayered copper heat sink and direct liquid-$\mathrm{N_2}$ immersion. The corresponding predicted single-fed dry-out length is approximately $2.7\,\mathrm{mm}$, equivalent to an ideal double-fed coolable width of approximately $5.3\,\mathrm{mm}$. The results indicate that local thermal performance is favorable, whereas lateral scalability is primarily constrained by capillary dry-out. An approximate capillary-viscous scaling relation provides a compact framework for comparing dry-out limits across working fluids and related wick geometries and for identifying strategies to extend capillary-fed transport.

While underground reservoirs emerge as essential elements to face global warming, these systems represent complex multi-physical and multiscale problems. The considered injection of fluids during hydrogen storage, carbon dioxide sequestration, or geothermal energy recovery involves a modification of the chemical equilibrium of the fluid in the porous reservoir. Chemical reactions can induce microstructural changes of the rock matrix, leading to a reduction of elastic properties of the material, and to potential settlement or stress redistribution. Consequently, it becomes pivotal to establish predictive behavior laws to describe the effect of chemical damage on elastic properties. Facing the difficulties to estimate experimentally the impact of chemical damage on mechanical properties, a Digital Rock Physics approach is proposed in this contribution. This numerical homogenization scheme is used to compare two distinct types of microstructure models: the first one consists in a Discrete Element Model, while the second one employs a continuous description. This continuous formulation is based on a Phase-Field description to predict the evolution of the microstructure subjected to chemical alterations and on the Fast Fourier Transform to estimate the macroscopic properties of the material. Finally, these frameworks establish different softening laws that can be used as constitutive ingredients for a cemented material during its weathering.

Polarizability predicts how an object responds to an incident electromagnetic field, the interactions between small particles, or the optical forces exerted upon them. Polarizability is responsible for the effective-medium properties of artificial materials or metasurfaces. Despite significant progress in all these areas, it is unclear what the limits of the strength of such interactions are or, more specifically, what the upper bounds on the polarizability of a given spatial region that an unknown and designed particle would occupy are. This work connects the electromagnetic field description via an integral equation with a dual formulation of quadratic programming to derive fundamental bounds on components of all four polarizability tensors or on their specific combinations. In particular, the work establishes an intuitive visualization of what strong polarizability means and how strong it can be. The developed fundamental bound also answers which materials and domains are best for the given demands on polarizability. These findings establish a versatile platform that can accommodate a wide range of demands on polarizable bodies, providing an absolute measure of their performance against which the results of human-powered or automated design procedures can be compared.

Freestanding thin films of quantum materials naturally develop sub-micrometer, nonuniform strain fields that strongly affect their electronic, magnetic, and structural properties in both equilibrium and nonequilibrium conditions. However, current methods to quantitatively resolve these mesoscopic features are primarily restricted to scanning probes, making it challenging for dynamical measurements such as single-shot imaging and femtosecond microscopy. Here, we present a new computational framework which we denote as bend contour electron tomography (BCET), which efficiently converts bend contours in transmission electron microscope images into quantitative two-dimensional maps of strain and topography. By iteratively minimizing a designed loss function between experimental and simulated bend contour images, BCET retrieves both surface morphology and in-plane strain tensor fields without requiring scanning or diffraction mapping. We applied BCET to freestanding SrTiO$_3$ thin films, demonstrating the successful reconstruction of the local strain distribution and curvature field with high fidelity. Our approach provides a quantitative framework for characterizing mesoscale structures in freestanding films using wide-field imaging, opening new avenues to investigate how spatial inhomogeneity governs phase transitions and nonequilibrium dynamics in two-dimensional quantum materials.

Modern inverse problems and adaptive simulations require efficient and differentiable representations of complex geometries that can be updated frequently for gradient-based optimization. Existing methods cannot simultaneously achieve both. We present Warp-Geo, a differentiable GPU-accelerated framework for reconstructing signed distance fields and evaluating surface normals from point clouds of complex 3D geometries. Warp-Geo combines uniform grids for GPU parallelism with implicit differentiation through the Poisson solve, enabling frequent geometry recomputation while maintaining full differentiability. We demonstrate Warp-Geo on forward SDF reconstruction, dynamic fluid–structure interaction with moving boundaries, and inverse shape optimization via automatic differentiation with gradient validation. GPU scalability analysis confirms practical feasibility for frequent geometry updates. The framework, implemented in Warp and JAX, enables seamless integration with differentiable physics solvers for end-to-end shape optimization.

Pretraining a neural PDE surrogate can reduce the amount of new CFD data needed when geometry or modeled physics changes. However, it remains unclear how different components of distribution shift affect this benefit. We pretrain a surrogate on 254,909 RANS solutions from one airfoil family and fine-tune it on a new family under two target settings with matched freestream ranges: the same Spalart-Allmaras (SA) modeling and SA with added $e^N$ transition modeling. At $N=1000$, the pretrained model matches the accuracy of a model trained from scratch on $3.25\times$ as many samples for the same-SA target, but $2.58\times$ as many for the transition-modeled target. By $N=5000$, this ordering reverses ($1.56\times$ versus $1.86\times$). At $N=1000$, sampling more distinct airfoils lowers error on both targets, but only for the same-SA target is the gain increase larger than the observed draw-to-draw variation ($3.3\times$ to $4.0\times$). These results show that pretraining value depends jointly on target-data budget, target-data coverage, and whether source and target differ in modeled physics.

We report improved global-minimum configurations for the classical Thomson problem of $N=60$, $61$, $92$, and $99$ repulsive Coulomb charges confined to a disk. By combining the quenched molecular dynamics (QMD) method with the fixed-border heuristic introduced by Amore and Zarate, we systematically obtain configurations with energies $E_{\mathrm{QMD}}(60)=2159.3584240930$, $E_{\mathrm{QMD}}(61)=2237.19264190$, $E_{\mathrm{QMD}}(92)=5358.35353314$, and $E_{\mathrm{QMD}}(99)=6254.83029083$, which improve upon the previously best-known values. For $N=60$, the Voronoi diagram of our configuration differs from the one reported earlier. The symmetries for $N=61$ $(C_{2})$ and $N=99$ $(D_{1})$ are confirmed and are consistent with the known symmetry patterns for these configurations. For $N=92$, whereas the previously reported Voronoi diagram has lower symmetry, our solution exhibits a clear $D_{1}$ (axial) symmetry of defects. The results are highly reproducible across multiple independent runs, providing strong evidence that these configurations are robust global-minimum candidates.

High-performance racing shoes rely on ultra-low-density elastomeric foams that undergo large, repeated deformations during running. Yet little is known about how their mechanical properties vary throughout the shoe or change with repeated use. Here, we characterize the midsole foam in an elite-level racing shoe from the heel, midfoot, and toe of a new shoe and a shoe worn for 300 miles. Microscopy reveals a characteristic pore length scale of 128+/-18 um and supports an approximately isotropic continuum description. We then quantify the mechanical response under tension, compression, and shear. Remarkably, despite 300 miles of real-world use, the foam retains its mechanical response across all three loading modes and shoe regions. Energy return remains largely unchanged, with values of 85-93% in tension and compression and 64-71% in shear. At the same time, we observe strong regional variations, with tensile and compressive stiffnesses 38-51% lower in the toe than in the heel. The foam also exhibits a pronounced tension-compression asymmetry in Poisson's ratio. Together, these findings reveal a spatially structured and mode-dependent mechanical response that remains largely preserved after 300 miles of use. This mechanical resilience may extend the functional lifetime of racing shoes, with implications for runners, replacement recommendations, and sustainability.

Studying the dynamical behavior of a system often depends on characterizing how it transitions between long-lived states. Because such transitions are rare, observing them usually requires specialized enhanced sampling techniques. Transition Path Sampling (TPS) is a well-established method for generating reactive trajectories, which is simple to implement and does not require the definition of a preconceived reaction coordinate. However, its efficiency is limited by its sequential nature and the resulting correlations between sampled paths. Previous work addressed this limitation by combining TPS with a sampling scheme based on conditioned Boltzmann Generators, a generative machine learning model capable of sampling a given target probability distribution. This approach produces uncorrelated transition paths but relies on an accurate reaction coordinate, which is rarely known in advance. Building on recent advances in committor learning, specifically on the Artificial Intelligence for Molecular Mechanism Discovery (AIMMD) method, in this work we introduce GenAIMMD, an iterative algorithm that actively and self-consistently learns the ideal reaction coordinate (the committor) and trains a conditioned Boltzmann Generator to sample from arbitrary bias windows along it. GenAIMMD thereby provides a correlation-free and fully parallelizable path sampling scheme that does not require prior knowledge of the system's transition mechanism. We apply GenAIMMD to a two-dimensional toy model and a higher-dimensional polymer system. In both cases, GenAIMMD succeeds in training the Boltzmann Generator and learning the committor. Benchmark results show a substantial increase in performance compared to standard TPS.

Monolayer tin selenide (SnSe) exhibits phase-dependent anharmonic lattice dynamics, yet their consequences for the thermoelectric power factor (PF) and point-defect tolerance remain unresolved. We combine density functional theory, the stochastic self-consistent harmonic approximation (SSCHA), and Boltzmann transport calculations including electron-phonon and electron-defect scattering to investigate monolayer $\alpha$-SnSe (Pnma) and $\beta$-SnSe (Cmcm). In dynamically stable $\alpha$-SnSe, SSCHA renormalizes the finite-temperature phonons without changing the qualitative n-type transport picture. In $\beta$-SnSe, SSCHA removes the harmonic soft-mode instability of the Cmcm phase at 800-1000 K, and thereby enables high-temperature transport calculations; LO/TO-2 is the principal electron-scattering channel. In the lower-density window near $10^{12}$ cm$^{-2}$, the n-type PF reaches 15-19 $\mu\mathrm{W}/(\mathrm{K}^{2}\cdot\mathrm{cm})$ at 800-900 K and exceeds the p-type PF primarily because of the higher electrical conductivity. Se vacancies ($V_{\mathrm{Se}}$) produce weaker electron-defect scattering than Sn vacancies ($V_{\mathrm{Sn}}$), and p-type transport is less defect tolerant than n-type transport in both phases. We define an operational critical defect concentration, $C_{\mathrm{crit}}$, at which the PF decreases by 15% relative to the corresponding defect-free value. The lowest $C_{\mathrm{crit}}$ is $8.841\times10^{-5}$ (approximately 88 ppm) for p-type $\alpha$-SnSe with $V_{\mathrm{Sn}}$; for n-type $\beta$-SnSe with $V_{\mathrm{Se}}$, the 15% threshold is not reached up to $5\times10^{-3}$ (5000 ppm). These results distinguish finite-temperature phonon renormalization in stable $\alpha$-SnSe from anharmonic stabilization in $\beta$-SnSe and provide defect-concentration limits for preserving the PF.

Computational materials databases have become central infrastructure for data-driven discovery of inorganic materials, yet their growth remains strikingly uneven across property families. This perspective synthesizes a systematic survey of mainstream density functional theory (DFT) software, the computational cost and credibility of nineteen material-property families, and the coverage of existing computational databases, into a coherent picture of where the field stands and where it should go. We show that the ecosystem of first-principles codes is methodologically mature: for nearly every property of technological interest, at least one production-grade code can compute it.The binding constraint is no longer methodological capability but the economics of trust - which properties can be computed cheaply enough, and accurately enough, to be harvested at database scale. Mapping database coverage onto a Gartner-style readiness cycle reveals a sharp divide: ground-state structure, energetics, elasticity, and topology have reached routine production, while nine property families - including NMR/EPR parameters, core-level spectra, electron-phonon properties, thermal conductivity, and quantum transport - remain without any systematic computational database. We argue that these blank zones define the scientific opportunity of the next decade, and we propose a three-horizon roadmap: consolidating coverage and interoperability in the near term, industrializing mid-cost properties through surrogate-accelerated workflows in the medium term, and conquering the high-cost frontier through machine-learned interatomic potentials, autonomous computing infrastructure, and community governance in the long term.

The performance synthesis calculation of aircraft engines relies on tabulated datasets for the simulation of the complex turbomachinery components at acceptable runtimes. Typically, these so-called performance maps express the change of the fluid's energetic state over the component in terms of the total pressure ratio and the isentropic efficiency of the process. However, the definitions of both parameters are different for compressors and turbines and the isentropic efficiency is in both cases not defined for a unity pressure ratio. Moreover, this parameter combination is not directly applicable when modelling the engine at the aircraft level, which is required for modern highly integrated aircraft design. As exergy analysis is an established tool for aircraft design asessment, novel performance maps are proposed that describe the change in exergy and anergy over a given turbomachine component. Both changes are expressed as non-dimensional parameters whose definition is consistent for compressors and turbines, and is compatible with a unity pressure ratio. It is shown that the presented exergy-anergy maps are obtainable on a standard turbomachine test bed or through transformation of existing maps. It is highlighted using examples, that the conversion between existing maps and exergy-anergy maps is completely lossless. The different operating regimes of the turbomachines are clearly distinguishable in the novel map representation, which allows an assessment of the map's physical consistency. It is therefore concluded, that the exergy-anergy maps are an important alternative to the performance map variants established today.

Permutation invariant polynomials (PIPs), together with related polynomial-based invariant descriptors such as fundamental invariants (FIs), are widely used in constructing high-fidelity global potential energy surfaces (PESs) of molecules that contain identical atoms. Although the monomial symmetrization approach (MSA) enables fast evaluation of PIPs through recursive factorization, it leads to deeply nested computational graphs that may require large memory and are inefficient in modern automatic differentiation frameworks. In this work, we introduce JaxPIP, a JAX-based implementation that reformulates PIP/FI evaluation into tensorized linear algebra operations. By replacing recursive factorization with dense matrix operations combined with log-exp transformation and segmented summation, the evaluation becomes regular and GPU-friendly. This allows efficient execution with just-in-time compilation and enables large-scale batch evaluation of energies and forces (as well as higher-order derivatives). The resulting architecture supports ensemble simulations such as quasi-classical trajectory (QCT), path-integral molecular dynamics (PIMD), and diffusion Monte Carlo (DMC) in a fully vectorized manner. As demonstrated in examples, JaxPIP provides a practical route for efficient simulations of molecular systems with scalable GPU execution.

This paper introduces the mollified-sharp decomposition, a probabilistic regularization of moving shocks in parametric reduced-order models. Each detected shock location is treated as a random variable with a prescribed probability density. Averaging over this artificial distribution replaces the localized pressure change by a smooth transition whose spatial extent is set by the density width, rather than by direct filtering of the pressure field. Each snapshot is decomposed exactly into a regularized mollified field and a local sharp correction that restores the shock. This probabilistic construction and exact additive split define the general method; the detector, kernel, treatment of multiple shocks, alignment coordinates, and regression are implementation choices. In the present realization, a calibrated indicator detects shocks, a compactly supported Wendland kernel mollifies them, and a peak-normalized and, where necessary, partitioned weight derived from each shock-location probability density defines the centroid, principal axes, and scales of its local alignment domain. Separate POD-GPR models represent the mollified field and aligned corrections, with additional regressions for shock presence and alignment. The method is demonstrated on the transonic airfoil pressure data of Catalani et al. (2023) and compared with a standard POD-GPR model constructed from the same data and POD-energy criterion. The mollified-sharp model reduces the mean test-set relative $L^2$ pressure error by 31.2% and the mean test-set surface-pressure-coefficient error by 33.2%, while also improving the predicted shock locations and pressure changes; the trade-off is a median online evaluation time 2.24 times as long. The construction is applicable in principle to other parameter-dependent fields with moving sharp features, such as moving material interfaces in multiphase flows.

Neural networks provide expressive representations for scientific computing. However, even sufficiently expressive networks can suffer training failure in weak-gradient regimes, limiting their practical use in quantum many-body physics and ab initio quantum chemistry. Here we derive an unbiased direct gradient estimator and introduce the adaptive minimum-variance phase (AMVP) estimator for neural-network variational optimization. By improving the signal-to-noise ratio of weak gradients, these methods enable reliable scientific calculations where training previously failed, while substantially reducing computational cost. The framework enables compact networks to outperform larger and fine-tuned default standard-estimator models with over an order of magnitude less GPU time on correlated flux models, and ultimately exceed the density matrix renormalization group (DMRG) accuracy. It further achieves chemical accuracy in N$_2$ bond breaking and, for the first time, in heavy-element I$_2$ with explicit spin-orbit coupling. These results demonstrate that gradient-estimator design expands the capabilities of neural-network variational methods for accurate scientific computing.

What can and cannot be predicted by a model when only limited information is available? We answer this question by constructing a ladder of information limits that organizes prediction tasks according to their information requirements: from memoryless and memory-augmented trajectory forecasts to event prediction, stationary statistics, and generative laws. The analysis, which is independent of the model structure or architecture, identifies the minimum attainable error along with the information limitations that give rise to it. The approach distinguishes information hidden in unresolved variables from information recovered through time-delayed observations and connects these contributions to Mori–Zwanzig memory. The ladder also reveals why Lyapunov growth alone cannot characterize reduced-order prediction error, why stationary statistics may remain predictable even when individual trajectories become unpredictable, and why stochasticity can represent (but cannot recover) missing information. Numerical studies of the Kuramoto–Sivashinsky equation and the Lorenz system illustrate these results across the ladder. This unified perspective provides a principled language for understanding the fundamental limitations of ROMs for a given prediction task and for clarifying whether improved predictions require richer observations, greater precision in the input, or additional memory.

We investigate pressure-dependent melting of ordered B2-NiAl using neural-network molecular dynamics with a Deep Potential interatomic model. Melting temperatures are determined from two-phase solid-liquid coexistence simulations over a broad pressure range, yielding the melting curve $T_m(P)$. Relative to available experimental and previous molecular-dynamics results, the present calculations predict a stronger increase of the melting temperature with pressure at elevated compression. To assess the thermodynamic consistency of the calculated melting line, we evaluate the enthalpy and volume changes upon melting and compare the Clapeyron slope with the derivative of the fitted $T_m(P)$ curve. The two estimates are in good agreement over most of the investigated pressure range, supporting the internal consistency of the coexistence results. To probe the character of melting, we perform a layer-resolved composition analysis of the coexistence configurations and find that the coexisting liquid remains essentially equiatomic at all studied pressures, with deviations of the aluminum fraction from the stoichiometric value not exceeding $5\times10^{-3}$. This provides direct atomistic evidence that melting of B2-NiAl remains congruent within the present model. Together, these results establish a thermodynamically consistent pressure-dependent melting description of B2-NiAl and clarify the character of its melting under compression.

Faster quantum gates can suppress the decoherence accumulated during a computation, but in superconducting processors stronger microwave pulses can also increase leakage, off-resonant excitation, stray-field errors, and crosstalk. This creates a central energy–speed–error tradeoff: can quantum state engineering make a gate parametrically faster without paying proportionally more drive energy? We address this question by treating the driving pulse as a quantum bosonic field rather than a classical waveform. For a finite-dimensional system coupled linearly to that field, we prove that fixed-fidelity gate transition rates grow at most as the square root of the pulse energy, under stated uniformity conditions on the coupling and accepted dynamics. The bound permits arbitrary pulse states, including squeezed and non-Gaussian states, as well as drive–system entanglement and back action; coherent Gaussian pulses attain its energy exponent. Thus squeezing or other state engineering alone cannot replace the square-root energy law of a conventional linear drive by the linear scaling allowed by general quantum speed limits. Achieving that improvement requires changing the interaction class, in addition to using a suitable nonclassical pulse, thereby identifying interaction nonlinearity as an essential resource for relaxing the practical gate-speed error tradeoff.

We present IB-Flows, an open-source, multi-GPU solver for the direct numerical simulation of incompressible fluid-structure interaction (FSI) problems. The code couples a second-order finite-difference fractional-step Navier-Stokes solver on a staggered Cartesian grid with two immersed boundary methods: a Lagrangian method based on moving least-squares (MLS) interpolation and an Eulerian sharp-interface method. Rigid bodies are advanced with a quaternion-based Newton-Euler solver, while deformable surfaces embedded in three-dimensional flows are described by a structural solver based on the interaction potentials; a predictor-corrector scheme provides either loose or strong fluid-structure coupling. Non-Newtonian fluids with shear-thinning or shear-thickening rheology are handled through a strain-rate-dependent viscosity, and the same framework accommodates subgrid-scale eddy-viscosity models for large-eddy simulations of turbulent flows. The solver is written in CUDA Fortran with MPI domain decomposition, so that the fluid, interpolation and structural kernels are all executed on the GPUs, while the distributed transposes required by the Poisson and implicit solvers are further accelerated by the cuDecomp library through GPU-aware communication. Strong- and weak-scaling tests show near-ideal intra-node scaling and good multi-node efficiency. The solver is validated against a set of benchmark problems spanning wall-bounded turbulence, rigid-body and deformable-body fluid-structure interaction, and a biomedical application. IB-Flows is intended as a reproducible reference implementation for immersed boundary FSI simulations on modern GPU clusters and as a transparent platform to be used for tackling multiphysics problems.

Plasmonic excitation of metallic nanostructures can generate energetic carriers capable of transferring charge into nearby adsorbed molecules, providing a possible pathway for driving chemical transformations. In many theoretical descriptions of this process, charge transfer is determined by the equilibrium electronic structure of the molecule and substrate together with the time-dependent carrier distribution produced during plasmon decay. However, the intense transient electric fields associated with localized surface plasmons can also dynamically perturb the molecular electronic structure itself on ultrafast timescales. In this work, we investigate how these time-dependent fields modify the molecular spectral function and influence charge injection at catalytic interfaces. Using a model of CO2 adsorbed on Au(111), we compute the real-time molecular Green's function within a correlated frontier-orbital active space under plasmon-like driving. We find that the driving field rapidly produces transient Floquet-type replica bands in the molecular density of states, opening additional resonant pathways for hot-carrier injection that are absent without external driving or in time-local descriptions. Coupling the evolving molecular spectrum to a time-dependent hot-electron distribution described within a two-temperature Sommerfeld framework, we predict large enhancements in quasiparticle spectral overlap that governs injection during plasmon dephasing. These results suggest that dynamically generated non-equilibrium spectral structure may play an important role in plasmon-assisted catalysis and provide a framework for studying driven catalytic interfaces beyond static electronic structure descriptions.

The test error of a model plotted against its number of parameters $d$ falls, peaks when the model can just fit the training data, and falls again, exhibiting the double descent phenomenon. We explain the phenomenon with statistical mechanics. The training trajectory of a stochastic gradient-based method is a particle wandering over the energy landscape of the training loss at an induced temperature $T$, and a run that has equilibrated visits every parameter vector of a given training loss equally often, the fundamental postulate of statistical mechanics, with probability given by the Boltzmann distribution. Because training starts at an initial point and has only finite time to diffuse, it carries an effective weight decay, which makes every parameter a quadratic degree of freedom. The equipartition theorem then distributes the energy among the $d$ degrees of freedom in shares of $T/2$, so at a fixed training loss adding parameters lowers the temperature and drives the Boltzmann distribution toward the stationary path. Finally, adding parameters can only lower the $L^2$ norm of the stationary path, so a solution sampled at fixed loss is less likely to be large with increasing $d$, effectively increasing weight regularization.

Direct quadrature of the Hadamard-reduced sixth-order time-convolutionless (TCL6) generator at all $N_t$ sampled times requires $O(N_t^3)$ operations at fixed system dimension. We derive an exact reduction for a finite-dimensional system coupled through one Hermitian operator to a stationary centered Gaussian bath. By separating fixed operator coefficients from scalar bath kernels, the complete TCL6 time series is reduced to cumulative sums and first-order recurrences, one-dimensional causal convolutions, and an exact dyadic recursion for interlocked histories. With the algorithm, evaluating the complete TCL6 time series requires $O(N_t\log^2 N_t)$ operations at fixed system dimension. In addition, we show that TCL$2n$ can be evaluated with $O(N_t\log^{n-1} N_t)$ complexity. A fixed finite Matsubara expansion of the bath correlation function permits $O(N_t\log N_t)$ evaluation at any fixed TCL order. These reductions enable fast long-time simulations of non-Markovian open quantum systems within the regime of validity of the TCL expansion.

In this letter, a novel 3-D Finite-Difference Time-Domain (FDTD) framework is proposed that circumvents the costly volumetric scaling of conventional FDTD methods. By representing the electromagnetic fields as low-rank Quantized Tensor Trains (QTT), the memory requirements scale logarithmically with system size. Moreover, the various numerical operations that constitute the FDTD scheme can be efficiently implemented in this format, with their computational cost also exhibiting a logarithmic complexity. The simulation of systems in open space is enabled by the inclusion of a uniaxial PML. A validation example demonstrates that the method achieves excellent accuracy compared to the traditional full-grid (FG) FDTD method, while significantly reducing the required computational resources. Memory savings of several orders of magnitude are obtained, highlighting the potential of the proposed framework for the simulation of large multiscale electromagnetic systems.