We propose a computational strategy that incorporates time-lagged sampling into the reverse process of diffusion-based machine learning models. This approach aims to maintain ensemble diversity and prevent ensemble collapse related to underestimation of uncertainties, particularly in situations with limited ensemble sizes. We demonstrate the method's effectiveness using the Lorenz '96 system as a theoretical testbed, focusing on its highly nonlinear regime without explicitly resolving multiscale interactions. Our approach facilitates stable forecast-analysis cycles at small ensemble sizes without requiring covariance inflation or localization by enhancing the ensemble representation of uncertainties inherent in diffusion models and stochastic dynamics. The proposed models maintain superior temporal coherence. While the probability density function is narrowed under severely undersampled ensembles, stochastic variants recover distributional accuracy as ensemble size increases, whereas deterministic variants show persistent limitations but provide substantial computational efficiency. The method generalizes successfully to systems with increased multiscale complexity and maintains stable long-term performance without drift. Results indicate that active diversity preservation in ensemble generation can enable robust data assimilation in undersampled situations, with potential applications in geophysical flows, plasma systems, and turbulence modeling.

Inertial fusion energy promises nearly unlimited, clean power, if it can be achieved. However, the design and engineering of fusion systems requires controlling and manipulating matter at extreme energies and timescales; the shock physics and radiation transport governing the physical behavior under these conditions are complex, requiring the development, calibration, and use of predictive multiphysics codes to navigate the highly nonlinear and multifaceted design landscape. We hypothesize that artificial intelligence reasoning models can be combined with physics codes and emulators to autonomously design fusion fuel capsules. In this article we construct a multiagent system where natural language is utilized to explore the complex physics regimes around fusion energy. The agentic system is capable of executing a high-order multiphysics inertial fusion computational code. We demonstrate the capacity of the multiagent design assistant to translate natural-language design goals into simulation execution, physics-emulator construction, analysis, and iterative refinement of capsule-geometry parameters, ultimately achieving simulated ignition within the computational design workflow.

We present experiments on chiral active polar particles, realized as vibrated granular rods, revealing evidence for “skipping orbits” at hard boundaries. These classical edge states exhibit a net circulation opposite to the particles' intrinsic rotation and lead to a pronounced accumulation at the boundary, stronger than for their achiral counterparts. The directed nature of these orbits provides a simple yet high-fidelity mechanism for chiral sorting—even for solitary particles, unlike in [T. Barois et al., Phys. Rev. Lett. 125, 238003 (2020)]. We propose a unified theoretical framework for boundary interactions of both chiral and achiral particles. In this model, an effective outward radial force, proportional to motility and chirality, explains the observed boundary-hugging. Our theory predicts, and our experiments confirm, a transition in the pairing of two particles of the same chirality, from apolar spinners to polar circle walkers, with increasing packing fraction of an ambient medium of beads.

This paper investigates the dynamic properties of a confined quasi-two-dimensional granular fluid at moderate densities, modeled within the framework of the Enskog kinetic equation. The confinement is not treated explicitly at the geometrical level; instead, it is incorporated in an effective way through the so-called $\mathrm{\Delta }$ model (a collisional model that includes energy injection through modified collision rules). The referred model is extended in this paper to account for the influence of an interstitial gas via a viscous drag force and a stochastic Langevin-like term. By applying the Chapman-Enskog method, the Navier-Stokes transport coefficients and the cooling rate are derived analytically considering the leading terms in a Sonine polynomial expansion. The study focuses on steady-state conditions and examines how the combined effects of inelastic collisions and external driving influence transport properties such as the viscosity and the thermal conductivity. Theoretical predictions for the steady temperature and the kurtosis are validated against direct simulation Monte Carlo results, showing excellent agreement. The findings reveal that the external driving significantly alters the transport coefficients compared to dry (no gas phase) granular systems, challenging previous assumptions that neglected these effects. Additionally, a linear stability analysis demonstrates that the homogeneous steady state is stable across the explored parameter space.

Phase reduction provides a principled relationship between high-dimensional oscillator dynamics and low-dimensional phase models such as the Kuramoto model. However, connecting analytical results derived from the Kuramoto framework to experimental data remains an open challenge, as many physical platforms violate the weak-coupling assumptions underlying standard phase reduction. Here, we present an experimental platform based on electronic quadrature oscillators whose phase dynamics are, under certain assumptions, equivalent to the Kuramoto model. Crucially, this equivalence is achieved for both weak and strong coupling through a nondiffusive coupling circuit, and the design approach supports arbitrary (weighted and directed) network topologies while preserving scalability and oscillation regularity. Across different coupling schemes, we demonstrate that the platform reliably reproduces phase transitions predicted by the Kuramoto model for both global and cluster synchronization. This cost-effective platform enables systematic experimental validation of analytical and numerical results for synchronization in complex networks.

The stability of periodic states under applied strain is investigated using the one-dimensional phase-field-crystal (PFC) model. We perform a linear stability analysis of periodic states within the one-mode approximation and identify a marginal zero-wave-number mode, corresponding to the global translational symmetry, which controls the onset of instability at long wavelengths. Two distinct instability mechanisms are found. At low mean densities, phase fluctuations in the complex amplitude dominate, leading to a phase (elastic) instability of the periodic state. At higher densities, amplitude softening and density fluctuations drives a spinodal-like melting and the onset of liquid-solid coexistence. The crossover between these two regimes occurs at a sharp threshold coinciding with the onset of coexistence, and is confirmed by direct numerical simulations of the full PFC dynamics. Our results clarify how mean density and strain control the loss of periodic order through competing elastic (phase) and amplitude-driven instabilities.

We present a polynomial-based smoothing and curvature estimation pipeline for extracting geometric descriptors from noisy manifolds produced by segmentation of biological surfaces, with application to thoracic aortic geometry from computed tomography angiography (CTA). The approach is a mesh-adapted variant of local polynomial regression: we fit low-order polynomials on Monge patches to smooth segmentation noise while preserving volume, then estimate curvatures from the same locally defined polynomials. Conceptually, our smoothing step generalizes classic Savitzky-Golay least-squares fitting to irregularly sampled surface neighborhoods in the style of Locally Estimated Scatterplot Smoothing, while the curvature step follows standard differential geometry on Monge patches. Per-facet curvatures are aggregated into scale-invariant integrated quantities including the total squared and integrated Gaussian curvatures, which serve as global shape descriptors. The result is a transparent, reproducible workflow that preserves interpretable surface features and yields physically meaningful descriptors suitable for downstream modeling. We validate on a cohort of 267 CTA scans from normal and dissected aortas. As an illustrative downstream task, we show that integrated curvature descriptors derived from our pipeline recover outcome-group separability comparable to prior analyses, while providing a richer per-facet curvature field for future modeling.

Strong-anisotropy magnetic multilayers composed of two identical ferromagnetic layers separated by a nonmagnetic spacer are modeled by the Ising model including intralayer dipolar interactions and interlayer exchange coupling (IEC), where frustration emerges from the competition between interactions acting at different length scales. The effects of the IEC on the phase behavior are systematically analyzed by Monte Carlo simulations using the Wang-Landau (WL) and short-time dynamics (STD) methods. For a fixed identical ratio between exchange and dipolar constants, the temperature-IEC phase diagram at zero external magnetic field is determined. It is demonstrated that the layers reproduce the same low- and high-temperature phases as the bidimensional Ising dipolar model (IDM), namely, stripe domains and a tetragonal liquid phase, respectively. The continuous character of the transition line between these phases is independently confirmed by both WL and STD approaches. Remarkably, a scaling law for the reduced critical temperature is uncovered, characterized by a shift exponent close to the critical susceptibility exponent ($\gamma$) of the bidimensional IDM. This finding generalizes the scaling predictions previously reported for the order-disorder transition in short-range Ising bilayers to systems with long-range dipolar interactions. Furthermore, STD results reveal that the critical temperature increases monotonically with the IEC strength, reaching values up to twice the critical temperature of the IDM. This behavior demonstrates a strong IEC-induced enhancement of the thermal stability of the stripe phase, thereby extending the short-range bilayer scenario to the dipolar case. Finally, the estimated critical exponents appear to belong to the universality class of the corresponding bidimensional IDM.

We systematically investigate the dispersion relations of nondegenerate dark-bright-bright solitons in a three-component Manakov model with repulsive interactions. We show that the vector solitons possess four distinct dispersion relation branches, comprising two positive-mass branches and two negative-mass branches. The energy-velocity dispersion relation of each pair of positive- and negative-mass branches forms a closed loop, resulting in two disjoint loops for the soliton's overall dispersion. All soliton branches share a common maximum speed (much lower than the sound speed), at which branch degeneracies emerge. The maximum speed is determined by the larger bright soliton particle number for the nondegenerate dark-bright-bright soliton, in contrast to the partially degenerate case. Linear stability analysis shows that all these branches are stable against weak perturbations. Extending these results to the four-component Manakov model yields eight distinct nondegenerate soliton branches. Combining these results, we conclude that for an $N$-component Manakov system, the nondegenerate solitons have $2^{N-1}$ distinct branches, of which half are positive-mass branches and the other half are negative-mass branches. Each pair of positive- and negative-mass branches forms a closed dispersion relation loop, so that the vector solitons have $2^{N-2}$ disjoint loops. These results reveal rich degeneracy structures in soliton dispersion relations and may motivate experiments on externally driven vector-soliton dynamics.

Synaptic plasticity is the biological foundation of learning and memory and a key inspiration for efficient, spike-based neuromorphic computing. In both contexts, spike-timing-dependent plasticity (STDP) is a crucial adaptive rule. However, while analytical tools exist for spiking networks with fixed connections, they fail when STDP is introduced, creating a long-standing theoretical impasse. In this work, we demonstrate how mean-field theory can successfully bridge this gap. We derive the first McKean-Vlasov mean-field limit for a network of stochastic spiking units with discrete-state STDP synapses, providing a rigorous, low-dimensional description of a system previously considered analytically intractable due to strong heterogeneity and adaptive couplings. While motivated by neuroscience—introducing plasticity into a stochastic Wilson-Cowan model—our framework establishes a general theoretical tool for studying the collective dynamics of large systems of interacting spiking (as opposed to rate-based) units with adaptation. This result provides a new methodological avenue in statistical physics for analyzing a wide class of complex systems with plastic interactions.

Cell spreading is a fundamental biophysical process that regulates cellular responses to extracellular signals, influencing essential functions such as migration and polarization. To quantitatively investigate how geometric constraints govern spreading dynamics, we develop a particle-based computational framework that integrates cellular deformation, stochastic adhesion kinetics, and actin-driven traction. The framework employs smoothed dissipative particle dynamics to model the fluid dynamics of both suspending medium and cytoplasm, coarse-grained vertex modeling for membrane mechanics, receptor-ligand binding for substrate adhesion, and a Brownian ratchet mechanism for traction generation. We validate the model against experiments involving human dental pulp stem cells confined to circular, square, and triangular geometries, and the simulations quantitatively reproduce geometry-driven spreading dynamics. Our results reveal that cell spreading is driven toward a dynamic mechanical equilibrium, where cell-substrate adhesion balances substrate repulsion at their contact interface, and in-plane traction forces at the periphery are counterbalanced by the resultant in-plane adhesive and sidewall repulsive forces, while internal deformation forces remain self-balanced. Moreover, the confinement area scales adhesion and traction forces through formed bond and generated leading particles, while acute vertices amplify traction via curvature-enhanced cytoskeletal contractility. Consequently, triangular confinement induces vertex-directed spreading through perimeter-dependent adhesion and localized traction, whereas circular geometries preserve directional isotropy, yielding uniform radial spreading. The computationally efficient framework provides mechanistic insights into how spatial constraints translate to cellular responses, although limitations persist in modeling nuclear mechanics, detailed cytoskeletal dynamics, and biochemical signaling.

Magnetic Janus colloids whose permanent dipole moments are radially shifted from the particle center exhibit broken interaction symmetry that alters self-assembly in fully three-dimensional suspensions. Using Brownian dynamics simulations of dilute, field-free systems, we examine how the radial dipolar shift $s$ and the dipolar coupling strength $\lambda$ affect nucleation, growth kinetics, orientational ordering, and cluster morphology. Across the simulated parameter range, aggregation follows a two-stage nucleation-growth scenario: an initial delayed nucleation stage is followed by sustained growth, after which the weighted mean cluster size obeys a long-time power law, $\langle N_c\rangle \sim t^z$, where the exponent $z$ depends on both $s$ and $\lambda$. For small shifts, head-to-tail alignment dominates, producing extended chainlike and looped aggregates. Increasing $s$ frustrates collinear bonding, broadens the bonded-pair orientational distribution, and promotes closure, branching, and compactification. At the largest simulated shifts, the cluster-size distribution shifts toward small, compact aggregates with partially compensated dipole orientations, and the growth exponent decreases systematically, indicating hindered coarsening. Comparison with previously reported quasi-two-dimensional aggregation shows that $s$ and $\lambda$ remain the governing control parameters, while fully three-dimensional motion enables out-of-plane rearrangements, higher local coordination, and three-dimensional cluster-closure pathways. These results show that, within the simulated dilute and field-free regime, radial dipole displacement provides a geometric control parameter for tuning aggregation pathways and morphology selection in three-dimensional magnetic Janus suspensions.

The Landau-de Gennes phenomenological theory predicts the existence of a tricritical point in nematic liquid crystals with negative dielectric anisotropy, at which the nematic-isotropic phase transition changes from first order to second order under a sufficiently strong electric field. In this study, we successfully demonstrated the existence of a tricritical point by measuring the transmitted light intensity near the nematic-isotropic phase boundary under an applied electric field using a polarizing microscope and a temperature-gradient cell that allows position-to-temperature mapping with high accuracy. Furthermore, we measured the response of the transmitted light intensity to a superimposed ac electric field, and discussed the results on the basis of the phenomenology.

In this work, we propose and study an extension of the Ziff-Gulari-Barshad (ZGB) model for heterogeneous catalysis that incorporates sulfur (S) alongside the reactive gaseous species CO and ${\mathrm{O}}_2$. By accounting for the formation and desorption of ${\mathrm{SO}}_2$ and COS, this model introduces reactions that alter the system's dynamic behavior. Through Monte Carlo simulations and site mean-field (SMF) approximations, we characterize the steady-state phase diagram which now presents new S-rich absorbing phases. We introduce an aggregation factor in the SMF approach to recover the continuous phase transition that is typically absent in standard mean-field treatments. Our results indicate that the second-order transition line between the active and the pure O absorbing phase remains belonging to the directed percolation universality class. Notably, we observe a promotion effect where small concentrations of sulfur enhance ${\mathrm{CO}}_2$ production by approximately 5% when compared to the pure ZGB model. This phenomenon is attributed to the efficient creation of vacant sites by sulfur-oxygen reactions, which optimizes the surface coverage and prevents CO poisoning.

We propose a noninteger-dimensional spatial model for anisotropic solids incorporating a $q$-deformed derivative operator inspired by the Tsallis nonadditive entropy framework. This generalization provides a unified and flexible analytical formalism for exploring anisotropic thermal properties. We derive explicit expressions for the electronic and phonon densities of states and for the phonon specific heat, showing how the Hausdorff dimension $\alpha$ and the deformation parameter $q$ affect the spectral and thermodynamic response. The model is applied to several solid-state materials and provides a close description of their measured specific-heat behavior over broad temperature ranges. The fitted results illustrate the ability of the regularized entropic formulation to describe departures from the standard Debye response. Finally, we relate the $q$-deformation qualitatively to disorder-controlled conformable kinetics, suggesting a possible connection to microscopic heterogeneity and effective memory signatures without assuming a universal quantitative relation between $q$ and the kinetic exponent $\mu$.

Swarmalator systems intertwine two forms of collective behavior, swarming and synchronization, leading to the emergence of specific space-time patterns. Scaling the model to real-world phenomena and technical applications is problematic due to the assumption of global coupling among all swarmalators, which is impractical under physical constraints on interaction range. Conversely, purely local coupling was shown to be infeasible. To address this gap, we introduce and evaluate the concept of multihop coupling for swarmalators, which preserves the locality of physical interactions but propagates state information throughout the network via hop-limited and probabilistic flooding. It is demonstrated that convergence to the original emergent patterns can be achieved in a reliable and fast manner while keeping overhead low. A practical guideline for selecting the range, hop limit, and forwarding probability is provided. The range required for convergence can be approximated by the connectivity threshold of random geometric graphs.

In this work we study statistical properties of the extreme eigenstates of the randomly weighted adjacency matrices of random graphs. We focus on two random graph models: Erdős-Rényi graphs and random geometric graphs. Indeed, the adjacency matrices of both graph models are diluted versions of the Gaussian orthogonal ensemble (GOE) of random-matrix theory such that a transition from the Poisson ensemble (PE) to the GOE is observed by increasing the graph average degree $\langle k\rangle$. First, we write expressions for the spectral density in terms of $\langle k\rangle$ for the regimes below and above the percolation threshold. Then we show that the distributions of the largest ${\lambda }_1$ and second-largest ${\lambda }_2$ eigenvalues approach the Tracy-Widom distribution of type 1 for $\langle k\rangle \gg 1$, while $\langle {\lambda }_1\rangle =\sqrt{2\langle k\rangle }$. Additionally, we demonstrate that the distributions of the normalized distance between ${\lambda }_1$ and ${\lambda }_2$, the distribution of the ratio between higher consecutive eigenvalue spacings, and the distributions of the inverse participation ratios of the extreme eigenstates display a clear PE-to-GOE transition as a function of $\langle k\rangle$. We conclude that any of these distributions can be effectively used to probe the delocalization transition of the graph models without the need of the full spectrum.