Papers

In the field of architecture, early detection of damage or degradation of building facades has become increasingly vital due to the need for continuous monitoring of structural integrity. Traditional methods, such as visual inspections, are being supplemented by technological advancements, especially in robotics, which offer innovative solutions for more efficient and precise inspections. This work focuses on the development of a five degree of freedom hybrid cable-driven parallel robot designed for vertical facade inspections. A detailed robot's design and CAD modeling, with a particular focus on a torque transmission mechanism that simplifies the motion of two cables using a single motor is presented. Two degrees of freedom are driven by cables, while the remaining three are driven by a Sarrus-type mechanism and a pan-tilt mechanism. The inverse kinematics models are also developed. A prototype is presented, involving additive manufacturing. A control system for tracking a zig-zag trajectory, commonly used in inspection tasks, was experimentally validated.

The emergence of generative artificial intelligence (GenAI) marks a potential inflection point in the way academic information is accessed, raising fundamental questions about the evolving role of search in student learning. This study examines this shift by analyzing longitudinal trends in physics-related search and page-view activity, using declines in traditional search behavior as a quantitative proxy for changes in independent information-seeking practices. We analyze Google Trends data for core concepts in Classical Mechanics and Electromagnetism across three academic years (2022-2025) in more than 20 countries, and complement this analysis with Wikipedia page-view data across seven major languages to establish platform independence. The results reveal a substantial, systematic, and persistent global decline in search and page-view activity across most examined physics topics. The magnitude of this decline is domain-dependent, with Mechanics-related content exhibiting sharper and more consistent reductions than Electromagnetism-related content. Pronounced geographic and linguistic heterogeneity is observed: while English-speaking regions show relative stability or only moderate declines, non-English-speaking regions exhibit substantially larger reductions in traditional, search-based information-seeking activity. Despite the overall decrease in volume, the seasonal structure characteristic of academic activity remains robust. Taken together, these findings indicate a redistribution of physics-related information-seeking behavior in academic contexts where generative tools are increasingly available.

We study large but finite neural networks that, in the thermodynamic limit, admit an exact low-dimensional mean-field description. We assume that the governing mean-field equations describing macroscopic quantities such as the mean firing rate or mean membrane potential are known, while their parameters are not. Moreover, only a single scalar macroscopic observable from the finite network is assumed to be measurable. Using time-series data of this observable, we infer the unknown parameters of the mean-field equations and reconstruct the dynamics of unobserved (hidden) macroscopic variables. Parameter estimation is carried out using the differential evolution algorithm. To remove the dependence of the loss function on the unknown initial conditions of the hidden variables, we synchronize the mean-field model with the finite network throughout the optimization process. We demonstrate the methodology on two networks of quadratic integrate-and-fire neurons: one exhibiting periodic collective oscillations and another displaying chaotic collective dynamics. In both cases, the parameters are recovered with relative errors below $1\%$ for network sizes exceeding 1000 neurons.

Optical computing provides a promising path toward energy-efficient machine learning, yet implementing nonlinear transformations without complex electronics or high-power sources remains challenging. Here, we demonstrate that continuous-wave multimode fiber laser cavities can function as nonlinear optical processors. Input images encoded as phase patterns on a spatial light modulator undergo high-dimensional transformation through the interplay of multimode interference and gain saturation dynamics. The cavity maps input data into spatially stable, class-separable intensity distributions, enabling a simple linear classifier to achieve accuracies of 85–99% across diverse benchmarks – including medical imaging and remote sensing – with orders of magnitude fewer trainable parameters than deep neural networks. Our results establish multimode fiber lasers as compact, low-power physical processors for scalable optical machine learning.

Cooperation is central to human societies and often unfolds within groups. Higher-order networks, such as hypergraphs, naturally represent these groups as hyperedges. Network structures and update rules, by which individuals revise their strategies, are the two fundamental components that shape the evolution of cooperation in structured populations. Yet while the effects of network structure have been studied extensively, update rules have been examined mostly through isolated models, leaving their relationships and the origins of their differing evolutionary outcomes poorly understood. Here we develop a unified framework for imitation dynamics on higher-order networks, parameterizing imitation-based update rules by the number of groups an individual samples and the number of peers consulted within each group. Under weak selection, we derive a closed-form condition for the success of cooperation in any multiplayer social dilemma on homogeneous hypergraphs, encompassing games with both linear and nonlinear payoff structures. The framework places previously disconnected update rules within a single family and reduces their effects on cooperation to one interpretable quantity, which we term information diversity. Update rules inducing higher information diversity promote cooperation more effectively, and we prove that this ordering holds strictly across the entire space of multiplayer social dilemmas. Simulations extend this principle to heterogeneous hypergraphs constructed both synthetically and from empirical data. Our framework provides a systematic way to represent, analyze, and compare update rules on higher-order networks, turning a fragmented collection of microscopic updating mechanisms into a tractable and interpretable theory.

This study presents the development of a compact gas-kinetic scheme using an arbitrary Lagrangian-Eulerian (ALE) formulation for structured meshes. Unlike the Eulerian formulation, the ALE approach effectively tracks flow discontinuities, such as shock waves and contact discontinuities. However, mesh motion alters the geometry and increases computational costs. To address this, two key strategies were introduced to reduce costs and enhance accuracy. The first strategy is to use the gas-kinetic scheme to construct a third-order gas-kinetic flux, rather than the Runge-Kutta method to achieve high-order time accuracy, which allows a single reconstruction and flux calculation per time step. This approach enables direct updates of both cell-averaged flow variables and their gradients using a time-accurate flux function, facilitating compact reconstruction. Second, the significant computational expense is spent on reconstruction, which requires recalculating the reconstruction matrix at each time step due to mesh changes. A simplified fourth-order compact reconstruction using a small matrix was used to mitigate this cost. The combination of fourth-order spatial reconstruction and third-order time-accurate flux evolution ensures both high resolution and computational efficiency in the ALE framework. The tests shows that the current reconstruction is 2.4x to 3.0x faster than the previous reconstruction. Additionally, a generalized ENO(GENO) method for handling discontinuities enhances the scheme's robustness. The numerical test cases, such as the Riemann problem, Sedov problem, Noh problem, and Saltzmann problem, demonstrated the robustness and accuracy of our method.

This study focuses on the modeling and dynamics of gravity-driven, axisymmetric thin liquid film flow along a conical surface. Spatial linear stability analysis is performed on the basis of a Benney-type equation derived for the present configuration. In particular, streamwise curvature of the free surface is found to exert a crucial influence on the stability threshold. For simulations of surface waves, a second-order low-dimensional model is developed under the long-wave assumption, achieving accuracy comparable to direct numerical simulations at far lower cost. With this model, the characteristics of both linear and nonlinear waves are examined. A key difference from the flow over a flat plate is the dependence of wave dynamics on radial distance from the cone apex. At relatively high flow rates, a transition from solitary to sinusoidal waves is observed, with the transition position correlating closely with the linear stability threshold. Within the parameter range investigated, quantitative results of the conical film flow are almost identical to those in the flat-plate case when local parameters are substituted, indicating that inertial effects of the conical geometry are negligible. The models and findings presented in this paper may aid the design and optimization of industrial processes such as film coating and liquid-film-based heat and mass transfer on conical surfaces.

Electromagnetically induced transparency (EIT), arising from quantum interference in coherently driven atomic systems, has inspired a variety of photonic analogues, such as coupled-resonator-induced transparency (CRIT) built on the quantum-state modelling using resonators. Although CRIT serves as a building block for slow light in photonic integrated circuits, recent advances in topological photonics motivate a further generalization of both EIT and CRIT using gauge-field degrees of freedom. Here, we propose generalized CRIT via a spinor representation with dual-channel gauge fields, enabling fully programmable CRIT featuring dynamical spectral engineering. We generalize the traditional EIT framework by introducing a spinor representation of bright- and dark-mode resonances, yielding a unified description of design parameters through universal unitary operations. Implementing a coupled-resonator building block that accesses the entire design space through dual-channel gauge fields, we demonstrate a programmable slow-light band in a one-dimensional CRIT lattice. These results address urgent needs in optical interconnects, such as tunable delay lines, reconfigurable synchronization, and linear frequency conversion.

Symmetry-driven wave physics in open systems, exemplified by parity-time (PT) symmetry, has extended the landscape of crystalline phases in materials science to include gain-loss media. Given the growing interest in engineering disorder for wave manipulation, such non-Hermitian crystals motivate the extension of non-Hermitian frameworks into the realm of correlated disorder. Here, we propose hyperuniformity and stealthiness in non-Hermitian systems as a generalization of PT-symmetric crystals to correlated disorder. We extend the scattering-microstructure correspondence to open systems, formulating non-Hermitian hyperuniformity and stealthiness that encompass their Hermitian counterparts. This approach, incorporating a statistical crystallography framework for non-Hermitian materials, demonstrates that real-imaginary cross-correlations of the material potential are irrelevant for achieving hyperuniformity but are essential for characterizing stealthiness, revealing unidirectional scattering phases that are inaccessible in Hermitian materials and in non-Hermitian crystals. By analysing the microstructural statistics of the resulting materials, our results, building on non-Hermitian wave physics, establish a connection to materials science, encompassing conventional descriptors of correlated disorder.

In an early theoretical work published in 1965, Belen'kii & Fradkin proposed a turbulent diffusivity model for Rayleigh–Taylor (RT) mixing. We review its derivation and present alternative arguments leading to the same final similarity equation. The original work then introduced an approximation that led to a simplified ordinary differential equation (ODE), which was used primarily to derive the important scaling result, $h \sim (\ln R)gt^2$. Here, we extend the analysis by examining the solutions to both the full similarity ODE and the simplified ODE in detail. It is shown that the full similarity equation captures many now well-known features of non-Boussinesq RT flows, including asymmetric spike and bubble growth and a systematic shift of velocity statistics toward the light-fluid side. Comparisons of the theoretical model with numerical and experimental studies show reasonable agreement in both spatial profiles and growth trends of mixing layer heights. We further show that a global mass correction applied to the simplified solution closely approximates the full solution, highlighting that, to leading order, RT mixing is governed by the competing dynamics between diffusion of $\ln \bar{\rho}$ and mass conservation.

The next-generation sensing and communication applications rely on high-frequency microwave generation with low-noise. The microwave photonic technology is promising by the practical application is limited by its complex architecture so far. Here, we demonstrate an optically locked low-noise photonic microwave oscillator, so that all the optical components are packaged within a small module of 166 mL, and low noise microwave generation is achieved at 10.4 GHz with single-sideband phase noise of -54 dBc/Hz at 10 Hz, -141 dBc/Hz at 10 kHz, and -162 dBc/Hz at 10 MHz offset. Above performance arises from a dual-laser self-injection-locking scheme to a single Fabry-Perot cavity with high Q exceeding 10^8, with over 20 dB common-mode noise suppression. The low-noise nature of such reference is coherently transferred to the X-band through a high-performance TFLN electro-optic comb chip, thereby overcoming long-standing barriers in photonic microwave integration to enable truly field-deployable low-noise microwave generation.

Dysprosium (Dy) is the most magnetic element on the periodic table, making it excellent for studying dipolar atom-atom interactions. We report on a simple Dy MOT that captures atoms directly from the thermal beam using a single diode-laser system to generate the light. Additionally, the atoms are magnetically confined by the quadrupole magnetic field that also facilitates the MOT. The MOT loading time is $\tau_\text{b} = 26~\text{ms}$. Atoms can decay to a dark state that is magnetically trapped. The time constant for loading into this magnetic trap is $\tau_\text{d} = 410~\text{ms}$. The total magnetically trapped population is $1.14\times10^{5}$ atoms, with $85\%$ residing in the dark states. The magnetically trapped atoms have a temperature of $28~\mu\text{K}$, significantly below the Doppler limit. This population fulfills the requirements for a range of future experiments.

Space-charge-limited (SCL) current density for time-invariant injection (under long-pulse condition) via the diode cathode is the maximum transportable density, while it can be leveraged higher when the injection pulselength becomes shorter than the transmit time for electrons (i.e., under short-pulse condition). However, both limits mentioned above apply for the time-invariant injection condition and the role of time-varying current density for injection remains elusive. In this paper, we numerically investigate the SCL electron flow with time-variant injection. Using particle-in-cell simulation, four different time-variant profiles for electron injection are enforced, and the maximum current densities are determined resulting from the space charge effect for various pulse lengths. We speculate that the time-variant density of injection via the diode cathode will contribute to transport enhancement in the short-pulse condition.

This paper provides a detailed and unified historical account of early debates over wave-function realism, the modern term for the view that the wave function of quantum theory is physically real. As this paper will show, the idea of physical waves associated with particles had its roots in work by Einstein and de Broglie, who both originally thought of these waves as propagating in three-dimensional physical space. De Broglie quickly turned this wave–particle duality into an early pilot-wave theory, on which a particle's associated “phase wave” piloted or guided the particle along its trajectory. Schrödinger built on de Broglie's phase-wave hypothesis to provide a comprehensive account of the nascent quantum theory. However, Schrödinger's new “undulatory mechanics” came at the cost of replacing de Broglie's phase waves propagating in physical space with a wave function propagating in a system's abstract configuration space. The present work will argue that this move from three-dimensional physical space to a many-dimensional configuration space was a decisive reason why the founders of quantum theory uniformly abandoned the physical reality of the wave function. This paper will further clarify that de Broglie introduced two distinct pilot-wave theories, and will then argue that it was Bohm's rediscovery of the second of these two pilot-wave theories over two decades later, as well as Bohm's vociferous defense of wave-function realism, that were responsible for resurrecting the idea of an ontological wave function. Finally, this paper will describe how the influence of Schrödinger and Bohmended up playing a central role in Everett's development of the “many worlds” interpretation.

In quantum theory, a weak value is a complex number with a somewhat technical definition: it is a ratio whose numerator is the matrix element of a self-adjoint operator and whose denominator is the inner product of a corresponding pair of state vectors. Weak values first appeared in the research literature in a pair of papers in 1987 and 1988, and were originally defined as the results of a special kind of experimental protocol involving non-disturbing measurements combined with an explicit form of post-selection. In the years since, subsequent papers on weak values have produced a number of important practical spin-offs, including new methods for signal amplification and quantum-state tomography. The present work is not concerned with those practical spin-offs, but with historical and ongoing attempts to assign weak values a transparent, single-system interpretation, as well as efforts that invoke weak values to make a number of exotic claims about the properties and behavior of individual quantum systems. This paper challenges these interpretational claims by arguing that they involve several forms of fallacious reasoning.

High precision portable and deployable frequency standards are required for modern navigation and communication technologies. Optical frequency standards are attractive for their improved stability over their microwave counterparts; however, increased complexities have anchored them in the laboratory. Sacrificing sensitivity of the most stable optical clocks has led to the recent development of deployable and portable optical frequency standards, leveraging hot atomic or molecular vapor. The short term limit for a majority of previous reports on two-photon rubidium standards is either the shot-noise or intermodulation limit hindering the one second fractional frequency stability to around $1\times10^{-13}/\sqrt{\tau}$. The answer for the shot-noise limit is to increase optical power and collected fluorescence, while the intermodulation limit solution requires improvements in laser linewidth, stimulated Brillouin scattering (SBS) lasers are known to reduce frequency noise, suppressing noise of the pump laser at high offset frequencies. We investigate an optical frequency standard based on the two-photon transition in $^{87}$Rb probed with a narrow linewidth photonic integrated circuit SBS laser with a quality factor over 130 million and instantaneous linewidth $<$ 10 Hz. The use of a narrow linewidth clock laser coupled with operating at higher optical intensities yields clock instabilities of $2\times10^{-14}$ at one second, currently the best reported short-term stability for a two-photon rubidium optical frequency standard.

The consistent embedding of Loop Quantum Gravity (LQG) effects within the Standard Model requires a rigorous understanding of how Planck-scale deformations manifest at low energies. While phenomenological approaches often introduce canonical deformations with multiple free parameters to capture these effects, the physical requirement of gauge symmetry in the framework of a covariant Effective Field Theory (EFT) imposes strict conditions on the allowed interaction structure. In this paper, we demonstrate that these conditions act as physical selection rules for admissible quantum gravity models. By applying non-Abelian Ward identities and a covariant operator mapping to the dimension-six operator basis, we derive a fundamental on-shell equivalence between kinetic and cubic interaction terms. As a paradigmatic application, we show that the Levy-Helayel-Neto (LHN) framework, a candidate effective description of LQG, satisfies this physical requirement only when its parameters obey the specific algebraic relation: theta_3 / theta_8 = -1/2 * [ 1 + theta_7 * (ell_P / L)^(2 + 2 Upsilon) ] + O(ell_P). It must be highlighted that this result advances the physical understanding of LQG phenomenology by revealing that the apparent freedom in defining the Hamiltonian is illusory; the parameters are bound by the necessity of preserving the gauge structure of the Standard Model.

Every generative model for crystalline materials harbors a critical structure size beyond which its outputs become unreliable; we call this the extrapolation frontier. Despite its consequences for nanomaterial design, this frontier has never been systematically measured. We introduce RADII, a radius-resolved benchmark of ~75,000 crystal-derived nanoparticle structures (33-11,298 atoms) that treats radius as a continuous scaling knob, tracing generation quality from in- to out-of-distribution under leakage-free splits. Each model is conditioned on target composition and atom count, isolating geometric extrapolation as the evaluation variable. RADII provides frontier-specific diagnostics: per-radius error profiles pinpoint each architecture's scaling ceiling, surface-interior decomposition separates boundary from bulk failures, and cross-metric sequencing reveals which aspect of structural fidelity breaks first. Benchmarking five state-of-the-art architectures, we find that: (i) well-behaved models degrade by ~13% in global positional error beyond training radii, while divergent models show poor fidelity across scales, with local bond fidelity ranging from negligible degradation to over 2x error growth; (ii) no two architectures share a failure sequence, revealing the frontier as a multi-dimensional surface shaped by model family; and (iii) well-behaved models follow the expected geometric scaling exponent alpha ~ 1/3, whose in-distribution fit predicts out-of-distribution error, making frontiers forecastable. Scaling MatterGen to its published parameter count stabilizes sampling but does not close the frontier, while DiffCSP remains unstable at published scale. These findings establish output scale as a first-class evaluation axis for geometric generative models. Code and data: https://github.com/KurbanIntelligenceLab/RADII.

We establish both local and global bifurcation results for traveling periodic solutions of the one-dimensional two-species Vlasov-Poisson equation. These solutions consist of strip-like regions of ions and electrons in phase space that propagate coherently and emerge from spatially homogeneous, velocity-dependent equilibrium layers. Depending on the geometry of the underlying equilibrium and on the selected Fourier mode, the bifurcation diagram exhibits either two or four solution branches. In all cases, the bifurcation is of pitchfork type; in symmetric configurations, the local structure near the equilibrium has a hyperbolic geometry. We further show that these locally constructed branches extend globally. This work extends the previous study by the same author of the purely electronic case, where the ions were modeled as an immobile neutralizing background. Allowing both species to evolve dynamically leads to a more intricate, higher-dimensional analysis. Finally, by means of an affine change of variables, we reveal a connection with the one-dimensional two-component Euler-Poisson system, which in turn enables the construction of traveling periodic waves of both small and large amplitude for that model as well.

Deep Brain Stimulation (DBS) is an effective treatment for neurological disorders but requires invasive surgery. This work presents a method for non-invasive DBS, based on microwave focusing of amplitude-modulated electric fields using an external antenna array of magnetic point dipoles. The proposed method combines iterative time reversal (iTR) and temporal interference (TI) optimization to jointly address electromagnetic field focusing and physiologically relevant neural stimulation. Antenna element positions, orientations, frequencies, amplitudes, and phases are optimized to localize stimulation within a target region. The method is evaluated in an anatomically realistic voxel head model with heterogeneous and lossy tissue properties. Systematic numerical studies, including perturbation analysis and statistical evaluation, demonstrate consistent spatial localization and robustness across all reported configurations. Safety is quantified using specific absorption rate (SAR), ensuring compliance with exposure limits. The study further provides insight into the influence of key parameters on field behavior and the associated trade-offs between focality, penetration, and safety in physiologically relevant stimulation. To the authors knowledge, this is the first study to combine iTR and TI optimization for microwave-based DBS in a realistic voxel head model, establishing a promising framework for safe non-invasive deep brain stimulation.

A new contextualised reading of Fritz Zwicky's 1933 article ”The redshift of extragalactic nebulae” about the virial analysis of the velocity dispersion of galaxies in the Coma cluster leads to a reconsideration of the traditional discourse on the introduction of dark matter. We argue that this component of matter was not only already on the stage of the scientific debates of the time, but also, in a more concealed form, played a central role in Zwicky's epistemic context. We thus reject the narration that dark matter is the result of a “naïve” astrophysical observation and emphasise the cosmological motivations that prompted Zwicky to presciently search for it. Moreover, with regard to its abundance, we argue that the discrepancy between the observed amount of luminous matter in the Coma Cluster and Zwicky's higher mass estimate derived from virial analysis was not, in fact, astonishing. What Zwicky described as a surprising excess of dark matter was of precisely the order of magnitude he had set out to identify. Consequently, we challenge the widespread view that dark matter was merely an ad hoc hypothesis introduced to rescue Newtonian theory. Instead, we suggest it may represent one of the earliest cosmological indications supporting a new emerging theory of gravitation: General Relativity. This reinterpretation contributes to ongoing debates in the philosophy of science concerning the epistemic status of ad hoc hypotheses.

This work is the second in a series focused on ferrofluid bend channel flows. Here, ferrofluid flows in bend channels are modeled using machine learning methods, based on data generated from the CFD simulation discussed in the first work in this series. Predicting convective heat transfer in ferrofluid flows influenced by magnetic fields is key to advancing thermal management in microscale and energy-intensive systems.

This study investigates ferrohydrodynamic heat transfer enhancement in a two-dimensional 90 degree bend channel through systematic parametric analysis of externally applied non-uniform magnetic fields, using Numerical CFD simulations.

Academic STEM evaluation can elicit anxiety, yet routine grading rarely captures how students semantically frame exams and wellbeing. We reconstruct these framings using behavioural forma mentis networks (BFMNs), that is, feature-rich networks of concepts linked by memory recalls and enriched with affective ratings and concreteness norms. We build BFMNs from 994 participants spanning STEM experts (N1 = 59), Italian high-schoolers (N2 = 206), physics undergraduates (N3 = 10), psychology undergraduates with math-anxiety levels (N4 = 301), and simulated students (N5 = 497) personified by a large language model (GPT-OSS 20B). Across all human groups, the concepts "exam" and "grade" were (i) perceived negatively, (ii) connected primarily to negatively valenced memory recalls, indicating a clustering of negative emotions around assessment, and (iii) framed through concepts eliciting fear and anticipation in most groups, including physics undergraduates (z-scores in the range [2.04, 2.53]). The semantic neighbourhoods of "anxiety" and "exam" overlapped three times more in human students than in GPT-based simulations, providing structural evidence of test anxiety in student populations. By contrast, experts displayed a neutral and more concrete network neighbourhood for "exam" (z = 1.87), with no clear trace of test anxiety. These negative assessment framings coexisted with positive representations of "wellbeing", which were rich in concrete associations in humans but linked to more abstract concepts in GPT digital twins. Overall, our results show that BFMNs offer a quantitative and interpretable framework to study academic anxiety and to distinguish human affective framing from current AI-based simulations

This study demonstrates the feasibility of a novel fabrication technique for high spatial resolution CsI: Tl scintillation detectors tailored for single photon emission computed tomography (SPECT) systems. Building upon our previously developed laser induced optical barrier (LIOB) method, which achieved high spatial resolution, excellent sensitivity, and 100% fabrication yield in CsI: Tl detectors, we extend this approach to a converging-pixel architecture. A CsI: Tl crystal array with converging pixels was designed and fabricated, featuring entrance-face pixels of 1.6x1.6 mm2 and photodetector side pixels of 2x2 mm2. To localize gamma-ray interactions, both the center of gravity (CoG) algorithm and a maximum-likelihood (ML) based decoding method were implemented. A custom built four axis motion platform was developed to deliver a finely collimated pencil beam at precisely controlled positions and angles across the array, enabling generation of a comprehensive dataset for prior knowledge and validation. The results demonstrate an energy resolution of 11.79+/-0.53% (collimated experiment) and a position localization accuracy of 1.00+/-0.42 mm (nearest neighbor interpolation), confirming that the proposed converging-pixel architecture, combined with statistical decoding algorithms, provides a promising path toward the development of high-performance SPECT detectors.

Accurate simulation of turbulent flows is fundamental to scientific and engineering applications. Direct numerical simulation (DNS) offers the highest fidelity but is computationally prohibitive, while existing data-driven alternatives struggle with stable long-horizon rollouts, physical consistency, and faithful simulation of small-scale structures. These challenges are particularly acute in three-dimensional (3D) settings, where the cubic growth of spatial degrees of freedom dramatically amplifies computational cost, memory demand, and the difficulty of capturing multi-scale interactions. To address these challenges, we propose a Physics-Enhanced Swin Transformer (PEST) for 3D turbulence simulation. PEST leverages a window-based self-attention mechanism to effectively model localized PDE interactions while maintaining computational efficiency. We introduce a frequency-domain adaptive loss that explicitly emphasizes small-scale structures, enabling more faithful simulation of high-frequency dynamics. To improve physical consistency, we incorporate Navier–Stokes residual constraints and divergence-free regularization directly into the learning objective. Extensive experiments on two representative turbulent flow configurations demonstrate that PEST achieves accurate, physically consistent, and stable autoregressive long-term simulations, outperforming existing data-driven baselines.

The wisdom of crowds has been shown to operate not only for factual judgments but also in matters of taste, where accuracy is defined relative to an individual's preferences. However, it remains unclear how different types of social signals should be selectively used in such domains. Focusing on a music preference dataset in which contributors provide both personal evaluations (Own) and estimates of population-level preferences (Estimated), we propose a routing framework for collective intelligence in taste. The framework specifies when contributors should speak, what they should report, and when silence is preferable. Using simulation-based aggregation, we show that prediction accuracy improves over an all-own baseline across a broad region of the parameter space, conditional on items where routing applies. Importantly, these gains arise only when silence is allowed, enabling second-order signals to function effectively. The results demonstrate that collective intelligence in matters of taste depends on principled signal routing rather than simple averaging.

Debris flows often exhibit coherent wave structures-shock-like roll waves on steeper slopes and weaker, more sinusoidal dispersive pulses on gentler slopes. Coarse-rich heads raise basal resistance, whereas fines-rich tails lower it; in gentle reaches, small-amplitude pulses can locally transport momentum across low-resistance segments. We focus on this gentle-slope, long-wave, low-amplitude regime, where the base-flow Froude number is order unity. In this limit, we obtain a Korteweg-de Vries (KdV) reduction from depth-averaged balances with frictional (Coulomb) and viscous-plastic basal options, using a curvature-type internal normal-stress closure in the long-wave small-k regime. Multiple-scale analysis yields effective nonlinear and dispersive coefficients. We also introduce a practical nonlinearity diagnostic that can be computed from observed crest speeds and flow thicknesses. When laboratory-frame crest celerity is available, we estimate an effective quadratic coefficient from the KdV speed-amplitude relation and report its ratio to the shallow-water reference. When only a depth-averaged first-surge speed and thickness are available, we use the same construction to form a velocity-based proxy and note its bias near zero. A Froude-slope diagram organizes published cases into a steep-slope roll-wave domain and a gentle-slope corridor where KdV pulses are admissible. Numerical solutions of the full depth-averaged model produce cnoidal and solitary waves that agree with the reduced KdV predictions within this corridor. We regard dispersive pulses as a regime-specific complement to roll-wave dynamics, offering a condition-dependent contribution to mobility on gentle reaches rather than a universal explanation for long runout.

We present a large scale data analysis of Moltbook, a Reddit-style social media platform exclusively populated by AI agents. Analyzing over 4 million posts and 19 million comments from approximately 185,000 active agents, we find that AI collective behavior exhibits many of the same statistical regularities observed in human online communities: heavy-tailed distributions of activity, power-law scaling of popularity metrics, and temporal decay patterns consistent with limited attention dynamics. However, we also identify key differences, including a sublinear relationship between upvotes and discussion size that contrasts with human behavior. These findings suggest that, while individual AI agents may differ fundamentally from humans, their emergent collective dynamics share structural similarities with human social systems.

We consider the extended Korteweg-de Vries (eKdV) equation as a model for long moderately nonlinear surface water waves and use it to describe the evolution of initial conditions generating solitary waves with and without significant dispersive radiation, as well as cases of pure dispersive radiation without any solitary waves. In the slow time formulation for the modelled solutions this equation also generates fast propagating resonant forward radiation due to the non-convexity of its linear dispersion curve, which is not present in the direct numerical simulations of the strongly nonlinear Serre parent system (also known as the Su-Gardner and Green-Naghdi equations). We show that the extended KdV-Whitham approximation and the slow space formulation of the eKdV equation are suitable regularisations of the eKdV equation in several cases of interest. Importantly, unlike the KdV-type equations, it can be used to model waves of moderate amplitude. Numerical comparisons are made between the Serre system and several respective reduced models, where simulations are initiated with an approximate soliton solution of the eKdV equation, constructed by use of Kodama-Fokas-Liu near-identity transformation to the KdV equation, as well as a generic localised initial condition.