Papers

Born effective charge, a fundamental quantity in lattice dynamics and ferroelectrics, provides a quantitative measure of linear polarization response to ionic displacements. However, it does not account for higher-order effects, which can play a significant role in certain materials, such as fluorite HfO$_2$. In this paper, we use the second-order mode effective charges defined with the second-order atomic dynamical charges as a measure of the dipole moments generated by nonpolar lattice distortions. Using first-principles calculations, we demonstrate that specific combinations of nonpolar phonons in many oxides can induce strongly aligned second-order polarizations, reaching magnitudes comparable to those of intrinsically polar modes even in the zero frequency limit, broadening the understanding of second-order effects, which have historically been emphasized for their dynamical effects at specific frequency ranges. Through a symmetry-based analysis of the charge density, we elucidate the microscopic origin of these effects, tracing them to variations in bond covalency and local electronic rearrangements. We also demonstrate large second-order mode effective charge in well-studied perovskites, highlighting the generality of these phenomena. Our results offer insights into the design principles of next-generation ferroelectric, piezoelectric and multifunctional materials from the higher-order contribution to polarization in crystalline solids.

We study the impact of electrical contact barriers in hybrid WSe$_2$-LiNbO$_3$-based acoustoelectric and acousto-photoelectric devices using a combination of scanning photocurrent and acousto-electric current spectroscopy. Static scanning photocurrent measurements provide a qualitative measure of the band-bending and spatial distribution of the Schottky barrier between semiconducting WSe$_2$ and gold electrodes whereas the surface acoustic wave-induced acousto-electric current reveals the height of the tunnelling barrier created by the van der Waals gap between the multilayered WSe$_2$ flake and the gold electrodes in addition to the Schottky barrier. The combination of both techniques shows a ten-fold increase in the photocurrent by the acoustic wave. Moreover, the observed spatial redistribution of the current is attributed to the interplay between the contact properties at the source and drain electrodes and the charge carrier dynamics induced by the surface acoustic wave. The ratio of acoustic wavelength to the electrical channel length is found to impact the SAW-induced charge carrier transport. For a channel length shorter than one acoustic wavelength, carriers undergo a seesaw-like motion which changes to charge conveyance for channel lengths comparable or exceeding the acoustic wavelength.

AlGaAs/GaAs coatings are being considered as coating candidates for gravitational-wave detectors. In this paper we investigate the birefringence properties of this crystalline semiconductor material by modulating the optical illumination on the mirror coating and monitoring the induced birefringence. While the measured low-frequency birefringence values align with previous studies, we observed a frequency-dependent behavior in the illumination-to-birefringence coupling, characterized by a pole increasing with illumination intensity and a gain at zero frequency (DC gain) decreasing with illumination intensity. We developed a generic theoretical model based on a master equation to characterize the measurement results by considering photon-induced electric fields and electro-optic effects. This model can fit the frequency and intensity dependencies of the induced birefringence. Additionally, this model predicts a generation-recombination noise (GR noise) will be observable in the coating birefringence. While the presented measurement cannot predict the exact level of GR noise, for the frequency band and spot sizes relevant for gravitational-wave detectors we expect GR noise to be white below the pole frequency, scale with power the same way laser shot noise does, and for fixed power be independent of spot size.

I define a notion of local causality LOC closely modelled on the Bell principle, construed as the condition that single case probabilities cannot be modified by actions at spacelike separation. The new principle, like that of Bell, forces Bell inequalities, but with two loopholes: one is retrocausation, known to Bell, but the other is nonuniqueness of remote outcomes, a loophole only for LOC, not for the Bell principle. I also set out a theory of physical probability, applicable to the Everett interpretation, in agreement with the Born rule, and therefore violating Bell inequalities. I show it is consistent with LOC. Surprisingly, both loopholes are exploited. I conclude not only that probability in the Everett interpretation involves no action at a distance, but that the observed violations of Bell inequalities is powerful evidence for many worlds.

A key aspect of learned partial differential equation (PDE) solvers is that the main cost often comes from generating training data with classical solvers rather than learning the model itself. Another is that there are clear axes of difficulty–e.g., more complex geometries and higher Reynolds numbers–along which problems become (1) harder for classical solvers and thus (2) more likely to benefit from neural speedups. Towards addressing this chicken-and-egg challenge, we study difficulty transfer on 2D incompressible Navier-Stokes, systematically varying task complexity along geometry (number and placement of obstacles), physics (Reynolds number), and their combination. Similar to how it is possible to spend compute to pre-train foundation models and improve their performance on downstream tasks, we find that by classically solving (analogously pre-generating) many low and medium difficulty examples and including them in the training set, it is possible to learn high-difficulty physics from far fewer samples. Furthermore, we show that by combining low and high difficulty data, we can spend 8.9x less compute on pre-generating a dataset to achieve the same error as using only high difficulty examples. Our results highlight that how we allocate classical-solver compute across difficulty levels is as important as how much we allocate overall, and suggest substantial gains from principled curation of pre-generated PDE data for neural solvers. Our code is available at https://github.com/Naman-Choudhary-AI-ML/pregenerating-pde

While direct numerical simulations (DNS) are the most accurate method for studying turbulence, their large computational cost restricts their use to idealized configurations and to Reynolds numbers well below those found in practical systems. A recently proposed method, Large/Small Eddy Simulation (L/SES), aims to overcome this limitation while still providing the solution fidelity comparable to that of DNS. L/SES represents a pair of coupled calculations: a lower-fidelity Large Eddy Simulation (LES), which captures the large-scale flow structure, and a high-fidelity Small-Eddy Simulation (SES) targeting a sub-region of interest of the LES, in which the small-scale dynamics is fully resolved. In this study, we demonstrate the accuracy and performance of L/SES in large Reynolds-number homogeneous isotropic turbulence (HIT) up to Taylor-scale Reynolds number approximately 600. Turbulence properties obtained with L/SES are shown to be in close agreement with the literature, both in terms of global characteristics, such as kinetic energy spectra and dissipative anomaly, as well as small-scale properties, such as higher-order moments of the velocity gradients up to the 10th order and probability density functions of the intermittent quantities. Also using simulations of HIT, we systematically investigate various method parameters and determine their optimal converged values. Finally, we discuss the computational cost of L/SES and demonstrate that it is approximately 3 orders of magnitude lower than for a traditional DNS at the highest Reynolds number considered here. This highlights the potential of L/SES as a discovery tool, which brings high-fidelity simulations of realistic flows into the realm of feasibility.

Cells within biological tissue are constantly subjected to dynamic mechanical forces. Measuring the internal stress of tissues has proven crucial for our understanding of the role of mechanical forces in fundamental biological processes like morphogenesis, collective migration, cell division or cell elimination and death. Previously, we have introduced Bayesian Inversion Stress Microscopy (BISM), which is relying on measuring cell-generated traction forces in vitro and has proven particularly useful to measure absolute stresses in confined cell monolayers. We further demonstrate the applicability and robustness of BISM across various experimental settings with different boundary conditions, ranging from confined tissues of arbitrary shape to monolayers composed of different cell types. Importantly, BISM does not require assumptions on cell rheology. Therefore, it can be applied to complex heterogeneous tissues consisting of different cell types, as long as they can be grown on a flat substrate. Finally, we compare BISM to other common stress measurement techniques using a coherent experimental setup, followed by a discussion on its limitations and further perspectives.

For several years, students at an art college, working with NASA astronomers, have produced animations inspired by research on black holes, dark matter and more. They can be whimsical or poetic but still constrained by scientific rigour. The animations are used for scientific outreach and are freely available. Our program received a positive assessment through an evaluation we undertook. We are now planning a mobile STEAM exhibition to engage teenagers from underrepresented communities who may not typically consider STE(A)M for their studies. "Science anxiety" has been reported to be a significant barrier to learning. Mixing animation with astronomy can stimulate interest in STEAM, making science engaging in an unconventional way. One component would be activities where participants create artistic responses to astronomy. We undertook a workshop at a local city-run school, specialising in the arts for ages 14-17, to brainstorm the art/science activities. There we gave short scientific presentations leading to art activities: a giant colouring wall with projected celestial phenomena, a stop-motion station, and colouring images of comet 67P to produce an animation. Surveys before and after the activities showed positive responses. The hand of the artist has long been an important concept in animation (Crafton 1991). In a film entitled "The Movements of the Universe", this concept is adapted to the hands of scientists. Combining animation, filmed interviews at NASA (including a Nobel prize winner), and the scientists' hands, bring unexpected feelings of dream and humour to the audience. In this paper we explore three different viewpoints of these activities from a scientist, an animator, and an animation student.

Considering the effects of glow discharge plasmas on plasma-facing materials and the applications such as coating, cleaning and surface treating, this work has been done to investigate the energy of ions of dc glow discharge plasmas. On the way towards this goal, a plasma chamber has been simulated via COMSOL Multiphysics software. Then, a gridded energy analyzer has been simulated and designed. In the next step, the analyzer has been constructed and tested to measure the energy of plasma ions. The devise contains a grid which is negatively biased to the same potential as the glow discharge cathode electrode. It discriminates plasma ions based on their energies, which are accelerated due to sheath potential drop before colliding with the cathode. The obtained energy distribution function from experiments has been compared to that of simulated plasma. The experimental results show that there are different groups of ions each in local thermal equilibrium in dc glow discharge plasmas.

Edge localized modes (ELMs) and runaway electrons (REs) pose significant challenges for all Tokamak devices and act as potent heat sources, potentially shortening the lifespan of plasma-facing materials (PFMs). These thermal loads can manifest in various detrimental effects, including melting, sputtering, cracking, blistering, and other forms of material degradation. While the ELMs are an intrinsic feature of H-mode operation in Tokamaks, runaway electrons pose a potential threat across all Tokamak device scales.In devices such as ITER, even with mitigation strategies, the ELMs can still impose considerable heat loads on the PFMs, reaching levels of approximately 1 MJ/m2. Various methods exist to experimentally simulate the heat load effects of ELMs on PFMs. In this study, the thermal loads from REs in small-scale Tokamaks are considered for this purpose.The presence of small-scale Tokamaks, exemplified by the Damavand experiment, facilitates the investigation of RE energy deposition. Analysis of the experimental data indicates that the REs populations generated by plasma instabilities within the Damavand Tokamak discharges exhibit heat densities on the order of MJ/m2 to 1 cm2 area PFMs. Furthermore, considering an average REs energy of 1 MeV and the prevailing discharge current, calculations indicate that the average total energy of the REs population per discharge is approximately 1 kJ, which is subsequently deposited on the Tokamak limiter. Considering the limiter's surface area and assuming that only 40% of the REs energy is transferred to it, the calculated heat load density reaches the order of MJ/m2, comparable to that observed during the ELMs events in large-scale fusion devices. This correspondence enables the Damavand Tokamak to investigate thermal loas effects of ELMs on PFMs in large-scale Tokamaks.

Nuclear fusion devices are constantly under the threat of malfunctions coming from the damages of plasma-facing materials due to being affected by thermal heat loads. The frequent heat loads during some transient events in large-scale Tokamaks have always been a great concern for researchers. In ITER, the heat load of GW/m2 is estimated to impose plasma-facing components during edge localized modes, beside the Tokamak steady state load which is about 20 MW/m2. Moreover, there are also other transient thermal loads occurring due to off-normal operation of ITER such as vertical displacement events or disruptions, at the orders of hundreds of MW/m2 and tens of GW/m2, respectively. These loads are great enough to result in severe damages of plasma-facing materials. In this study, the facture of tungsten material under the heat loads of Tokamaks is simulated and the results are presented.

The origin of the superior high-temperature strength of {\gamma}-TiAl with high Nb addition remains highly controversial, largely due to the unclear role of Nb atoms. Using large-scale hybrid Monte Carlo and molecular dynamics simulations with a self-developed neural network potential,we show that Nb atoms predominantly occupy Ti sites and form short-range order with neighboring Al atoms, but a non-negligible fraction also occupies Al sites (NbAl) and promotes the formation of antisite defects (TiAl). Both the NbAl and TiAl antisites exceptionally reduce stacking fault energies and facilitate deformation twinning, thereby enhancing plasticity. Meanwhile, these substitutional and antisite defects also increase the Peierls stress of both screw and edge dislocations, which hinders dislocation motion to cause pronounced solid-solution strengthening. This work provides mechanistic insights into the dual role of Nb in enhancing both strength and ductility in {\gamma}-TiAl and further offers guidance for defect and composition engineering in advanced alloy systems.

In the study of quantum chaos, `out of time ordered correlators' (OTOCs) are commonly used to quantify the rate at which quantum information is scrambled. This rate has been conjectured by Maldecena et al. to obey a universal, temperature dependent bound. Recent studies have shown that instantons, delocalised structures that dominate tunnelling statistics over barriers, reduce the growth rate of OTOCs. For the case of the symmetric double well, this reduction ensures the bound is maintained for OTOCs generated using ring polymer molecular dynamics (RPMD), a method with approximate dynamics but exact quantum statistics. In this report we set out to further understand the role of the instanton in the enforcement of the Maldacena bound and test whether RPMD is sufficient to satisfy the bound. We also investigate the impact of coherence on the flattening of of OTOCs by contrasting bounded with scattering systems. For the scattering system we observe a significantly smaller OTOC growth rate than that of the analogous bounded system, and a flattening in growth rate as time progresses. We attribute the first effect to influence of the Boltzmann operator, and the second to interference caused by anharmonicity of the potential. In our studies of RPMD, we find counterexamples showing that it is not sufficient to satisfy the bound. We develop a theory for OTOCs using (analytically-continued) Matsubara dynamics, revealing significantly different dynamical behaviour around the instanton compared to the predictions of RPMD. The instanton is found to be stationary in all coordinates but its collective angle $\Phi_0$, and fluctuations about it no longer resemble that of classical dynamics on a first order saddle as in RPMD.

Large animal groups – bird flocks, fish schools, insect swarms – are often assumed to form by gradual aggregation of sparsely distributed individuals. Using a mathematically precise framework based on time-varying directed interaction networks, we show that this widely held view is incomplete. The theory demonstrates that large moving groups do not arise by slow accumulation; instead, they emerge through the rapid merging of multiple pre-existing subgroups that are simultaneously activated under high-density conditions. The key mechanism is topological: the long-term interaction structure of any moving group contains a single dominant strongly connected component (SCC). This dominant SCC determines the collective velocity – both speed and direction – of the entire group. When two subgroups encounter one another, the trailing subgroup aligns with – and ultimately inherits – the velocity of the dominant SCC of the leading subgroup. Repeated merging events naturally generate large groups whose speed is predicted to be lower than the mean speed of the original subgroups. The same dynamics explain several universal empirical features: broad neighbour-distance distributions, directional asymmetry in neighbour selection, and the characteristic narrow-front, wide-rear geometry of real flocks. The framework yields testable predictions for STARFLAG-style 3D datasets, offering a unified explanation for the formation, maintenance, and geometry of coordinated animal groups.

The method of image charges is a powerful and elegant technique in electrostatics, commonly used to determine the electric field generated by point charges near conductors of various shapes. While standard problems focus on single charges interacting with conductors, the behavior of multipoles in such configurations has received comparatively less attention, particularly beyond the well-studied case of a flat plane. In this paper, we explore the image formation of electric dipoles and quadrupoles near a conducting sphere and uncover a wonderful result: the image of a given multipole is not necessarily of the same type. Instead, alongside the expected multipole image, the resulting image configuration also includes lower-order multipole contributions. This finding broadens the understanding of electrostatic images and offers new insights into the interaction of multipoles with conducting boundaries.

We perform direct numerical simulation of the incompressible Navier-Stokes equation with forcing at different spatial dimensions and measure turbulent and chaotic properties. Lyapunov exponents, $\lambda$, decrease with dimension, and $\lambda < 0$ for all simulations in six-dimensions up to $Re = 40$. These six-dimensional simulations display non-Gaussian statistics and other behavior similar to well developed turbulence despite their lack of chaos. Further, we find that small scale perturbations do not extend to the largest scales and that this terminal scale between correlation and decorrelation shrinks with dimension. We theorize that this change is related to the increased rate of vortex stretching. We find the interplay between turbulent and chaotic properties changes with increasing dimension.

The flow of a colloidal solution between two parallel disks rotating with the same angular velocity about two non-coincident axes was studied. The problem has been approached from two perspectives, the first wherein the stress is expressed in terms of a power-law of kinematical quantities, and the second wherein we consider a non-standard model where the symmetric part of the velocity gradient is given by a power-law of the stress. For a range of power-law exponents, the class of models are non-invertible. By varying the material and geometric parameters, changes in the flow behaviour at different Reynolds numbers were analysed. We find that pronounced boundary layers develop even at low Reynolds numbers based on the power-law exponents. The new class of stress power-law fluids and fluids that exhibit limiting stress also show the ability to develop boundary layers.

Robotic continuous control tasks impose stringent demands on the energy efficiency and latency of computing architectures due to their high-dimensional state spaces and real-time interaction requirements. Conventional electronic computing platforms face computational bottlenecks, whereas the fusion of photonic computing and spiking reinforcement learning (RL) offers a promising alternative. Here, we propose a novel computing architecture based on photonic spiking RL, which integrates the Twin Delayed Deep Deterministic policy gradient (TD3) algorithm with spiking neural network (SNN). The proposed architecture employs an optical-electronic hybrid computing paradigm wherein a silicon photonic Mach-Zehnder interferometer (MZI) chip executes linear matrix computations, while nonlinear spiking activations are performed in the electronic domain. Experimental validation on the Pendulum-v1 and HalfCheetah-v2 benchmarks demonstrates the system capability for software-hardware co-inference, achieving a control policy reward of 5831 on HalfCheetah-v2, a 23.33% reduction in convergence steps, and an action deviation below 2.2%. Notably, this work represents the first application of a programmable MZI photonic computing chip to robotic continuous control tasks, attaining an energy efficiency of 1.39 TOPS/W and an ultralow computational latency of 120 ps. Such performance underscores the promise of photonic spiking RL for real-time decision-making in autonomous and industrial robotic systems.

There exists a significant scale gap between photonic neural network integrated chips and neural networks, which hinders the deployment and application of photonic neural network. Here, we propose hardware-aware lightweight spiking neural networks (SNNs) architecture tailored to our photonic neuromorphic chips, and conducts hardware-software collaborative computing for solving patter classification tasks. Here, we employed a simplified Mach-Zehnder interferometer (MZI) mesh for performing linear computation, and 16-channel distributed feedback lasers with saturable absorber (DFB-SA) array for performing nonlinear spike activation. Both photonic neuromorphic chips based on the MZI mesh and DFB-SA array were designed, optimized and fabricated. Furthermore, we propose a lightweight spiking neural network (SNN) with discrete cosine transform to reduce input dimension and match the input/output ports number of the photonic neuromorphic chips. We demonstrated an end-to-end inference of an entire layer of the lightweight photonic SNN. The hardware-software collaborative inference accuracy is 90% and 80.5% for MNIST and Fashion-MNIST datasets, respectively. The energy efficiency is 1.39 TOPS/W for the MZI mesh, and is 987.65 GOPS/W for the DFB-SA array. The lightweight architecture and experimental demonstration address the challenge of scale mismatch between the photonic chip and SNN, paving the way for the hardware deployment of photonic SNNs.

This paper examines the linearized stability of plane Couette flow for stress-power law fluids, which exhibit non-monotonic stress-strain rate behavior. The constitutive model is derived from a thermodynamic framework using a non-convex rate of dissipation potential. Under velocity boundary conditions, the system may admit three steady-state solutions. Linearized stability analysis reveals that the two solutions on ascending constitutive branches are unconditionally stable, while the solution on the descending branch is unconditionally unstable. For mixed traction-velocity boundary conditions, the base state is unique. Stability depends solely on whether the prescribed traction lies on an ascending (stable) or descending (unstable) branch of the constitutive curve. The results demonstrate that flow stability in these complex fluids is fundamentally governed by both boundary conditions and constitutive non-monotonicity.

Vickers et al. (2025) established that 58.20% of astrobiology experts believe intelligent extraterrestrial life likely exists, providing the first empirical baseline for public comparison. We surveyed 6,114 highly educated and scientifically engaged individuals (77.60% bachelor's degree+; 67.99% high-to-very-high scientific engagement) to assess their beliefs about extraterrestrial intelligent life existence: (1) personal beliefs, (2) perceived social circle beliefs, (3) perceived expert beliefs, and (4) responses to expert consensus revelation. Results showed 95.01% believed extraterrestrial intelligent life exists, with 62.59% holding definitive rather than probable convictions. Participants exhibited massive pluralistic ignorance, a 'cosmic closet', underestimating social circle beliefs by 46.07 percentage points despite near-universal personal conviction. Participants also exhibited a novel 'conviction intensity gap': while overestimating expert belief prevalence (67.63% vs. 58.20% actual), they underestimated expert conviction strength, perceiving only 21.10% as holding definitive beliefs. Experimental revelation of actual consensus (N = 5,106; 83.51% passed manipulation check) produced negligible personal belief change (d = -0.11) and small social belief change (d = 0.14). These findings demonstrate that consensus misperception operates along two dimensions, prevalence and intensity, and that even scientifically engaged audiences resist belief revision via expert consensus information.

Machine learning (ML) models have recently been used to reconstruct electric field distributions from EFISH signal profiles-the 'inverse EFISH problem'. This addresses the line-of-sight EFISH inaccuracy caused by the Gouy phase shift in focused beams. A key benefit of this approach is that the accuracy of the reconstructed profile can be directly checked via a 'forward transform' of the EFISH equation. Motivated by this latest success, the present study introduces a novel ML model with markedly improved performance. Based on a more powerful operator-learning architecture, it goes beyond the ANNs and CNNs employed previously. Termed Decoder-DeepONet (DDON), its main strength is learning function-to-function mappings, essential for recovering electric field profiles of unknown shape. The superior performance of DDON is exemplified via a comparison with our published CNN model and the feasibility of a classical mathematical method, as well as its application to both discharge simulations and experimental EFISH data from a nanosecond pulsed discharge. In almost all cases, the DDON model exhibits better generalizability, higher prediction accuracy, and wider applicability. Furthermore, the intrinsic nature of this operator-learning architecture renders it less sensitive to the exact location(s) of the acquired data, enabling electric field reconstruction even with seemingly 'incomplete' input profiles–an issue often accompanying poor signal sensitivity. We also employ Integrated Gradients (IG) to identify the signal regions most critical to reconstruction accuracy, providing guidance on the optimal sampling window for EFISH acquisition. Overall, we believe that the DDON model is a robust and comprehensive model which can be readily applied to reconstruct 'bell-shaped' electric field profiles with an existing axis of symmetry, especially in non-equilibrium plasmas.

We investigate whether the structural connectivity of urban road networks helps explain dengue incidence in Recife, Brazil (2015–2024). For each neighborhood, we compute the average communicability curvature, a graph-theoretic measure capturing the ability of a locality to influence others through multiple network paths. We integrate this metric into Negative Binomial models, fixed-effects regressions, SAR/SAC spatial models, and a hierarchical INLA/BYM2 specification. Across all frameworks, curvature is the strongest and most stable predictor of dengue risk. In the BYM2 model, the structured spatial component collapses ($\phi \approx 0$), indicating that functional network connectivity explains nearly all spatial dependence typically attributed to adjacency-based CAR terms. The results show that dengue spread in Recife is driven less by geographic contiguity and more by network-mediated structural flows.

Quantum networking seeks to enable global entanglement distribution through terrestrial and free space channels; however, the exponential loss in these channels necessitates quantum repeaters with efficient, long lived quantum memories (QMs). Space based architectures, particularly satellite assisted links, offer a path to truly global connectivity, yet they demand QMs that are compatible with orbital factors such as infrared radiation and the unique challenges of operating aboard a satellite. In this work, we propose a multimode quantum memory (MMQM) for low Earth orbit (LEO) repeaters based on the atomic frequency comb (AFC) protocol. Our design integrates a hybrid alkali noble gas ensemble in an optical cavity, using alkali atoms for strong photon matter coupling and noble gas nuclear spins for minutes to hours coherence, all without the need for cryogenics. The architecture natively supports temporal and spectral multiplexing, enabling the storage of 100 modes to parallelize probabilistic operations and overcome light limited round trip times. Representative link budgets at $h = 500$ km with realistic apertures, $\eta_{\text{mem}}\gtrsim 70%$, and $t_{\text{buffer}}$ of several minutes predict improvements of up to two orders of magnitude in per pass success probability and instantaneous SKR relative to a memoryless dual downlink, with clear scaling in $N$. Our contributions are (i) a non cryogenic, space ready multimode memory, (ii) a systems analysis coupling mode count, storage time, and orbital geometry to achievable rate, and (iii) a near term implementation roadmap. Together, these results indicate feasibility with current to near term technology and provide a practical path toward a high rate, space enabled quantum internet.

Flagellated microswimmers B. Subtilis dispersed in a nematic phase of a lyotropic chromonic liquid crystal form a living liquid crystal (LLC). The combination of the passive and active components allows us to analyze how the active component transitions from the shear-imposed alignment into topological turbulence. The lateral extension of the experimental cell is 1000 times larger than the 10-micrometer thickness, to avoid the effect of lateral confinement on the dynamic patterns. The surface anchoring is azimuthally degenerate to avoid permanent anisotropy. After the shear cessation, active forces produced by the microswimmers trigger self-amplifying bend undulations. The amplitude of undulations grows with time and then saturates, while the wavelength increases only slightly. Strong bend stresses at the extrema of undulations are released by nucleating disclination pairs, which multiply to produce topological turbulence. The nucleating pairs and the symmetry axes of the +1/2 disclinations are orientationally ordered. In highly active LLCs, this alignment causes a nonmonotonous time dependence of the disclinations' number, as a fast-moving +1/2 disclination of one pair annihilates a -1/2 disclination of the neighboring pair. The spectra of elastic and kinetic energies exhibit distinct wavevector dependencies caused by the energy cost of the deformed passive nematic background. The study demonstrates how applied shears and a passive viscoelastic background affect the dynamics of active matter. Combined with the previously determined viscoelastic properties of the lyotropic chromonic liquid crystal, the results complete a comprehensive description of the experimentally assessable type of active matter.

Data assimilation (DA) is a cornerstone of scientific and engineering applications, combining model forecasts with sparse and noisy observations to estimate latent system states. Classical high-dimensional DA methods, such as the ensemble Kalman filter, rely on Gaussian approximations that are violated for complex dynamics or observation operators. To address this limitation, we introduce DAISI, a scalable filtering algorithm built on flow-based generative models that enables flexible probabilistic inference using data-driven priors. The core idea is to use a stationary, pre-trained generative prior that first incorporates forecast information through a novel inverse-sampling step, before assimilating observations via guidance-based conditional sampling. This allows us to leverage any forecasting model as part of the DA pipeline without having to retrain or fine-tune the generative prior at each assimilation step. Experiments on challenging nonlinear systems show that DAISI achieves accurate filtering results in regimes with sparse, noisy, and nonlinear observations where traditional methods struggle. The code for DAISI is available at https://github.com/Erik-Wikingsson/DAISI.

We present a simple variational framework for planar elastica that enables distributed energies, such as gravitational loading or magnetic body torques, to be incorporated in a modular and unified manner. The formulation is based on expressing all load induced contributions directly at the level of the energy functional, which avoids the force balance constructions used in classical treatments such as Wang (1986) and makes the inclusion of additional physical effects straightforward. The resulting planar energy functional yields compact governing equations in which the contributions of individual load types remain clearly separated. We demonstrate that the framework reproduces the classical heavy elastica equations exactly and naturally accommodates magnetic energy terms commonly used in hard magnetic rod models. Although mathematically elementary, the formulation provides a clean and extensible structure for describing planar rod deformations under general distributed loads.

Nitrogen vacancy (NV) centers in diamond have precipitated profound advances in microwave detection, manifesting themselves both in spatial resolution and sensitivity. However, typical methods based on Rabi oscillations are subject to detunings due to thermal and magnetic fluctuations and/or gradients, which introduce systematic errors and render the measurements susceptible to environmental perturbations. Here, we propose and demonstrate a novel approach for determining both the magnitude and direction of microwaves, by exploiting the spectral line broadening effect in the optically detected magnetic resonance of NV centers. This method eliminates the requirement of aligning the MW frequency to the spin transitions and is therefore immune to variations and inhomogeneities of the magnetic field and temperature, providing an optimal tool for fast imaging applications. With this method, we achieved wide-field imaging of near field microwaves generated with a microscale $\rm{\Omega}$-pattern antenna with a resolution of 800\,nm. Combining with the vector detection using multi-axis NVs, a full reconstruction of the vector microwave fields is obtained. Besides, our scheme also exhibits excellent linearity over a broad range of MW amplitudes, and the scale is theoretically calculated to be more than four orders. Our results augment the applicability of diamond-based microwave devices in applications under complex scenarios, especially where large dynamic range, fast test speed, and high spatial resolution are demanded.

This work explores the possibility of applying stochastic quantum mechanics to curved spacetimes, with an emphasis on the Schwarzschild black hole. After reviewing the fundamental concepts of this approach, the quantum stochastic equations are extended to curved spacetime using a fully covariant treatment. Subsequently, the Klein-Gordon equation is solved for scalar perturbations, and the resulting stochastic trajectories are analyzed by varying parameters such as angular momentum, particle frequency, and computational integration time. In conclusion, we find that the trajectories are influenced by gravitational fluctuations in spacetime and that, depending on the variation of the fundamental parameters, different types of stochastic trajectories are obtained.

Machine learning (ML) has shown significant promise in studying complex geophysical dynamical systems, including turbulence and climate processes. Such systems often display sensitive dependence on initial conditions, reflected in positive Lyapunov exponents, where even small perturbations in short-term forecasts can lead to large deviations in long-term outcomes. Thus, meaningful inference requires not only accurate short-term predictions, but also consistency with the system's long-term attractor that is captured by the marginal distribution of state variables. Existing approaches attempt to address this challenge by incorporating spatial and temporal dependence, but these strategies become impractical when data are extremely sparse. In this work, we show that prior knowledge of marginal distributions offers valuable complementary information to short-term observations, motivating a distribution-informed learning framework. We introduce a calibration algorithm based on normalization and the Kernelized Stein Discrepancy (KSD) to enhance ML predictions. The method here employs KSD within a reproducing kernel Hilbert space to calibrate model outputs, improving their fidelity to known physical distributions. This not only sharpens pointwise predictions but also enforces consistency with non-local statistical structures rooted in physical principles. Through synthetic experiments-spanning offline climatological CO2 fluxes and online quasi-geostrophic flow simulations-we demonstrate the robustness and broad utility of the proposed framework.