We present cmKAN, a versatile framework for color matching. Given an input image with colors from a source color distribution, our method effectively and accurately maps these colors to match a target color distribution in both supervised and unsupervised settings. Our framework leverages the spline capabilities of Kolmogorov-Arnold Networks (KANs) to model the color matching between source and target distributions. Specifically, we developed a hypernetwork that generates spatially varying weight maps to control the nonlinear splines of a KAN, enabling accurate color matching. As part of this work, we introduce a first large-scale dataset of paired images captured by two distinct cameras and evaluate the efficacy of our and existing methods in matching colors. We evaluated our approach across various color-matching tasks, including: (1) raw-to-raw mapping, where the source color distribution is in one camera's raw color space and the target in another camera's raw space; (2) raw-to-sRGB mapping, where the source color distribution is in a camera's raw space and the target is in the display sRGB space, emulating the color rendering of a camera ISP; and (3) sRGB-to-sRGB mapping, where the goal is to transfer colors from a source sRGB space (e.g., produced by a source camera ISP) to a target sRGB space (e.g., from a different camera ISP). The results show that our method outperforms existing approaches by 37.3% on average for supervised and unsupervised cases while remaining lightweight compared to other methods. The codes, dataset, and pre-trained models are available at: this https URL

A torus-shaped sail consists of a reflective membrane attached to an inflatable torus-shaped rim. The sail's deployment from its stowed configuration is initiated by introducing inflation pressure into the toroidal rim with an attached circular flat membrane coated by heat-sensitive materials that undergo thermal desorption (TD) from a solid to a gas phase. Our study of the deployment and acceleration of the sail is split into three steps: at a particular heliocentric distance a torus-shaped sail is deployed by a gas inflated into the toroidal rim and the membrane is kept flat by the pressure of the gas; under heating by solar radiation, the membrane coat undergoes TD and the sail is accelerated via TD of coating and solar radiation pressure (SRP); when TD ends, the sail utilizes thrust only from SRP. We study the stability of the torus-shaped sail and deflection and vibration of the flat membrane due to the acceleration by TD and SRP. The stability of the toroidal rim is addressed.

In traditional neural network architectures, a multilayer perceptron (MLP) is typically employed as a classification block following the feature extraction stage. However, the Kolmogorov-Arnold Network (KAN) presents a promising alternative to MLP, offering the potential to enhance prediction accuracy. In this paper, we propose the replacement of linear and convolutional layers of traditional networks with KAN-based counterparts. These modifications allowed us to significantly increase the per-pixel classification accuracy for hyperspectral remote-sensing images. We modified seven different neural network architectures for hyperspectral image classification and observed a substantial improvement in the classification accuracy across all the networks. The architectures considered in the paper include baseline MLP, state-of-the-art 1D (1DCNN) and 3D convolutional (two different 3DCNN, NM3DCNN), and transformer (SSFTT) architectures, as well as newly proposed M1DCNN. The greatest effect was achieved for convolutional networks working exclusively on spectral data, and the best classification quality was achieved using a KAN-based transformer architecture. All the experiments were conducted using seven openly available hyperspectral datasets. Our code is available at https://github.com/f-neumann77/HyperKAN.

Donut-shaped laser radiation, carrying orbital angular momentum, namely optical vortex, recently was shown to provide vectorial mass transfer, twisting transiently molten material and producing chiral micro-scale structures on surfaces of different bulk materials upon their resolidification. In this paper, we show for the first time that nanosecond laser vortices can produce chiral nanoneedles (nanojets) of variable size on thin films of such plasmonic materials, as silver and gold films, covering thermally insulating substrates. Main geometric parameters of the produced chiral nanojets, such as height and aspect ratio, were shown to be tunable in a wide range by varying metal film thickness, supporting substrates, and the optical size of the vortex beam. Donut-shaped vortex nanosecond laser pulses, carrying two vortices with opposite handedness, were demonstrated to produce two chiral nanojets twisted in opposite directions. The results provide new important insights into fundamental physics of the vectorial laser-beam assisted mass transfer in metal films and demonstrate the great potential of this technique for fast easy-to-implement fabrication of chiral plasmonic nanostructures.

We study double prompt $J/\psi$ hadroproduction within the nonrelativistic-QCD factorization formalism adopting the parton Reggeization approach to treat initial-state radiation in a gauge invariant and infrared-safe way. We present first predictions for the cross section distributions in the transverse momenta of the subleading $J/\psi$ meson and the $J/\psi$ pair. Already at leading order in $\alpha_s$, these predictions as well as those for the total cross section and its distributions in the invariant mass $m_{\psi\psi}$ and the rapidity separation $|Y|$ of the $J/\psi$ pair nicely agree with recent ATLAS and CMS measurements, except for the large-$m_{\psi\psi}$ and large-$|Y|$ regions, where the predictions substantially undershoot the data. In the latter regions, BFKL resummation is shown to enhance the cross sections by up to a factor of two and so to improve the description of the data.

Precision experiments, such as the search for electric dipole moments of charged particles using radiofrequency spin rotators in storage rings, demand for maintaining the exact spin resonance condition for several thousand seconds. Synchrotron oscillations in the stored beam modulate the spin tune of off-central particles, moving it off the perfect resonance condition set for central particles on the reference orbit. Here we report an analytic description of how synchrotron oscillations lead to non-exponential decoherence of the radiofrequency resonance driven up-down spin rotations. This non-exponential decoherence is shown to be accompanied by a nontrivial walk of the spin phase. We also comment on sensitivity of the decoherence rate to the harmonics of the radiofreqency spin rotator and a possibility to check predictions of decoherence-free magic energies.

Deep fake technology became a hot field of research in the last few years. Researchers investigate sophisticated Generative Adversarial Networks (GAN), autoencoders, and other approaches to establish precise and robust algorithms for face swapping. Achieved results show that the deep fake unsupervised synthesis task has problems in terms of the visual quality of generated data. These problems usually lead to high fake detection accuracy when an expert analyzes them. The first problem is that existing image-to-image approaches do not consider video domain specificity and frame-by-frame processing leads to face jittering and other clearly visible distortions. Another problem is the generated data resolution, which is low for many existing methods due to high computational complexity. The third problem appears when the source face has larger proportions (like bigger cheeks), and after replacement it becomes visible on the face border. Our main goal was to develop such an approach that could solve these problems and outperform existing solutions on a number of clue metrics. We introduce a new face swap pipeline that is based on FaceShifter architecture and fixes the problems stated above. With a new eye loss function, super-resolution block, and Gaussian-based face mask generation leads to improvements in quality which is confirmed during evaluation.

In the report we propose six new implementations of ruCLIP model trained on our 240M pairs. The accuracy results are compared with original CLIP model with Ru-En translation (OPUS-MT) on 16 datasets from different domains. Our best implementations outperform CLIP + OPUS-MT solution on most of the datasets in few-show and zero-shot tasks. In the report we briefly describe the implementations and concentrate on the conducted experiments. Inference execution time comparison is also presented in the report.

Laser irradiation of various materials including metals, polymers and semiconductors with vortex beams was previously shown to twist transiently molten matter providing the direct easy-to-implement way to obtain chiral surface relief. Specifically for metals, this effect was attributed to transfer of an optical angular momentum (OAM) carrying by the vortex beam. In this Letter, we report the formation of twisted metal nanoneedles on the surface of silver and gold metal films under their irradiation with zero-OAM laser beam having spiral-shape lateral intensity distribution. Our comparative experiments clearly demonstrate, for the first time, that the formation process of chiral nanoneedles on the surface of plasmonic-active metals is mainly governed by the temperature-gradient induced chiral thermocapillary mass transfer rather that the OAM driven rotation of the transiently molten matter.

A wide range of practically important problems is nowadays efficiently solved using artificial neural networks. This gave momentum to intensive development of their optical implementations, among which, the so-called diffractive neural networks (DNNs) constituted by a set of phase diffractive optical elements (DOEs) attract considerable research interest. In the practical implementation of DNNs, one of the standing problems is the requirement for high positioning accuracy of the DOEs. In this work, we address this problem and propose a method for the design of DNNs for image classification, which takes into account the positioning errors (transverse shifts) of the DNN elements. In the method, the error of solving the classification problem is represented by a functional depending on the phase functions of the DOEs and on random vectors describing their transverse shifts. The mathematical expectation of this functional is used as an error functional in the gradient method for calculating the DNN taking into account the transverse shifts of the DOEs. It is shown that the calculation of the derivatives of this functional corresponds to the DNN training method, in which the DOEs have random transverse shifts. Using the proposed gradient method, DNNs are designed that are robust to transverse shifts of the DOEs and enable solving the problem of classifying handwritten digits at a visible wavelength. Numerical simulations demonstrate good performance of the designed DNNs at transverse shifts of up to 17 wavelengths.

We study the optical properties of lowest-energy carbon allotropes in the infrared, visible and ultra-violet ranges of light in the general gradient approximation of the density functional theory. In our calculations we used the all-electron approach as well as the pseudo-potential approximation. In the infrared range, the complex dielectric functions, infrared and Raman spectra have been calculated using CRYSTAL14 program. The electronic properties and energy-dependent dielectric functions in the visible and ultraviolet ranges have been calculated using VASP program. We have described with a good accuracy experimentally known optical properties of cubic diamond crystal. Using obtained set of relevant parameters for calculations, we have predicted optical constants, dielectric functions and Raman spectra for the lowest-energy hypothetical carbon allotropes and lonsdaleite.

We theoretically describe the optical computation of the divergence of a two-dimensional vector field, which is composed by the transverse electric field components of an incident light beam. The divergence is computed in reflection at oblique incidence of light on a layered structure. We show that in the particular case of a linearly polarized incident beam, the layered structure implementing the divergence operator also allows one to compute the gradient and perform the isotropic differentiation. As an example of a layered structure computing the divergence, we propose a metal-dielectric multilayer consisting of two pairs of metal and dielectric layers on a metal substrate. The presented numerical simulation results of the designed multilayer confirm that the divergence operator is computed with high accuracy. We also demonstrate the possibility of using the designed structure for optical directional differentiation and computation of the gradient and Laplace operators.

We report a three-stage synthesis of a hybrid metal-carbon 2D material, in which cobalt atoms are covalently embedded in the graphene-like carbon (GLC) matrix. The resulting material (CoGLC) exhibits a distinctive XRD pattern indicative of the ordered arrangement of cobalt atoms in the layers. Furthermore, we demonstrate the fabrication of surfactant-free conductive inks from CoGLC via electrochemical exfoliation, making it a promising candidate for applications in in flexible electronics, spintronics or electrocatalysis.

The hadroproduction of single isolated photon at high energies is studied within the framework of the NLO${}^\star$ approximation of the Parton Reggeization Approach based on the modified Multi-Regge limit of the hard scattering QCD amplitudes. The contribution from the LO subprocess $Q \bar Q \to \gamma$ and the NLO tree-level corrections $Q R \to \gamma q$ and $Q \bar Q \to \gamma g$ are considered. To avoid specific double counting in rapidity between tree-level corrections to the hard scattering coefficient and unintegrated PDFs, a corresponding subtraction scheme is proposed. We demonstrated self-consistency of the Parton Reggeization Approach using the subtraction scheme. The results of calculations are compared with data from various Collaborations in the wide energy range $\sqrt{S} = 24 \ {\rm GeV} - 13 \ {\rm TeV}$. We obtained a quite satisfactory agreement with data up to $p_T^\gamma / \sqrt{S} \simeq 0.2 - 0.3$. Predictions for the future SPD NICA experiment were discussed.

Torsional Alfv\'en waves in coronal plasma loops are usually considered to be non-collective, i.e. consist of cylindrical surfaces evolving independently, which significantly complicates their detection in observations. This non-collective nature, however, can get modified in the nonlinear regime. To address this question, the propagation of nonlinear torsional Alfv\'en waves in straight magnetic flux tubes has been investigated numerically using the astrophysical MHD code Athena++ and analytically, to support numerical results, using the perturbation theory up to the second order. Numerical results have revealed that there is radially uniform induced density perturbation whose uniformity does not depend on the radial structure of the mother Alfv\'en wave. Our analysis showed that the ponderomotive force leads to the induction of the radial and axial velocity perturbations, while the mechanism for the density perturbation is provided by a non-equal elasticity of a magnetic flux tube in the radial and axial directions. The latter can be qualitatively understood by the interplay between the Alfv\'en wave perturbations, external medium, and the flux tube boundary conditions. The amplitude of these nonlinearly induced density perturbations is found to be determined by the amplitude of the Alfv\'en driver squared and the plasma parameter $\beta$. The existence of the collective and radially uniform density perturbation accompanying nonlinear torsional Alfv\'en waves could be considered as an additional observational signature of Alfv\'en waves in the upper layers of the solar atmosphere.

We theoretically demonstrate the possibility of generating a spatiotemporal optical vortex (STOV) beam in a dielectric slab waveguide. The STOV is generated upon reflection of a spatiotemporal optical pulse from an integrated metal-dielectric structure consisting of metal strips "buried" in the waveguide. For describing the interaction of the incident pulse with the integrated structure, we derive its "vectorial" spatiotemporal transfer function (TF) describing the transformation of the electromagnetic field components of the incident pulse. We show that if the TF of the structure corresponds to the TF of a spatiotemporal differentiator with a $\pi/2$ phase difference between the terms describing temporal and spatial differentiation, then the envelope of the reflected pulse will contain an STOV in all nonzero components of the electromagnetic field. The obtained theoretical results are in good agreement with the results of rigorous numerical simulation of the STOV generation using a three-strip metal-dielectric integrated structure. We believe that the presented results pave the way for the research and application of STOV beams in the on-chip geometry.

The Leding-Logarithmic (LL) gluon Sudakov formfactor is derived from rapidity-ordered BFKL evolution with longitudinal-momentum conservation. This derivation further clarifies the relation between High-Energy and TMD-factorizations and can be extended beyond LL-approximation as well.

In the article, we study associated production of prompt $J/\psi(\Upsilon)$ and $D-$mesons in the improved color evaporation model using the high-energy factorization approach as it is realized in the Monte-Carlo event generator KaTie. The modified Kimber-Martin-Ryskin-Watt model for unintegrated parton distribution functions is used. We predict cross sections for associated $J/\psi(\Upsilon)$ and $D-$meson hadroproduction via the single and double parton scattering mechanisms using the set of model parameters which has been fixed early for description of prompt single and pair heavy quarkonium production at the LHC energies. We found the results of calculations agree with the LHCb Collaboration data at the energies $\sqrt{s} = 7,8$ TeV and we present theoretical predictions for the energy $\sqrt{s} = 13$ TeV .

We study a large-$p_T$ three-photon production in proton-proton collisions at the LHC. We use the leading order (LO) approximation of the parton Reggeization approach consistently merged with the next-to-leading order corrections originated from the emission of additional jet. For numerical calculations we use the parton-level generator KaTie and modified KMR-type unintegrated parton distribution functions which satisfy exact normalization conditions for arbitrary $x$. We compare our prediction with data from ATLAS collaboration at the center-of-mass energy $\sqrt{s}=8$ TeV. We find that the inclusion of the real next-to-leading-order corrections leads to a good agreement between our predictions and data with the same accuracy as for the next-to-next-to-leading calculations based on the collinear parton model of QCD. At higher energies ($\sqrt{s}=13$ and 27 TeV) parton Reggeization approach predicts larger cross sections, up to $\sim 15$ \% and $\sim 30$ \%, respectively.