Institute for Informatics and Automation Problems

We performed extensive all-atom molecular dynamics simulations to investigate the interaction dynamics and orientation of levofloxacin, a newer fluoroquinolone antibiotic, with anionic (SDS) and cationic (CTAB) micelles. The maximum drug/micelle ratio (loading capacity and entrapment efficiency) was estimated for anionic and cationic micelles. High encapsulation efficiencies were observed for SDS (average: ∼80%). In contrast, for CTAB micelles, the efficiency was ∼8%, indicating that the binding of levofloxacin molecules to SDS micelles is significantly higher than that to CTAB micelles. Tilted orientations were observed for levofloxacin in SDS micelles (∼48–51°) and in CTAB micelles (∼40–42°), where the positively charged piperazine group is anchored to anionic headgroups. In contrast, the negatively charged carboxylic group is close to cationic headgroups. Calculating the relative binding energies, we found that levofloxacin binds more strongly to SDS than CTAB. Due to π–π interactions and hydrogen bonding, the formation of concerted columnar stacks of levofloxacin was also recorded for both anionic and cationic micelles.

Oleyl-capped nanoparticles have been used in nonpolar dispersions and can form ordered assemblies, for which an understanding of their interactions from a theoretical perspective is relevant. Long-range comprehensive molecular dynamics runs (1000 ns) are performed on oleyl-capped gold nanoclusters in different environments: hexane as a nonpolar solvent, and ethanol and pentanol as polar solvents. The molecular dynamics results in ethanol medium demonstrate that oleyl-capped nanoclusters tend to form attractive interactions with themselves via hydrocarbon interdigitation of their oleyl ligands, which leads to their aggregation. On the contrary, the attractive interactions between these nanoclusters are compensated by the interactions with hexane molecules, so that the nanoclusters keep separated from each other, leading to stable dispersions. The behavior of pentanol with the oleyl-capped nanoclusters indicates presumably more similarities to hexane, despite of being a polar solvent. These three simulation cases provide an insightful overview about the stabilization effects of oleyl-capped nanoparticles in organic solvents.

Autonomous driving systems usually depend on RGB images but face difficulties in low-visibility conditions. Thermal infrared (TIR) imaging offers an alternative that performs well in such environments. However, TIR is hindered by low contrast, indistinct boundaries, and the absence of color, making it less compatible with standard algorithms. Recent advances in thermal infrared colorization are promising, but they face challenges such as semantic distortions, frame instability, and issues with small objects. We introduce the frame-based Weber-law-driven Generative Adversarial Network (FWGAN), a new model that converts thermal images into visible-spectrum images. FWGAN improves object recognition by maintaining textures, increasing temporal consistency, and enabling reliable detection in different scenarios. Tests on benchmark datasets like FLIR and KAIST show that FWGAN's enhancements in semantic accuracy, object detection, and visual quality significantly push forward autonomous driving technology.

Distributed machine learning has become increasingly vital as data sources continue to expand geographically. Traditional ensemble methods such as AdaBoost demonstrate impressive predictive capabilities but often require frequent synchronization across nodes, resulting in significant communication overhead. This paper introduces a novel paradigm of infrequent synchronization in which nodes perform multiple rounds of local AdaBoost before exchanging partial or complete model updates. The potential advantages include reduced communication costs, the ability to handle intermittent connectivity, and competitive accuracy compared to fully synchronized approaches. A real-world use case in the trucking industry is presented to demonstrate the feasibility and value of this new approach. The paper concludes by outlining future directions and the expected impact on communication-efficient distributed learning.

Medical image segmentation is a critical task in healthcare, particularly for disease detection and proper treatment planning. Deep learning models achieve high performance in medical image analysis. This paper presents the effectiveness of the new PDE-UNet architecture, inspired by the applications of partial differential equations (PDEs) in neural networks, to enhance medical image segmentation performance. Experiments were conducted on brain tumor MRI images from the BraTS2020 dataset and compared with the traditional UNet architecture.

This paper provides a comprehensive review of our previously obtained results on exactly solvable lattice models, with a primary focus on the Abelian sandpile model, dimer model, loop-erased random walks and their connections to the enumeration of spanning trees.

A cap set in a projective or affine geometry over a finite field is a set of points no three of which are collinear.We construct complete cap sets with sizes 274432, 13991936, and 30294016 in affine geometries (, ), (, ), and (, ), respectively.

An acoustoplasma discharge is experimentally investigated when low-intensity acoustic oscillations are far from the intensity region where shock waves exist. The frequency range of 20–50 kHz is considered, when non-axial modes, rather than a plane longitudinal wave, propagate in the discharge tube. The discharge trajectory in the tube is shown to take the form of a planar sinusoid. The calculated and experimentally measured spectra of acoustic oscillations arising during discharge current modulation and sound dispersion for non-axial modes are presented. The dispersion is found to be significant. The spectrum of non-axial modes has a band structure. In acoustoplasma, the transition from one frequency band to the next is characterized by a phase transition. Within each band, dispersion can be either normal or anomalous, depending on the frequency. When moving from one band to the next, dispersion becomes anomalous.

Nowadays, X-ray sources are advanced tools used to probe the properties of matter and dynamic systems. Large scale accelerator based light source facilities have been constructed around the Globe to meet the scientific, technological and societal needs. Development of compact pocket-style X-ray systems meeting safety regulations is a complementary task requiring stable, reliable, and reproducible sources of X-rays with controllable parameters. In this research we employed a new approach for pyroelectric accelerator operation. Periodically varying temperature of LiTaO3 crystal enabled us to generate stable quasi-continuous X-ray beam. We managed to observe and characterise the avalanche process at pyroelectric effect. This phenomenon is responsible for multiplication of electrons at the anode while producing X-rays. We have demonstrated the great potential of this compact X-ray source by performing reliable energy-dispersive X-ray spectroscopy for element analysis of two solid samples.