arXiv subjects

Martin Schoen

Publications and source records attributed to Martin Schoen.

14 recordsLinked to original sources

arXiv2025

Leveraging Machine Learning for Early Detection of Lung Diseases

A combination of traditional image processing methods with advanced neural networks concretes a predictive and preventive healthcare paradigm. This study offers rapid, accurate, and non-invasive diagnostic solutions that can significantly impact patient outcomes, particularly in areas with limited access to radiologists and healthcare resources. In this project, deep learning methods apply in enhancing the diagnosis of respiratory diseases such as COVID-19, lung cancer, and pneumonia from chest x-rays. We trained and validated various neural network models, including CNNs, VGG16, InceptionV3, and EfficientNetB0, with high accuracy, precision, recall, and F1 scores to highlight the models' reliability and potential in real-world diagnostic applications.

eess.IV↗

arXiv2025

Stochastic Zeroth-Order Method for Computing Generalized Rayleigh Quotients

The maximization of the (generalized) Rayleigh quotient is a central problem in numerical linear algebra. Conventional algorithms for its computation typically rely on matrix-adjoint products, making them sensitive to errors arrissing from adjoint mismatches. To address this issue, we introduce a stochastic zeroth-order Riemannian algorithm that maximizes the generalized Rayleigh quotient without requiring adjoint or matrix inverse computations. We provide theoretical convergence guarantees showing that the iterates converge to the set of global maximizers of the (generalized) Rayleigh quotient and the norm of the Riemannian gradient vanishes at a sublinear rate with probability one. Our theoretical results are supported by numerical experiments, which demonstrate the excellent performance of the proposed method compared to state-of-the-art algorithms.

math.OC↗

arXiv2019

Structure and Dynamics of Solvated Polymers near a Silica Surface: On the Different Roles Played by Solvent

Whereas it is experimentally known that the inclusion of nanoparticles in hydrogels can lead to a mechanical reinforcement, a detailed molecular understanding of the adhesion mechanism is still lacking. Here we use coarse-grained molecular dynamics simulations to investigate the nature of the interface between silica surfaces and solvated polymers. We show how differences in the nature of the polymer and the polymer--solvent interactions can lead to drastically different behavior of the polymer--surface adhesion. Comparing explicit and implicit solvent models, we conclude that this effect cannot be fully described in an implicit solvent. We highlight the crucial role of polymer solvation for the adsorption of the polymer chain on the silica surface, the significant dynamics of polymer chains on the surface, and details of the modifications in the structure solvated polymer close to the interface.

cond-mat.soft↗

arXiv2015

The "magic" angle in the self-assembly of colloids suspended in a nematic host phase

Using extensive Monte Carlo (MC) simulations of colloids immersed in a nematic liquid crystal we compute an effective interaction potential via the local nematic director field and its associated order parameter. The effective potential consists of a local Landau-de Gennes (LdG) and a Frank elastic contribution. Molecular expressions for the LdG expansion coefficients are obtained via classical density functional theory (DFT). The DFT result for the LdG parameter $A$ is improved by locating the phase transition through finite-size scaling theory. We consider effective interactions between a pair of homogeneous colloids with Boojum defect topology. In particular, colloids attract each other if the angle between their center-of-mass distance vector and the far-field nematic director is about $30^{\circ}$ which settles a long-standing discrepancy between theory and experiment. Using the effective potential in two-dimensional MC simulations we show that self-assembled structures formed by the colloids are in excellent agreement with experimental data.

cond-mat.soft↗

arXiv2014

Experimental evidence of enhanced transport in supernematics

Deviations of molecular shapes from spherical symmetry may give rise to a variety of novel phenomena, including their dynamic behavior. It has recently been predicted [Mazza \textit{et al}. Phys. Rev. Lett. \textbf{105}, 227802 (2010)] that liquid crystals in the reentrant nematic phases may show unexpectedly high rates of translational displacements. Using broadband dielectric spectroscopy and a single-component thermotropic liquid crystal we explore molecular dynamics in the vicinity of the reentrant nematic transition and find the formation of a maximum in the mobilities upon changing temperature. The occurrence of the high mobility states are found to be history-dependent due to the formation of metastable mesophases on heating. The experimental results are further supported by computer simulation. Our results contribute to the clarification of an ongoing controversy on the dynamics of reentrant nematics.

cond-mat.soft↗

arXiv2012

Nonequilibrium domain formation by pressure fluctuations

Fluctuations in thermal many-particle systems reflect fundamental dynamical processes in both equilibrium and nonequilibrium (NEQ) physics. In NEQ systems \cite{ritort} fluctuations are important in a variety of contexts ranging from pattern formation \cite{hohenberg,vdbroeck} to molecular motors \cite{schaller,kierfeld,narayan,prost}. Here, we address the question if and how fluctuations may be employed to characterize and control pattern formation in NEQ nanoscopic systems. We report computer simulations of a liquid crystal system of prolate molecules (mesogens) sandwiched between flat walls, and exposed to a time-dependent external field. We find that a switchable smectic domain forms for sufficiently high frequency. Although pressure and temperature are too low to induce an equilibrium smectic phase, the fluctuations of the pressure in the NEQ steady state match the pressure fluctuations characteristic of the equilibrium smectic phase. Furthermore, the {\it wall-normal} pressure fluctuations give rise to a {\it tangential} "fluctuation-vorticity" tensor that specifies the symmetry-breaking direction of the smectic layers. Our calculations demonstrate a novel method through which nanomaterials with a high degree of molecular order may be manufactured in principle.

cond-mat.stat-mech↗

arXiv2011

Structure and dynamics of reentrant nematics: Any open questions after almost 40 years?

Liquid crystals have attracted enormous interest because of the variety of their phases and richness of their application. The interplay of general physical symmetries and specific molecular features generates a myriad of different phenomena. A surprising behavior of liquid crystals is the reentrancy of phases as temperature, pressure, or concentration are varied. Here, we review the main experimental facts and the different theoretical scenarios that have guided the understanding of bulk reentrant nematics. Recently, some computer simulations of a system confined to nanoscopic scales have found new dynamical features of the reentrant nematic phase. We discuss this prediction in relation with the available experimental evidence on reentrant nematics and with the dynamics of liquids in strongly confined environments.

cond-mat.stat-mech↗

arXiv2011

The role of string-like, supramolecular assemblies in reentrant supernematic liquid crystals

Using a combination of isothermal-isobaric Monte Carlo and microcanonical molecular dynamics we investigate the relation between structure and self-diffusion in various phases of a model liquid crystal using the Gay-Berne-Kihara potential. These molecules are confined to a mesoscopic slit-pore with atomically smooth substrate surfaces. As reported recently [see M. G. Mazza {\em et al.}, Phys. Rev. Lett. {\bf 105}, 227802 (2010)], a reentrant nematic (RN) phase may form at sufficiently high pressures/densities. This phase is characterized by a high degree of nematic order and a substantially enhanced self-diffusivity in the direction of the director $\hat{\bm{n}}$ which exceeds that of the lower-density nematic and an intermittent smectic A phase by about an order of magnitude. Here we demonstrate that the unique transport behavior in the RN phase may be linked to a confinement-induced packing effect which causes the formation of supramolecular, string-like conformations. The strings consist of several individual molecules that are capable of travelling in the direction of $\hat{\bm{n}}$ as individual "trains" consisting of chains of molecular "cars". Individual trains run in parallel and may pass each other at sufficiently high pressures.

cond-mat.soft↗

arXiv2010

Entropy-driven enhanced self-diffusion in confined reentrant supernematics

We present a molecular dynamics study of reentrant nematic phases using the Gay-Berne-Kihara model of a liquid crystal in nanoconfinement. At densities above those characteristic of smectic A phases, reentrant nematic phases form that are characterized by a large value of the nematic order parameter $S\simeq1$. Along the nematic director these "supernematic" phases exhibit a remarkably high self-diffusivity which exceeds that for ordinary, lower-density nematic phases by an order of magnitude. Enhancement of self-diffusivity is attributed to a decrease of rotational configurational entropy in confinement. Recent developments in the pulsed field gradient NMR technique are shown to provide favorable conditions for an experimental confirmation of our simulations.

cond-mat.soft↗

arXiv2010

Frustration of nanoconfined liquid crystals due to hybrid substrate anchoring

We present Monte Carlo simulations of liquid-crystalline material confined to a nanoscopic slit-pore. The simulations are carried out under isothermal conditions in a specialized isostress ensemble in which N fluid molecules are exposed to a compressional stress acting on the fluid in directions parallel with the substrate planes. Fluid-fluid and fluid-substrate interactions are modelled as in our previous work [M. Greschek et al., Soft Matter, 2010, DOI:10.1039/B924417D). We study several anchoring mechanisms at the solid substrate by introducing an anchoring function (0<g(u)<1) that depends on the orientation of a fluid molecule relative to the substrate plane; g(u) "switches" the fluid-substrate attraction on or off. Here we focus on various heterogeneous (i.e., hybrid) anchoring scenarios imposing different anchoring functions at the opposite substrates. As in our previous study we compute the isostress heat capacity which allows us to identify states at which the confined fluid undergoes a structural transformation. The isotropic-nematic transformation turns out to be nearly independent of the specific anchoring scenario. This is because the director in the nematic phase assumes a direction that is a compromise between the directions enforced by the competing anchoring scenarios at either substrate. On the contrary, at lower compressional stresses molecules prealign in specific directions that depend on details of the anchoring scenario.

cond-mat.soft↗

arXiv1999

Attractive forces between anisotropic inclusions in the membrane of a vesicle

The fluctuation-induced interaction between two rod-like, rigid inclusions in a fluid vesicle is studied by means of canonical ensemble Monte Carlo simulations. The vesicle membrane is represented by a triangulated network of hard spheres. Five rigidly connected hard spheres form rod-like inclusions that can leap between sites of the triangular network. Their effective interaction potential is computed as a function of mutual distance and angle of the inclusions. On account of the hard-core potential among these, the nature of the potential is purely entropic. Special precaution is taken to reduce lattice artifacts and the influence of finite-size effects due to the spherical geometry. Our results show that the effective potential is attractive and short-range compared with the rod length L. Its well depth is of the order of κ/10, where κis the bending modulus.

cond-mat.soft↗

arXiv1999

Absence of simulation evidence for critical depletion in slit-pores

Recent Monte Carlo simulation studies of a Lennard-Jones fluid confined to a mesoscopic slit-pore have reported evidence for ``critical depletion'' in the pore local number density near the liquid-vapour critical point. In this note we demonstrate that the observed depletion effect is in fact a simulation artifact arising from small systematic errors associated with the use of long range corrections for the potential truncation. Owing to the large near-critical compressibility, these errors lead to significant changes in the pore local number density. We suggest ways of avoiding similar problems in future studies of confined fluids.

cond-mat.stat-mech↗

arXiv1994

Pre-Galaxy Formation: A Non-Linear Analysis of the Evolution of Cosmological Perturbations

A higher-order analysis of the evolution of cosmological perturbations in a Friedman universe is given by using the PMF method. The essence of the PMF approach is to choose a gauge where all fluctuations of the density, the pressure, and the four-velocity vanish. In that gauge, even in higher orders, the perturbation field equations simplify considerably; they can be decoupled and - for simple equations of state - also be solved analytically. We give the solution for the dust universe up to third order. Comparison of these solutions strongly supports the conjecture that in general instable perturbations grow much faster than they do according to the first-order analysis. However, perturbations with very large spatial extension behave differently; they grow only moderatly. Thus, an upper boundary of the region of instability seems to exist.

gr-qc↗

arXiv1993

Operative Time Definition and Principal Uncertainity

Arguments are given that time must be defined in an operative manner,i.e., by constructing devices which can serve as clocks.The investigation of such devices leads to the conclusion that there is a principal uncertainity of time if one considers periods which are not large compared with the Planck time. Thus,according to the old (classical) concept,time cannot be well-defined at this scale.The uncertainity of time leads to a breakdown of Special and General relativity in the Planck regime;the same happens with causality. We present arguments that the classical concept of time,which treats t simply as a real parameter,must be replaced by a new one.

gr-qc↗