arXiv subjects

Mark Wilson

Publications and source records attributed to Mark Wilson.

14 recordsLinked to original sources

arXiv2026

Aligning LLM-Simulated and Human Examinees for Psychometric Calibration: A Cognitive Diagnostic Profiling Approach

Psychometric calibration for educational tests typically requires costly human response data. Large language models (LLMs) simulated examinees offer a promising route to early calibration, but their responses are too accurate and too uniform. We propose Cognitive Diagnostic Profiling (CDP), a zero-shot framework that prompts LLMs to simulate plausible examinees with diverse cognitive profiles: binary attribute-mastery patterns are rendered as natural-language profiles and sampled under an uninformative or an informative distribution. Using the Tatsuoka fraction-subtraction dataset (536 examinees, 15 items, five attributes), we evaluated eight LLM configurations under no-profile, uninformative-CDP, and informative-CDP conditions, assessing alignment with human examinees at the ability-distribution, mastery-profile, and item-difficulty levels. CDP improved all three levels: distributional overlap rose across configurations; weighted correlations between profile-level scores and human profile expectations reached 0.92 to 0.98; and item-difficulty recovery improved in rank order and absolute alignment, most for reasoning-enabled models; in the strongest case, Gemini 3.0 Flash (Thinking), one-parameter logistic (1PL) difficulty Spearman correlations rose from 0.24 to 0.86 and 0.90 and the root-mean-square error (RMSE) fell from 6.31 to 1.30 and 0.90; the informative condition helped most where profile-level alignment was strong. CDP brings LLM-simulated examinees into closer psychometric alignment with human examinees, making them practical for operational test development.

cs.CY↗

arXiv2026

Grain-boundary-mediated kinetic arrest in graphite-to-diamond transformation

The graphite-to-diamond transition exhibits striking variability under high-pressure, high-temperature (HPHT) conditions, producing diamond, graphitic phases, or metastable, mixed diamond-graphite nanocomposites despite similar synthesis conditions. Existing atomistic models, largely based on idealised single-crystal graphite, do not explain the persistence of partially transformed intermediate states under HPHT conditions. Here, using large-scale molecular dynamics simulations, we show that precursor grain structure governs graphite-to-diamond transformation pathways by decoupling diamond nucleation from cooperative transformation propagation. Grain boundaries first facilitate local sp$^3$ nucleation, after which diamond growth propagates within individual grains but becomes arrested at crystallographically mismatched grain boundaries. As a result, structurally heterogeneous graphite stabilizes kinetically arrested mixed sp$^2$-sp$^3$ states, whereas large or single-crystalline domains favour cooperative bulk transformation into diamond. Our findings identify structural heterogeneity as a missing control parameter alongside pressure and temperature, reframing metastable transformation products as kinetically trapped states arising from precursor microstructure rather than thermodynamic intermediates. Precursor crystallinity therefore emerges as a practical control parameter governing graphite-to-diamond transformation pathways.

cond-mat.mtrl-sci↗

arXiv2026

Mapping Multimodal Pilot Stress and Fatigue During Flight Sessions

This study analyzes patterns of stress and exhaustion among student pilots throughout flight training using a combination of physiological and self-reported measurements. The Perceived Stress Scale (PSS-10) was used to measure perceived stress and exhaustion before and after each flight, while physiological data, including heart rate (HR), electrodermal activity (EDA), skin temperature, and acceleration, were continuously recorded during flight sessions. To identify recurring patterns in arousal and workload, physiological signals were preprocessed and analyzed across the flight stages. The findings indicate a buildup of workload-related weariness over time, as evidenced by steady increases in EDA and skin temperature across flights, as well as post-flight increases in self-reported exhaustion. Heart rate responses were more event-specific, with brief spikes during high-demand phases of flight. Overall, the findings demonstrate the value of combining physiological signals with subjective reports to identify patterns of stress and fatigue during real-world flight training and highlight the potential of data-driven approaches for monitoring pilot well-being.

cs.HC↗

arXiv2024

Toward Mitigating Sex Bias in Pilot Trainees' Stress and Fatigue Modeling

While researchers have been trying to understand the stress and fatigue among pilots, especially pilot trainees, and to develop stress/fatigue models to automate the process of detecting stress/fatigue, they often do not consider biases such as sex in those models. However, in a critical profession like aviation, where the demographic distribution is disproportionately skewed to one sex, it is urgent to mitigate biases for fair and safe model predictions. In this work, we investigate the perceived stress/fatigue of 69 college students, including 40 pilot trainees with around 63% male. We construct models with decision trees first without bias mitigation and then with bias mitigation using a threshold optimizer with demographic parity and equalized odds constraints 30 times with random instances. Using bias mitigation, we achieve improvements of 88.31% (demographic parity difference) and 54.26% (equalized odds difference), which are also found to be statistically significant.

cs.LG↗

arXiv2022

Exploring the configurational space of amorphous graphene with machine-learned atomic energies

Two-dimensionally extended amorphous carbon ("amorphous graphene") is a prototype system for disorder in 2D, showing a rich and complex configurational space that is yet to be fully understood. Here we explore the nature of amorphous graphene with an atomistic machine-learning (ML) model. We create structural models by introducing defects into ordered graphene through Monte-Carlo bond switching, defining acceptance criteria using the machine-learned local, atomic energies associated with a defect, as well as the nearest-neighbor (NN) environments. We find that physically meaningful structural models arise from ML atomic energies in this way, ranging from continuous random networks to paracrystalline structures. Our results show that ML atomic energies can be used to guide Monte-Carlo structural searches in principle, and that their predictions of local stability can be linked to short- and medium-range order in amorphous graphene. We expect that the former point will be relevant more generally to the study of amorphous materials, and that the latter has wider implications for the interpretation of ML potential models.

physics.chem-ph↗

arXiv2019

Structural transitions in dense disordered silicon from quantum-accurate ultra-large-scale simulations

Structurally disordered materials continue to pose fundamental questions, including that of how different disordered phases ("polyamorphs") can coexist and transform from one to another. As a widely studied case, amorphous silicon (a-Si) forms a fourfold-coordinated, covalent random network at ambient conditions, but much higher-coordinated, metallic-like phases under pressure. However, a detailed mechanistic understanding of the liquid-amorphous and amorphous-amorphous transitions in silicon has been lacking, due to intrinsic limitations of even the most advanced experimental and computational techniques. Here, we show how machine-learning (ML)-driven simulations can break through this long-standing barrier, affording a comprehensive, quantum-accurate, and fully atomistic description of all relevant liquid and amorphous phases of silicon. Combining a model system size of 100,000 atoms (ten-nanometre length scale) with a prediction accuracy of a few meV per atom, our simulations reveal a remarkable, three-step transformation sequence for a-Si under increasing external pressure. First, up to 10-11 GPa, polyamorphic low- and high-density amorphous (LDA and HDA) regions are found to coexist, rather than appearing sequentially. Then, we observe a structural collapse into a distinct, very-high-density amorphous (VHDA) phase at 12-13 GPa, reminiscent of the dense liquid but being formed at a much lower temperature. Finally, our simulations indicate the transient nature of this VHDA phase: it rapidly nucleates crystallites at 13-16 GPa, ultimately leading to the formation of a poly-crystalline, simple-hexagonal structure, consistent with experiments but not seen in earlier simulations.

cond-mat.mtrl-sci↗

arXiv2017

Refining glass structure in two dimensions

Recently determined atomistic scale structures of near-two dimensional bilayers of vitreous silica (using scanning probe and electron microscopy) allow us to refine the experimentally determined coordinates to incorporate the known local chemistry more precisely. Further refinement is achieved by using classical potentials of varying complexity; one using harmonic potentials and the second employing an electrostatic description incorporating polarization effects. These are benchmarked against density functional calculations. Our main findings are that (a) there is a symmetry plane between the two disordered layers; a nice example of an emergent phenomenon, (b) the layers are slightly tilted so that the Si-O-Si angle between the two layers is not $180^{\circ}$ as originally thought but rather $175 \pm 2 ^{\circ}$ and (c) while interior areas that are not completely imagined can be reliably reconstructed, surface areas are more problematical. It is shown that small crystallites that appear are just as expected statistically in a continuous random network. This provides a good example of the value that can be added to disordered structures imaged at the atomic level by implementing computer refinement.

cond-mat.dis-nn↗

arXiv2014

Ring statistics of silica bilayers

The recent synthesis and characterization of bilayers of vitreous silica has produced valuable new information on ring sizes and distributions. In this paper, we compare the ring statistics of experimental samples with computer generated samples. The average ring size is fixed at six by topology, but the width, skewness and other moments of the distribution of ring edges are characteristics of particular samples. We examine the Aboav-Weaire law that quantifies the propensity of smaller rings to be adjacent to larger rings, and find similar results for available experimental samples which however differ somewhat from computer-generated bilayers currently. We introduce a new law for the areas of rings of various sizes.

cond-mat.mtrl-sci↗

arXiv2013

Modeling vitreous silica bilayers

We computer model a free-standing vitreous silica bilayer which has recently been synthesized and characterized experimentally in landmark work. Here we model the bilayer using a computer assembly procedure that starts from a single layer of amorphous graphene, generated using a bond switching algorithm from an initially crystalline graphene structure. Next each bond is decorated with an oxygen atom and the carbon atoms are relabeled as silicon. This monolayer can be now thought of as a two dimensional network of corner sharing triangles. Next each triangle is made into a tetrahedron, by raising the silicon atom above each triangle and adding an additional singly coordinated oxygen atom at the apex. The final step is to mirror reflect this layer to form a second layer and then attach the two layers together to form the bilayer. We show that this vitreous silica bilayer has the additional macroscopic degrees of freedom to easily form a network of identical corner sharing tetrahedra if there is a symmetry plane through the center of the bilayer going through the layer of oxygen ions that join the upper and lower layers. This has the consequence that the upper rings lie exactly above the lower rings, which are tilted in general. The assumption of a network of perfect corner sharing tetrahedra leads to a range of possible densities that we have previously characterized in three dimensional zeolites as a flexibility window. Finally, using a realistic potential, we have relaxed the bilayer to determine the density, and other structural characteristics such as the Si-Si pair distribution functions and the Si-O-Si bond angle distribution, which are compared to the experimental results obtained by direct imaging.

cond-mat.dis-nn↗

arXiv2011

The Aggregation Kinetics of a Simulated Telechelic Polymer

We investigate the aggregation kinetics of a simulated telechelic polymer gel. In the hybrid Molecular Dynamics (MD) / Monte Carlo (MC) algorithm, aggregates of associating end groups form and break according to MC rules, while the position of the polymers in space is dictated by MD. As a result, the aggregate sizes change every time step. In order to describe this aggregation process, we employ master equations. They define changes in the number of aggregates of a certain size in terms of reaction rates. These reaction rates indicate the likelihood that two aggregates combine to form a large one, or that a large aggregate splits into two smaller parts. The reaction rates are obtained from the simulations for a range of temperatures. Our results indicate that the rates are not only temperature dependent, but also a function of the sizes of the aggregates involved in the reaction. Using the measured rates, solutions to the master equations are shown to be stable and in agreement with the aggregate size distribution, as obtained directly from simulation data. Furthermore, we show how temperature induced variations in these rates give rise to the observed changes in the aggregate distribution that characterizes the sol-gel transition.

cond-mat.soft↗

arXiv2009

Asymptotic expansions of oscillatory integrals with complex phase

We consider saddle point integrals in d variables whose phase function is neither real nor purely imaginary. Results analogous to those for Laplace (real phase) and Fourier (imaginary phase) integrals hold whenever the phase function is analytic and nondegenerate. These results generalize what is well known for integrals of Laplace and Fourier type. The method is via contour shifting in complex d-space. This work is motivated by applications to asymptotic enumeration.

math.CO↗

arXiv2009

Topological changes at the jamming and gel transition of a reversible polymeric network

We investigate the network topologies of an ensemble of telechelic polymers. The telechelic polymers serve as links between nodes, which consist of aggregates of their telechelic endgroups. Our analysis shows that the degree distribution is bimodal and consists of two Poissonian distributions with different average degrees. The number of nodes in each of them as well as the distribution of links depends on temperature. By comparing the eigenvalue spectra of the simulated network gels with those of reconstructed networks, the most likely ttopology at each temperature is determined.

cond-mat.soft↗

arXiv2007

The Space Infrared Interferometric Telescope (SPIRIT): High-resolution imaging and spectroscopy in the far-infrared

We report results of a recently-completed pre-Formulation Phase study of SPIRIT, a candidate NASA Origins Probe mission. SPIRIT is a spatial and spectral interferometer with an operating wavelength range 25 - 400 microns. SPIRIT will provide sub-arcsecond resolution images and spectra with resolution R = 3000 in a 1 arcmin field of view to accomplish three primary scientific objectives: (1) Learn how planetary systems form from protostellar disks, and how they acquire their inhomogeneous composition; (2) characterize the family of extrasolar planetary systems by imaging the structure in debris disks to understand how and where planets of different types form; and (3) learn how high-redshift galaxies formed and merged to form the present-day population of galaxies. Observations with SPIRIT will be complementary to those of the James Webb Space Telescope and the ground-based Atacama Large Millimeter Array. All three observatories could be operational contemporaneously.

astro-ph↗

arXiv2004

Asymptotics of multivariate sequences, II: multiple points of the singular variety

We consider a multivariate generating function F(z), whose coefficients are indexed by d-tuples of nonnegative integers: F(z) = sum_r a_r z^r where z^r denotes the product of z_j^{r_j} over j = 1, ..., d. Suppose that F(z) is meromorphic in some neighborhood of the origin in complex d-space. Let V be the set where the denominator of F vanishes. Effective asymptotic expansions for the coefficients can be obtained by complex contour integration near points of V. In the first article in this series, we treated the case of smooth points of V. In this article we deal with multiple points of V. Our results show that the central limit (Ornstein-Zernike) behavior typical of the smooth case does not hold in the multiple point case. For example, when V has a multiple point singularity at the point (1, ..., 1), rather than a_r decaying on the order of |r|^{-1/2} as |r| goes to infinity, a_r is a polynomial plus a rapidly decaying term.

math.CO↗