Search

We propose a technique for performing deductive qualitative data analysis at scale on text-based data. Using a natural language processing technique known as text embeddings, we create vector-based representations of texts in a high-dimensional meaning space within which it is possible to quantify differences as vector distances. To apply the technique, we build off prior work that used topic modeling via Latent Dirichlet Allocation to thematically analyze 18 years of the Physics Education Research Conference proceedings literature. We first extend this analysis through 2023. Next, we create embeddings of all texts and, using representative articles from the 10 topics found by the LDA analysis, define centroids in the meaning space. We calculate the distances between every article and centroid and use the inverted, scaled distances between these centroids and articles to create an alternate topic model. We benchmark this model against the LDA model results and show that this embeddings model recovers most of the trends from that analysis. Finally, to illustrate the versatility of the method we define 8 new topic centroids derived from a review of the physics education research literature by Docktor and Mestre (2014) and re-analyze the literature using these researcher-defined topics. Based on these analyses, we critically discuss the features, uses, and limitations of this method and argue that it holds promise for flexible deductive qualitative analysis of a wide variety of text-based data that avoids many of the drawbacks inherent to prior NLP methods., Comment: 33 pages, 13 figures

Backpropagation-optimized artificial neural networks, while precise, lack robustness, leading to unforeseen behaviors that affect their safety. Biological neural systems do solve some of these issues already. Thus, understanding the biological mechanisms of robustness is an important step towards building trustworthy and safe systems. Unlike artificial models, biological neurons adjust connectivity based on neighboring cell activity. Robustness in neural representations is hypothesized to correlate with the smoothness of the encoding manifold. Recent work suggests power law covariance spectra, which were observed studying the primary visual cortex of mice, to be indicative of a balanced trade-off between accuracy and robustness in representations. Here, we show that unsupervised local learning models with winner takes all dynamics learn such power law representations, providing upcoming studies a mechanistic model with that characteristic. Our research aims to understand the interplay between geometry, spectral properties, robustness, and expressivity in neural representations. Hence, we study the link between representation smoothness and spectrum by using weight, Jacobian and spectral regularization while assessing performance and adversarial robustness. Our work serves as a foundation for future research into the mechanisms underlying power law spectra and optimally smooth encodings in both biological and artificial systems. The insights gained may elucidate the mechanisms that realize robust neural networks in mammalian brains and inform the development of more stable and reliable artificial systems.

This Letter introduces an approach for precisely designing surface friction properties using a conditional generative machine learning model, specifically a diffusion denoising probabilistic model (DDPM). We created a dataset of synthetic surfaces with frictional properties determined by molecular dynamics simulations, which trained the DDPM to predict surface structures from desired frictional outcomes. Unlike traditional trial-and-error and numerical optimization methods, our approach directly yields surface designs meeting specified frictional criteria with high accuracy and efficiency. This advancement in material surface engineering demonstrates the potential of machine learning in reducing the iterative nature of surface design processes. Our findings not only provide a new pathway for precise surface property tailoring but also suggest broader applications in material science where surface characteristics are critical.

Friction is the force resisting relative motion of objects. The force depends on material properties, loading conditions and external factors such as temperature and humidity, but also contact aging has been identified as a primary factor. Several aging mechanisms have been proposed, including increased "contact quantity" due to plastic or elastic creep and enhanced "contact quality" due to formation of strong interfacial bonds. While proposed mechanisms for frictional aging have been dependent upon the presence of a normal force, this factor is not a fundamental prerequisite for the occurrence of aging. In light of this, we present a novel demonstration of a substantial frictional aging effect within a cubic silicon carbide system, even when a normal force is entirely absent. Our observations indicate that the time-evolution of the frictional aging effect follows a logarithmic behavior, which is a pattern that has been previously observed in numerous other materials. To explain this behavior, we provide a derivation that is rooted in basic statistical mechanics, demonstrating that surface diffusion, a phenomenon that serves to minimize surface energy in the interface region, can account for the observed behavior. Upon application of a normal force, the friction force is enhanced owing to the presence of plastic creep. Although aging resulting from plastic and elastic creep is widely recognized and incorporated into most friction laws, diffusion-driven aging has received comparatively less attention. The ultimate objective is to develop or redesign friction laws by incorporating the microscopic behavior, with the potential to enhance their effectiveness.

This article reports on a study investigating how computational essays can be used to redistribute epistemic agency--cognitive control and responsibility over one's own learning--to students in higher education STEM. Computational essays are a genre of scientific writing that combine live, executable computer code with narrative text to present a computational model or analysis. The study took place across two contrasting university contexts: an interdisciplinary data science and modeling course at Michigan State University, USA, and a third-semester physics course at the University of Oslo, Norway. Over the course of a semester, computational essays were simultaneously and independently used in both courses, and comparable datasets of student artifacts and retrospective interviews were collected from both student populations. These data were analyzed using a framework which operationalized the construct of epistemic agency across the dimensions of programming, inquiry, data analysis and modeling, and communication. Based on this analysis, we argue that computational essays can be a useful tool in redistributing epistemic agency to students within higher education science due to their combination of adaptability and disciplinary authenticity. However, we also argue that educational contexts, scaffolding, expectations, and student backgrounds can constrain and influence the ways in which students choose to take up epistemic agency., Comment: Submitted to the Journal of Research in Science Teaching

This article describes a curricular innovation designed to help students experience authentic physics inquiry with an emphasis on computational modeling and scientific communication. The educational design centers on a new type of assignment called a computational essay, which was developed and implemented over the course of two semesters of an intermediate electricity and magnetism course at the University of Oslo, Norway. We describe the motivation, learning goals, and scaffolds used in the computational essay project, with the intention that other educators will be able to replicate and adapt our design. We also report on initial findings from this implementation, including key features of student-written computational essays, student reflections on the inquiry process, and self-reported conceptual and attitudinal development. Based on these findings, we argue that computational essays can serve a key role in introducing students to open-ended, inquiry-based work and setting the foundation for future computational research and studies., Comment: Submitted to the European Journal of Physics

In nature and experiments, a large variety of rupture speeds and front modes along frictional interfaces are observed. Here, we introduce a minimal model for the rupture of homogeneously loaded interfaces with velocity strengthening dynamic friction, containing only two dimensionless parameters; $\bar \tau$ which governs the prestress, and $\bar \alpha$ which is set by the dynamic viscosity. This model contains a large variety of front types, including slow fronts, sub-Rayleigh fronts, super-shear fronts, slip pulses, cracks, arresting fronts and fronts that alternate between arresting and propagating phases. Our results indicate that this wide range of front types is an inherent property of frictional systems with velocity strengthening branches.

When a rigid rough solid slides on a rigid rough surface, it experiences a random motion in the direction normal to the average contact plane. Here, through simulations of the separation at single-point contact between self-affine topographies, we characterize the statistical and spectral properties of this normal motion. In particular, its rms amplitude is much smaller than that of the equivalent roughness of the two topographies, and depends on the ratio of the slider's lateral size over a characteristic wavelength of the topography. In addition, due to the non-linearity of the sliding contact process, the normal motion's spectrum contains wavelengths smaller than the smallest wavelength present in the underlying topographies. We show that the statistical properties of the normal motion's amplitude are well captured by a simple analytic model based on the extreme value theory framework, extending its applicability to sliding-contact-related topics.

The rupture of dry frictional interfaces occurs through the propagation of fronts breaking the contacts at the interface. Recent experiments have shown that the velocities of these rupture fronts range from quasi-static velocities proportional to the external loading rate to velocities larger than the shear wave speed. The way system parameters influence front speed is still poorly understood. Here we study steady-state rupture propagation in a one-dimensional (1D) spring-block model of an extended frictional interface, for various friction laws. With the classical Amontons--Coulomb friction law, we derive a closed-form expression for the steady-state rupture velocity as a function of the interfacial shear stress just prior to rupture. We then consider an additional shear stiffness of the interface and show that the softer the interface, the slower the rupture fronts. We provide an approximate closed form expression for this effect. We finally show that adding a bulk viscosity on the relative motion of blocks accelerates steady-state rupture fronts and we give an approximate expression for this effect. We demonstrate that the 1D results are qualitatively valid in 2D. Our results provide insights into the qualitative role of various key parameters of a frictional interface on its rupture dynamics. They will be useful to better understand the many systems in which spring-block models have proved adequate, from friction to granular matter and earthquake dynamics., Comment: 16 pages, 10 figures, 1 table

The transition from stick to slip at a dry frictional interface occurs through the breaking of the junctions between the two contacting surfaces. Typically, interactions between the junctions through the bulk lead to rupture fronts propagating from weak and/or highly stressed regions, whose junctions break first. Experiments find rupture fronts ranging from quasi-static fronts with speeds proportional to external loading rates, via fronts much slower than the Rayleigh wave speed, and fronts that propagate near the Rayleigh wave speed, to fronts that travel faster than the shear wave speed. The mechanisms behind and selection between these fronts are still imperfectly understood. Here we perform simulations in an elastic 2D spring--block model where the frictional interaction between each interfacial block and the substrate arises from a set of junctions modeled explicitly. We find that a proportionality between material slip speed and rupture front speed, previously reported for slow fronts, actually holds across the full range of front speeds we observe. We revisit a mechanism for slow slip in the model and demonstrate that fast slip and fast fronts have a different, inertial origin. We highlight the long transients in front speed even in homogeneous interfaces, and we study how both the local shear to normal stress ratio and the local strength are involved in the selection of front type and front speed. Lastly, we introduce an experimentally accessible integrated measure of block slip history, the Gini coefficient, and demonstrate that in the model it is a good predictor of the history-dependent local static friction coefficient of the interface. These results will contribute both to building a physically-based classification of the various types of fronts and to identifying the important mechanisms involved in the selection of their propagation speed., Comment: 29 pages, 21 figures

The Casimir-Polder force is an important long range interaction involved in adsorption and desorption of molecules in fluids. We explore Casimir-Polder interactions between methane molecules in water, and between a molecule in water near SiO2 and hexane surfaces. Inclusion of the finite molecular size in the expression for the Casimir-Polder energy leads to estimates of the dispersion contribution to the binding energies between molecules and between one molecule and a planar surface., Comment: 6 pages, 5 figures

To study the microscopic origins of friction, we build a framework to describe the collective behaviour of a large number of individual micro-junctions forming a macroscopic frictional interface. Each micro-junction can switch in time between two states: A pinned state characterized by a displacement-dependent force, and a slipping state characterized by a time-dependent force. Instead of tracking each micro-junction individually, the state of the interface is described by two coupled distributions for (i) the stretching of pinned junctions and (ii) the time spent in the slipping state. We show how this framework represents an overarching structure for important models existing in the friction literature. We then use it to study systematically the effect of the time-scale that controls the duration of the slipping state. We first find the steady-state friction force as a function of the sliding velocity. As the framework allows for a whole family of micro-junction behaviour laws, we show how these laws can be chosen to obtain monotonic (strengthening or weakening) or non-monotonic velocity dependence at the macroscale. By then considering transient situations, we predict that the macroscopic static friction coefficient is strongly influenced by the way the interface was prepared, in particular by the slip dynamics of the previous sliding event. We also show that slow slip spontaneously occurs in the framework for a wide range of behaviour laws., Comment: 20 pages, 10 figures

The failure of the population of micro-junctions forming the frictional interface between two solids is central to fields ranging from biomechanics to seismology. This failure is mediated by the propagation along the interface of various types of rupture fronts, covering a wide range of velocities. Among them are so-called slow fronts, which are recently discovered fronts much slower than the materials' sound speeds. Despite intense modelling activity, the mechanisms underlying slow fronts remain elusive. Here, we introduce a multi-scale model capable of reproducing both the transition from fast to slow fronts in a single rupture event and the short-time slip dynamics observed in recent experiments. We identify slow slip immediately following the arrest of a fast front as a phenomenon sufficient for the front to propagate further at a much slower pace. Whether slow fronts are actually observed is controlled both by the interfacial stresses and by the width of the local distribution of forces among micro-junctions. Our results show that slow fronts are qualitatively different from faster fronts. Since the transition from fast to slow fronts is potentially as generic as slow slip, we anticipate that it might occur in the wide range of systems in which slow slip has been reported, including seismic faults., Comment: 35 pages, 5 primary figures, 6 supporting figures. Post-print version with improvements from review process included

Recovery of oil from oil shales and the natural primary migration of hydrocarbons are closely related processes that have received renewed interests in recent years because of the ever tightening supply of conventional hydrocarbons and the growing production of hydrocarbons from low permeability tight rocks. Quantitative models for conversion of kerogen into oil and gas and the timing of hydrocarbon generation have been well documented. However, lack of consensus about the kinetics of hydrocarbon formation in source rocks, expulsion timing and how the resulting hydrocarbons escape from or are retained in the source rocks motivates further investigation. In particular, many mechanisms for the transport of hydrocarbons from the source rocks in which they are generated into adjacent rocks with higher permeabilities and smaller capillary entry pressures have been proposed, and a better understanding of this complex process (primary migration) is needed. To characterize these processes it is imperative to use the latest technological advances. In this study, it is shown how insights into hydrocarbon migration in source rocks can be obtained by using sequential high resolution synchrotron X-ray tomography. Three-dimensional (3D) images of several immature "shale" samples were constructed at resolutions close to 5 micrometers. This is sufficient to resolve the source rock structure down to the grain level, but very fine grained silt particles, clay particles and colloids cannot be resolved. Samples used in this investigation came from the R-8 unit in the upper part of the Green River Shale, which is organic rich, varved, lacustrine marl formed in Eocene Lake Uinta, United States of America. One Green River Shale sample was heated in-situ up to 400{\deg}C as X-ray tomography images were recorded. The other samples were scanned before and after heating at 400{\deg}C. During the heating phase, the organic matter was decomposed, and gas was released. Gas expulsion from the low permeability shales was coupled with formation of microcracks. The main technical difficulty was numerical extraction of microcracks that have apertures in the 5 to 30 micrometer range (with 5 micrometers being the resolution limit) from a large 3D volume of X-ray attenuation data. The main goal of the work presented here is to develop a methodology to process these 3D data and image the cracks. This methodology is based on several levels of spatial filtering and automatic recognition of connected domains. Supportive petrographic and thermogravimetric data were an important complement to this study. An investigation of the strain field using two-dimensional image correlation analyses was also performed. As one application of the four-dimensional (4D, space + time) microtomography and the developed workflow, we show that fluid generation was accompanied by crack formation. Under different conditions, in the subsurface, this might provide paths for primary migration., Comment: 12 pages, 10 figures

We study the dynamic behaviour of 1D spring-block models of friction when the external loading is applied from a side, and not on all blocks like in the classical Burridge-Knopoff-like models. Such a change in the loading yields specific difficulties, both from numerical and physical viewpoints. To address some of these difficulties and clarify the precise role of a series of model parameters, we start with the minimalistic model by Maegawa et al. (Tribol. Lett. 38, 313, 2010) which was proposed to reproduce their experiments about precursors to frictional sliding in the stick-slip regime. By successively adding (i) an internal viscosity, (ii) an interfacial stiffness and (iii) an initial tangential force distribution at the interface, we manage to (i) avoid the model's unphysical stress fluctuations, (ii) avoid its unphysical dependence on the spatial resolution and (iii) improve its agreement with the experimental results, respectively. Based on the behaviour of this improved 1D model, we develop an analytical prediction for the length of precursors as a function of the applied tangential load. We also discuss the relationship between the microscopic and macroscopic friction coefficients in the model., Comment: 13 pages, 14 figures, accepted in Tribology Letters

We describe a 2D spring-block model for the transition from static to kinetic friction at an elastic slider/rigid substrate interface obeying a minimalistic friction law (Amontons-Coulomb). By using realistic boundary conditions, a number of previously unexplained experimental results on precursory micro-slip fronts are successfully reproduced. From the analysis of the interfacial stresses, we derive a prediction for the evolution of the precursor length as a function of the applied loads, as well as an approximate relationship between microscopic and macroscopic friction coefficients. We show that the stress build-up due to both elastic loading and micro-slip-related relaxations depend only weakly on the underlying shear crack propagation dynamics. Conversely, crack speed depends strongly on both the instantaneous stresses and the friction coefficients, through a non-trivial scaling parameter., Comment: 5 pages, 4 figures, accepted for publication in Physical Review Letters

To better understand the mechanisms of fracture pattern development and fluid escape in low permeability rocks, we performed time-resolved in situ X-ray tomography imaging to investigate the processes that occur during the slow heating (from 60\degree; to 400\degree;C) of organic-rich Green River shale. At about 350\degree;C cracks nucleated in the sample, and as the temperature continued to increase, these cracks propagated parallel to shale bedding and coalesced, thus cutting across the sample. Thermogravimetry and gas chromatography revealed that the fracturing occurring at ~350\degree;C was associated with significant mass loss and release of light hydrocarbons generated by the decomposition of immature organic matter. Kerogen decomposition is thought to cause an internal pressure build up sufficient to form cracks in the shale, thus providing pathways for the outgoing hydrocarbons. We show that a 2D numerical model based on this idea qualitatively reproduces the experimentally observed dynamics of crack nucleation, growth and coalescence, as well as the irregular outlines of the cracks. Our results provide a new description of fracture pattern formation in low permeability shales., Comment: 28 pages, 7 figures