Search

Magnet schools employ a variety of methods in order to enroll a diverse student body, such as targeted outreach, free and accessible transportation, encouraging choice across school districts, intentional school siting, and employing equitable lottery-based admissions policies. Many magnets also offer innovative programs around an attractive and relevant theme, like experiential learning, STEM, or fine arts. The U.S. Department of Education's 2024 Notice Inviting Applications for the Magnet Schools Assistance Program (MSAP) includes a new Competitive Preference Priority 5 (CPP5) that relates to Promoting Equity in Student Access to Educational Resources and Opportunities. CPP5 encourages applicants to "examin[e]…sources of inequity and inadequacy and implement responses" that help address these inequities. Importantly, CPP5 acknowledges the relationship between community segregation and school segregation, and encourages applicants to propose projects designed to tackle this longstanding challenge. Applicants are encouraged to propose projects designed to "increas[e] student racial or socioeconomic diversity, through developing or implementing evidencebased policies or strategies," specifically focused on: (1) interdistrict coordination; (2) cross-agency collaboration, such as with housing or transportation authorities; and/or (3) improvements to school assignment or admissions policies. This fact sheet provides ideas and related resources for applicants seeking to respond to CPP5's call for "cross-agency collaboration, such as with housing or transportation authorities."

Cloud computing has grown in importance in recent years which has led to a significant increase in Data Centre (DC) network requirements. A major driver of this change is virtualisation, which allows computing resources to be deployed on a large scale. However, traditional DCs, with their network topology and proliferation of network endpoints, are struggling to meet the flexible, centrally managed requirements of cloud computing applications. Software-Defined Networks (SDN) promise to offer a solution to these growing networking requirements by separating control functions from data routing. This shift adds more flexibility to networks but also introduces new security issues. This article presents a framework for evaluating security of SDN architectures. In addition, through an experimental study, we demonstrate how this framework can identify the threats and vulnerabilities, calculate their risks and severity, and provide the necessary measures to mitigate them. The proposed framework helps administrators to evaluate SDN security, address identified threats and meet network security requirements., Comment: pp. 66-91

In this study, we report a tetragonal perovskite structure of SrIrO$_3$ (P4/mmm, a = 3.9362(9) \r{A}, c = 7.880(3) \r{A}) synthesized at 6 GPa and 1400 $\deg$C, employing the ambient pressure monoclinic SrIrO$_3$ with distorted 6H structure as a precursor. The crystal structure of tetragonal SrIrO3 was evaluated on the basis of single crystal and powder X-ray diffraction. A cubic indexing was observed attributed to overlooked superlattice reflections. Weak fractional peaks in the H and K dimensions suggest possible structure modulation by oxygen defects. Magnetization study reveals weak paramagnetic behavior down to 2 K, indicative of the interplay between spin-orbit coupling, electron correlations, and crystal electric field. Additionally, measurements of electrical resistivity display metallic behavior with an upturn at about 54 K, ascribed to weak electron localization and possible structural defects., Comment: 31 pages, 12 figures

Data-driven, machine learning (ML) models of atomistic interactions are often based on flexible and non-physical functions that can relate nuanced aspects of atomic arrangements into predictions of energies and forces. As a result, these potentials are as good as the training data (usually results of so-called ab initio simulations) and we need to make sure that we have enough information for a model to become sufficiently accurate, reliable and transferable. The main challenge stems from the fact that descriptors of chemical environments are often sparse high-dimensional objects without a well-defined continuous metric. Therefore, it is rather unlikely that any ad hoc method of choosing training examples will be indiscriminate, and it will be easy to fall into the trap of confirmation bias, where the same narrow and biased sampling is used to generate train- and test- sets. We will demonstrate that classical concepts of statistical planning of experiments and optimal design can help to mitigate such problems at a relatively low computational cost. The key feature of the method we will investigate is that they allow us to assess the informativeness of data (how much we can improve the model by adding/swapping a training example) and verify if the training is feasible with the current set before obtaining any reference energies and forces -- a so-called off-line approach. In other words, we are focusing on an approach that is easy to implement and doesn't require sophisticated frameworks that involve automated access to high-performance computational (HPC).

Understanding the nature of irradiation damage often requires a multi-scale and multi-physics approach, i.e. it requires a significant amount of information from experiments, simulations and phenomenological models. This paper focuses on the initial stages of irradiation damage, namely neutron-induced displacement cascades in zirconium, as nuclear-grade zirconium alloys are widely used in fuel assemblies. We provide results of large-scale molecular dynamics (MD) simulations based on existing inter-atomic potentials and the two-temperature model to include the effect of electron-phonon coupling. Our data can be used directly in higher scale methods. Furthermore, we analysed summary statistics associated with defect production, such as the number of defects produced, their distribution and the size of clusters. As a result, we have developed a generative model of collision cascades. The model is hierarchical, as well as stochastic, i.e. it includes the variance of the considered features. This development had three main objectives: to establish a sufficient descriptor of a cascade, to develop an interpolator of data obtained from high-fidelity simulations, and to demonstrate that the statistical model of the data can generate representative distributions of primary irradiation defects. The results can be used to generate synthetic inputs for longer length- and time-scale models, as well as to build fast approximations relating dose, damage and irradiation conditions.

Imaging mass cytometry (IMC) is a relatively new technique for imaging biological tissue at subcellular resolution. In recent years, learning-based segmentation methods have enabled precise quantification of cell type and morphology, but typically rely on large datasets with fully annotated ground truth (GT) labels. This paper explores the effects of imperfect labels on learning-based segmentation models and evaluates the generalisability of these models to different tissue types. Our results show that removing 50% of cell annotations from GT masks only reduces the dice similarity coefficient (DSC) score to 0.874 (from 0.889 achieved by a model trained on fully annotated GT masks). This implies that annotation time can in fact be reduced by at least half without detrimentally affecting performance. Furthermore, training our single-tissue model on imperfect labels only decreases DSC by 0.031 on an unseen tissue type compared to its multi-tissue counterpart, with negligible qualitative differences in segmentation. Additionally, bootstrapping the worst-performing model (with 5% of cell annotations) a total of ten times improves its original DSC score of 0.720 to 0.829. These findings imply that less time and work can be put into the process of producing comparable segmentation models; this includes eliminating the need for multiple IMC tissue types during training, whilst also providing the potential for models with very few labels to improve on themselves. Source code is available on GitHub: https://github.com/kimberley/ISBI2024., Comment: International Symposium on Biomedical Imaging (ISBI) 2024 Submission

Microstructure is key to controlling and understanding the properties of metallic materials, but traditional approaches to describing microstructure capture only a small number of features. To enable data-centric approaches to materials discovery, allow efficient storage of microstructural data and assist in quality control in metals processing, we require more complete descriptors of microstructure. The concept of microstructural fingerprinting, using machine learning (ML) to develop quantitative, low-dimensional descriptors of microstructures, has recently attracted significant attention. However, it is difficult to interpret conclusions drawn by ML algorithms, which are commonly referred to as "black boxes". Here we explore variational autoencoders (VAEs), which can be trained to produce microstructural fingerprints in a continuous latent space. VAEs enable the reconstruction of images from fingerprints, allowing us to explore how key features of microstructure are encoded. We develop a VAE architecture based on ResNet18 and train it on Ti-6Al-4V optical micrographs as an example of an industrially important alloy where microstructural control is critical to performance. The latent space is explored in several ways, including by supplying interpolated and randomly perturbed fingerprints to the trained decoder and via dimensionality reduction to explore the distribution of microstructural features within the latent space of fingerprints. We show that the VAE fingerprints exhibit smooth, interpolable behaviour with stability to local perturbations, supporting their suitability as general purpose descriptors for microstructure. We also show that key properties of the microstructures are strongly correlated with position in the latent space, supporting the use of VAE fingerprints for quantitative exploration of process-structure-property relationships.

We used density-functional-theory simulations to examine the structural and electronic properties of the $\Sigma 180{\deg}(100)[001] $ grain boundary in monoclinic ZrO$_2$, which is a very low-energy (0.06Jm$^{-2}$) twin boundary present in experimental oxide texture maps, with suggested special properties. This equilibrium structure was compared with a metastable structure (with a boundary energy of 0.32Jm$^{-2}$), which was considered to be representative of a more general oxide boundary. The interaction of oxygen vacancies, substitutional Sn and Nb defects (substituting host Zr sites) with both structures - and their effect on the boundary properties - were examined. We found that the equilibrium structure energetically favours V$_\textrm{O}^{2+}$ and Nb$_\textrm{Zr}^{2-}$, whereas the metastable structure favours V$_\textrm{O}^{2+}$ and Sn$_\textrm{Zr}^{2-}$. Tin was further found to bind strongly with oxygen vacancies in both structures, and introduce gap states in the band gap of their electronic structure. Sn$_\textrm{Zr}^{2-}$ was, however, found to increase the segregation preference of V$_\textrm{O}^{2+}$ for the metastable structure, which might contribute to increased oxygen and electron transport down this interface, and therefore other more general boundaries, compared to the equilibrium structure of the studied monoclinic twin boundary.

The importance of cloud computing has grown over the last years, which resulted in a significant increase of Data Center (DC) network requirements. Virtualisation is one of the key drivers of that transformation and enables a massive deployment of computing resources, which exhausts server capacity limits. Furthermore, the increased network endpoints need to be handled dynamically and centrally to facilitate cloud computing functionalities. Traditional DCs barely satisfy those demands because of their inherent limitations based on the network topology. Software-Defined Networks (SDN) promise to meet the increasing network requirements for cloud applications by decoupling control functionalities from data forwarding. Although SDN solutions add more flexibility to DC networks, they also pose new vulnerabilities with a high impact due to the centralised architecture. In this paper we propose an evaluation framework for assessing the security level of SDN architectures in four different stages. Furthermore, we show in an experimental study, how the framework can be used for mapping SDN threats with associated vulnerabilities and necessary mitigations in conjunction with risk and impact classification. The proposed framework helps administrators to evaluate the network security level, to apply countermeasures for identified SDN threats, and to meet the networks security requirements., Comment: In Proceedings of the 13th International Conference on Cloud Computing and Services Science (CLOSER 2023). Prague, Czech Republic

Zirconium alloys are widely used as the fuel cladding material in pressurised water reactors, accumulating a significant population of defects and dislocations from exposure to neutrons. We present and interpret synchrotron microbeam X-ray diffraction measurements of proton-irradiated Zircaloy-4, where we identify a transient peak and the subsequent saturation of dislocation density as a function of exposure. This is explained by direct atomistic simulations showing that the observed variation of dislocation density as a function of dose is a natural result of the evolution of the dense defect and dislocation microstructure driven by the concurrent generation of defects and their subsequent stress-driven relaxation. In the dynamic equilibrium state of the material developing in the high dose limit, the defect content distribution of the population of dislocation loops, coexisting with the dislocation network, follows a power law with exponent $\alpha \approx 2.2$. This corresponds to the power law exponent of $\beta \approx 3.4$ for the distribution of loops as a function of their diameter that compares favourably with the experimentally measured values of $\beta$ in the range $ 3 \leq \beta \leq 4$.

Poverty & Race Research Action Council's (PRRAC's) March 2021 policy brief, "Mixed income neighborhoods and integrated schools: Linking HUD's Choice Neighborhoods Initiative with the Department of Education's Magnet Schools Assistance Program" (ED611507), highlighted an important opportunity for interagency collaboration, encouraging an explicit connection between HUD's Choice Neighborhoods Initiative (CNI) and the Department of Education's Magnet Schools Assistance Program (MSAP). As the authors advocated with both agencies, if funding connecting MSAP and public housing redevelopment were incentivized by both HUD and the Department of Education, applicants could create plans for their communities that include a vision for safe, accessible housing, as well as neighborhoods and schools with diversity of race and income. The purpose of this update is to determine: (1) whether HUD and the Department of Education incorporated any of these recommendations to encourage the connection of MSAP and CNI or other public housing redevelopment projects; and (2) whether applicants responded to such encouragement.

The Choice Neighborhoods Initiative (CNI), launched in 2010, provides funds for public housing authorities and other local entities and aims to redevelop distressed housing projects and transform the neighborhoods surrounding them into "mixed-income, high opportunity communities." From its inception, the CNI has included a strong focus on school improvement and coordination, although the responses of most applicants have been less than robust. Despite the CNI's emphasis on transforming neighborhoods of poverty into mixed income neighborhoods, this economic integration goal has rarely been applied to the local public school--indeed, with a few notable exceptions the U.S. Department of Housing and Urban Development (HUD) and CNI applicants have generally assumed that the local school will remain a high poverty school. This policy brief reviews the goals of the CNI and the track record of CNI school improvement efforts, as reflected in successful applications for CNI funding. The authors suggest that HUD and the Department of Education have missed a major opportunity for cross-agency collaboration, which could be achieved by linking magnet school funding and school integration planning grants with Choice Neighborhoods implementation and planning grants.

Finding efficient means of fingerprinting microstructural information is a critical step towards harnessing data-centric machine learning approaches. A statistical framework is systematically developed for compressed characterisation of a population of images, which includes some classical computer vision methods as special cases. The focus is on materials microstructure. The ultimate purpose is to rapidly fingerprint sample images in the context of various high-throughput design/make/test scenarios. This includes, but is not limited to, quantification of the disparity between microstructures for quality control, classifying microstructures, predicting materials properties from image data and identifying potential processing routes to engineer new materials with specific properties. Here, we consider microstructure classification and utilise the resulting features over a range of related machine learning tasks, namely supervised, semi-supervised, and unsupervised learning. The approach is applied to two distinct datasets to illustrate various aspects and some recommendations are made based on the findings. In particular, methods that leverage transfer learning with convolutional neural networks (CNNs), pretrained on the ImageNet dataset, are generally shown to outperform other methods. Additionally, dimensionality reduction of these CNN-based fingerprints is shown to have negligible impact on classification accuracy for the supervised learning approaches considered. In situations where there is a large dataset with only a handful of images labelled, graph-based label propagation to unlabelled data is shown to be favourable over discarding unlabelled data and performing supervised learning. In particular, label propagation by Poisson learning is shown to be highly effective at low label rates.

The population of dislocation defects in a crystalline material strongly influences its properties, so the ability to analyse this population in experimental samples is of great utility. As a complement to direct counting in the transmission electron microscope, quantitative analysis of x-ray diffraction line profiles is an important tool. This is an indirect approach to quantification and so requires careful validation of the physical models that underly the inferential process. Here we undertake to directly evaluate the ability of line profile analysis to quantify aspects of the dislocation and stacking fault populations by exploiting atomistic models of deformed copper single crystals. We directly analyse these models to determine exact details of the defect content (our "ground truth"). We then generate theoretical line profiles for the models and analyse them using the same procedures used in experimental analysis. This leads to inferred measures of the defect content which we are able to compare with the exact data. We show that line profile analysis is able to provide sound predictions of both dislocation density and stacking fault fraction across two orders of magnitude. We further show how the outer cut-off radius in the mean-square strain of a dislocation distribution invoked by Warren and Averbach corresponds to the cell size in an artificially constructed restrictedly-random distribution of dislocations according to the model of Wilkens. Overall, our results lend important new support to the use of line profile analysis for the quantification of line and planar defects in crystalline materials.

Quantitative measurements of extended defects in crystalline materials are important in understanding material behaviour. X-ray line profile analysis provides a complement to direct counting in the electron microscope, but is an indirect method and requires validation. Previous studies have focused on comparing x-ray analysis to electron microscopy results. Instead, we use simulated defective material with known defect content and apply line profile analysis to calculated diffraction profiles to directly show that line profile analysis can reliably quantify dislocations and stacking faults.