Search

Determining whether a state's young people are on track for a life of opportunity is a difficult task for governors and state leaders. States can be both awash in data and unable to easily access and use that data to inform policy. State longitudinal data systems that meaningfully connect workforce, higher education, K-12, and early childhood outcome data from multiple state agencies over time are foundational to strong decision-making by state leaders. State leaders are charged with determining how to advance educational and workforce opportunities, particularly for historically underserved populations, including rural students, students living in poverty, English-language learners, and students of color. Policymakers need reliable, relevant, and transparent data from PK-12 to the workforce to inform decisions and allocate resources. Researchers need access to data to assess current conditions and to identify possible solutions. Finally, the public -- including parents -- needs easily accessible, user-friendly information about readiness and opportunity for young people in the state. Well-governed and well-structured state longitudinal data systems can help. Today, most states have disparate data systems that are functionally siloed off by education level, like early childhood or K-12 or higher education. Meaningfully connected and well-governed data systems make it easier to see the relationships between metrics like early-grade reading scores, high school graduation, and wage data. Disaggregated outcome data about the youngest learners up to young adults provide invaluable insights to policymakers, researchers, and parents. This report examines how state leaders should prioritize state longitudinal data systems in legislation and policymaking.

Texas depends on a talented workforce to fuel its economy, and Texans depend on our public education system to prepare them for the jobs of today and tomorrow. But unfortunately, workforce and public education systems often operate independently, making innovation and improvement difficult. Between now and 2036, over 70% of the jobs in Texas will require a postsecondary credential. But far too few Texas students will possess those credentials. Only 22% of Texas eighth graders currently earn a postsecondary degree or credentials within six years of high school graduation. This article examines the Texas Tri-Agency Workforce Initiative. It brings together the Texas Education Agency (TEA), the Texas Higher Education Coordinating Board (THECB), and the Texas Workforce Commission (TWC). The Tri-Agency is focused on three priorities: (1) Pathways: Support efficient and flexible pathways to earning degrees, certificates, and other credentials linked to high-wage, in-demand jobs; (2) Support: Ensure students receive the support necessary to succeed at all stages of their education in their transitions to the workforce; and (3) Infrastructure: Create a robust infrastructure for interagency collaboration around common goals, data, and processes to ensure improved student outcomes and meet employers' needs.

Convolutional neural networks (CNNs) achieve prevailing results in segmentation tasks nowadays and represent the state-of-the-art for image-based analysis. However, the understanding of the accurate decision-making process of a CNN is rather unknown. The research area of explainable artificial intelligence (xAI) primarily revolves around understanding and interpreting this black-box behavior. One way of interpreting a CNN is the use of class activation maps (CAMs) that represent heatmaps to indicate the importance of image areas for the prediction of the CNN. For classification tasks, a variety of CAM algorithms exist. But for segmentation tasks, only one CAM algorithm for the interpretation of the output of a CNN exist. We propose a transfer between existing classification- and segmentation-based methods for more detailed, explainable, and consistent results which show salient pixels in semantic segmentation tasks. The resulting Seg-HiRes-Grad CAM is an extension of the segmentation-based Seg-Grad CAM with the transfer to the classification-based HiRes CAM. Our method improves the previously-mentioned existing segmentation-based method by adjusting it to recently published classification-based methods. Especially for medical image segmentation, this transfer solves existing explainability disadvantages., Comment: Accepted for publication at the Journal of Machine Learning for Biomedical Imaging (MELBA) https://melba-journal.org/2024:023

The coupling between excitons and phonons across adjacent layers has been experimentally observed in various heterostructures of layered materials. Yet the precise mechanism underlying this phenomenon remains elusive. Using the WSe$_2$@hBN heterostructure as an example, we study the origin of the interlayer exciton-phonon coupling and its signature in resonant Raman scattering through first-principles calculations. Our study emphasizes the central role of crystal symmetries in the interlayer exciton-phonon scattering processes, which are responsible for the anomalous resonant Raman intensities of the in-plane and the out-of-plane hBN phonon modes. We find that the deformation potential induced by the hBN phonon interacts with the hybridized hole density of WSe$_2$ excitons near the hBN interface, leading to interlayer exciton-phonon coupling., Comment: 5 Pages, 3 figures

Doped semiconductors are capable of exhibiting metallic-like properties ranging from superconductivity to tunable localized surface plasmon resonances. Diamond is a wide-bandgap semiconductor that is rendered electronically active by incorporating a hole dopant, boron. While the effects of boron doping on the electronic band structure of diamond are well-studied, any link between charge carriers and plasmons, which could facilitate optical applications, has never been shown. Here, we report intervalence plasmons in boron-doped diamond, defined as collective electronic excitations between the valence subbands, opened up by the presence of holes. Evidence for these low energy excitations is provided by scanning transmission electron microscope-valence electron energy loss spectroscopy and photoinduced force infrared spectroscopy. The measured loss and absorbance spectra are subsequently reproduced by first-principles calculations based on the contribution of intervalence band transitions to the dielectric function. Remarkably, the calculations also reveal that the real part of the dielectric function exhibits a resonance characteristic of metallicity (narrow-banded negative values of the dielectric function). The energy of the zero-crossing and the position of the loss peak are found to coincide, and both increase with the carrier density. Our results provide insight into a new mechanism for inducing plasmon-like behavior in doped semiconductors from intervalence band transitions, and the possibility of attaining such properties in diamond, a key emerging material for biomedical and quantum information technologies.

This paper investigates the relationship between scientists' cognitive profile and their ability to generate innovative ideas and gain scientific recognition. We propose a novel author-level metric based on the semantic representation of researchers' past publications to measure cognitive diversity both at individual and team levels. Using PubMed Knowledge Graph (PKG), we analyze the impact of cognitive diversity on novelty, as measured by combinatorial novelty indicators and peer labels on Faculty Opinion. We assessed scientific impact through citations and disruption indicators. We show that the presence of exploratory individuals (i.e., cognitively diverse) is beneficial in generating distant knowledge combinations, but only when balanced by a significant proportion of exploitative individuals (i.e., cognitively specialized). Furthermore, teams with a high proportion of exploitative profiles tend to consolidate science, whereas those with a significant share of both profiles tend to disrupt it. Cognitive diversity between team members appears to be always beneficial to combining more distant knowledge. However, to maximize the relevance of these distant combinations of knowledge, maintaining a limited number of exploratory individuals is essential, as exploitative individuals must question and debate their novel perspectives. These specialized individuals are the most qualified to extract the full potential of novel ideas and integrate them within the existing scientific paradigm.

Coastal ecosystems are increasingly experiencing anthropogenic pressures such as climate heating, CO2 increase, metal and organic pollution, overfishing and resource extraction. Some resulting stressors are more direct like fisheries, others more indirect like ocean acidification, yet they jointly affect marine biota, communities and entire ecosystems. While single-stressor effects have been widely investigated, the interactive effects of multiple stressors on ecosystems are less researched. In this study, we review the literature on multiple stressors and their interactive effects in coastal environments across organisms. We classify the interactions into three categories: synergistic, additive, and antagonistic. We found phytoplankton and mollusks to be the most studied taxonomic groups. The stressor combinations of climate warming, ocean acidification, eutrophication, and metal pollution are the most critical for coastal ecosystems as they exacerbate adverse effects on physiological traits such as growth rate, basal respiration, and size. Phytoplankton appears to be most sensitive to interactions between metal and nutrient pollution. In nutrient-enriched environments, the presence of metals considerably affects the uptake of nutrients, and increases respiration costs and toxin production in phytoplankton. For mollusks, warming and low pH are the most lethal stressors. The combined effect of heat stress and ocean acidification leads to decreased growth rate, shell size, and acid-base regulation capacity in mollusks. However, for a holistic understanding of how coastal food webs will evolve with ongoing changes, we suggest more research on ecosystem-level responses. This can be achieved by combining in-situ observations from controlled environments (e.g. mesocosm experiments) with modelling approaches.

This paper introduces a multilingual dataset of COVID-19 vaccine misinformation, consisting of annotated tweets from three middle-income countries: Brazil, Indonesia, and Nigeria. The expertly curated dataset includes annotations for 5,952 tweets, assessing their relevance to COVID-19 vaccines, presence of misinformation, and the themes of the misinformation. To address challenges posed by domain specificity, the low-resource setting, and data imbalance, we adopt two approaches for developing COVID-19 vaccine misinformation detection models: domain-specific pre-training and text augmentation using a large language model. Our best misinformation detection models demonstrate improvements ranging from 2.7 to 15.9 percentage points in macro F1-score compared to the baseline models. Additionally, we apply our misinformation detection models in a large-scale study of 19 million unlabeled tweets from the three countries between 2020 and 2022, showcasing the practical application of our dataset and models for detecting and analyzing vaccine misinformation in multiple countries and languages. Our analysis indicates that percentage changes in the number of new COVID-19 cases are positively associated with COVID-19 vaccine misinformation rates in a staggered manner for Brazil and Indonesia, and there are significant positive associations between the misinformation rates across the three countries., Comment: Accepted to EMNLP 2023 (Main conference), 9 pages, 5 figures

Sizable hyperpolarisation, i.e. an imbalance of the occupation numbers of nuclear spins in a sample deviating from thermal equilibrium, is needed in various fields of science. For example, hyperpolarised tracers are utilised in magnetic resonance imaging in medicine (MRI) and polarised beams and targets are employed in nuclear physics to study the spin dependence of nuclear forces. Here we show that the quantum interference of transitions induced by radio-wave pumping with longitudinal and radial pulses are able to produce large polarisations at small magnetic fields. This method is easier than established methods, theoretically understood and experimentally proven for beams of metastable hydrogen atoms in the keV energy range. It should also work for a variety of samples at rest. Thus, this technique opens the door for a new generation of polarised tracers, possibly low-field MRI with better spatial resolution or the production of polarised fuel to increase the efficiency of fusion reactors by manipulating the involved cross sections., Comment: 22 pages, 9 figures, splitted into two parts: main paper and methods

Quantised lattice vibrations (i.e., phonons) in solids are robust and unambiguous fingerprints of crystal structures and of their symmetry properties. In metals and semimetals, strong electron-phonon coupling may lead to so-called Kohn anomalies in the phonon dispersion, providing an image of the Fermi surface in a non-electronic observable. Kohn anomalies become prominent in low-dimensional systems, in particular in graphene, where they appear as sharp kinks in the in-plane optical phonon branches. However, in spite of intense research efforts on electron-phonon coupling in graphene and related van der Waals heterostructures, little is known regarding the links between the symmetry properties of optical phonons at and near Kohn anomalies and their sensitivity towards the local environment. Here, using inelastic light scattering (Raman) spectroscopy, we investigate a set of custom-designed graphene-based van der Waals heterostructures, wherein dielectric screening is finely controlled at the atomic layer level. We demonstrate experimentally and explain theoretically that, depending exclusively on their symmetry properties, the two main Raman modes of graphene react differently to the surrounding environment. While the Raman-active near-zone-edge optical phonons in graphene undergo changes in their frequencies due to the neighboring dielectric environment, the in-plane, zone-centre optical phonons are symmetry-protected from the influence of the latter. These results shed new light on the unique electron-phonon coupling properties in graphene and related systems and provide invaluable guidelines to characterise dielectric screening in van der Waals heterostructures and moir\'e superlattices., Comment: 22 pages, main manuscript with 3 figures, appendix with 5 figures, supplemental material with 6 figures

We study two-dimensional noble metal chalcogenides, with composition {Cu, Ag, Au}2{S, Se, Te}, crystallizing in a snub-square lattice. This is a semi-regular two-dimensional tesselation formed by triangles and squares that exhibits geometrical frustration. We use for comparison a square lattice, from which the snub-square tiling can be derived by a simple rotation of the squares. The mono-layer snub-square chalcogenides are very close to thermodynamic stability, with the most stable system (Ag2Se) a mere 7 meV/atom above the convex hull of stability. All compounds studied in the square and snub-square lattice are semiconductors, with band gaps ranging from 0.1 to more than 2.5 eV. Excitonic effects are strong, with an exciton binding energy of around 0.3 eV. We propose the Cu (001) surface as a possible substrate to synthesize Cu2Se, although many other metal and semiconducting surfaces can be found with very good lattice matching.

The introduction of the spatial Lambda-Fleming-Viot model (LV) in population genetics was mainly driven by the pioneering work of Alison Etheridge, in collaboration with Nick Barton and Amandine V\'eber about ten years ago (1,2). The LV model provides a sound mathematical framework for describing the evolution of a population of related individuals along a spatial continuum. It alleviates the "pain in the torus" issue with Wright and Mal\'ecot's isolation by distance model and is sampling consistent, making it a tool of choice for statistical inference. Yet, little is known about the potential connections between the LV and other stochastic processes generating trees and the spatial coordinates along the corresponding lineages. This work focuses on a version of the LV whereby lineages move infinitely rapidly over infinitely small distances. Using simulations, we show that the induced LV tree-generating process is well approximated by a birth-death model. Our results also indicate that Brownian motions modelling the movements of lineages along birth-death trees do not generally provide a good approximation of the LV due to habitat boundaries effects that play an increasingly important role in the long run. Finally, we describe efficient algorithms for fast simulation of the backward and forward in time versions of the LV model.

Cloud-native architectures are often based on microservices and combine different aspects that aim to leverage the capabilities of cloud platforms for software development. Cloud-native architectural characteristics like patterns and best practices aim to design, develop, deploy, and operate such systems efficiently with minimal time and effort. However, architects and developers are faced with the challenge of applying such characteristics in a targeted manner to improve selected quality attributes. Hence, we aim to investigate relationships, or more specifically impacts, between architectural characteristics of cloud-native applications, and quality aspects. The architectural characteristics in consideration are based on our recently proposed quality model for cloud-native software architectures. To validate its elements and revise this literature-based quality model, we conducted a questionnaire-based survey among 42 software professionals. While the survey results reinforce the quality model to a fair extent, they also indicate parts requiring a revision. Thus, as an additional contribution, we present an updated version of the quality model incorporating the survey results. Practitioners will benefit from our work when designing and developing cloud-native applications in a quality-oriented way. Researchers will moreover profit from our specifically developed questionnaire-based survey tool, which allows surveying complex structures like a hierarchical quality model., Comment: 10 pages, 5 figures, accepted at 17th IEEE International Conference on Service-Oriented System Engineering (2023)

The focused ion beam (FIB) is a powerful tool for the fabrication, modification and characterization of materials down to the nanoscale. Starting with the gallium FIB, which was originally intended for photomask repair in the semiconductor industry, there are now many different types of FIB that are commercially available. These instruments use a range of ion species and are applied broadly in materials science, physics, chemistry, biology, medicine, and even archaeology. The goal of this roadmap is to provide an overview of FIB instrumentation, theory, techniques and applications. By viewing FIB developments through the lens of the various research communities, we aim to identify future pathways for ion source and instrumentation development as well as emerging applications, and the scope for improved understanding of the complex interplay of ion-solid interactions. We intend to provide a guide for all scientists in the field that identifies common research interests and will support future fruitful interactions connecting tool development, experiment and theory. While a comprehensive overview of the field is sought, it is not possible to cover all research related to FIB technologies in detail. We give examples of specific projects within the broader context, referencing original works and previous review articles throughout., Comment: This publication is based upon work from the COST Action FIT4NANO CA19140, supported by COST (European Cooperation in Science and Technology) https://www.cost.eu/. Financial support from COST Action CA19140 is acknowledged http://www.fit4nano.eu/ Version 3 has many text and language edits as well as layout tuning but no substantial new content

Despite its simple crystal structure, layered boron nitride features a surprisingly complex variety of phonon-assisted luminescence peaks. We present a combined experimental and theoretical study on ultraviolet-light emission in hexagonal and rhombohedral bulk boron nitride crystals. Emission spectra of high-quality samples are measured via cathodoluminescence spectroscopy, displaying characteristic differences between the two polytypes. These differences are explained using a fully first-principles computational technique that takes into account radiative emission from ``indirect'', finite-momentum, excitons via coupling to finite-momentum phonons. We show that the differences in peak positions, number of peaks and relative intensities can be qualitatively and quantitatively explained, once a full integration over all relevant momenta of excitons and phonons is performed., Comment: Main: 6 pages and 4 figures, Supplementary: 6 pages and 7 figures

Although point defects in solids are one of the most promising physical systems to build functioning qubits, it remains challenging to position them in a deterministic array and to integrate them into large networks. By means of advanced ab initio calculations we show that undissociated screw dislocations in cubic 3C-SiC, and their associated strain fields, could be used to create a deterministic pattern of relevant point defects. Specifically, we present a detailed analysis of the formation energies and electronic structure of the divacancy in 3C-SiC when located in the vicinity of this type of dislocations. Our results show that the divacancy is strongly attracted towards specific and equivalent sites inside the core of the screw dislocations, and would form a one-dimensional arrays along them. Furthermore, we show that the same strain that attracts the divacancy allows the modulation of the position of its electronic states and of its charge transition levels. In the case of the neutral divacancy, we find that these modulations result in the loss of its potential as a qubit. However, these same modulations could transform defects with no potential as qubits when located in bulk, into promising defects when located inside the core of the screw dislocations. Since dislocations are still mostly perceived as harmful defects, our findings represent a technological leap as they show that dislocations can be used as active building blocks in future defect-based quantum computers.

The decarbonization of for example the energy or heat sector leads to the transformation of distribution grids. The expansion of decentralized energy resources and the integration of new consumers due to sector coupling (e.g. heat pumps or electric vehicles) into low voltage grids increases the need for grid expansion and usage of flexibilities in the grid. A high observability of the current grid status is needed to perform these tasks efficiently and effectively. Therefore, there is a need to increase the observability of low voltage grids by installing measurement technologies (e.g. smart meters). Multiple different measurement technologies are available for low voltage grids which can vary in their benefit to observation quality and their installation costs. Therefore, Bayernwerk Netz GmbH and E.DIS AG in cooperation with E-Bridge Consulting GmbH and the Institute for High Voltage Equipment and Grids, Digitalization and Energy Economics (IAEW) investigated the effectiveness of different strategies for the smartification of low voltage grids. This paper presents the methodology used for the investigation and exemplary results focusing on the impact of intelligent cable distribution cabinets and smart meters on the quality of the state estimation., Comment: In proceedings of the 13th "Internationale Energiewirtschaftstagung" (IEWT2023), February 15-17, 2023, Vienna, Austria

Graphene is an ideal platform to study the coherence of quantum interference pathways by tuning doping or laser excitation energy. The latter produces a Raman excitation profile that provides direct insight into the lifetimes of intermediate electronic excitations and, therefore, on quantum interference, which has so far remained elusive. Here, we control the Raman scattering pathways by tuning the laser excitation energy in graphene doped up to 1.05eV, above what achievable with electrostatic doping. The Raman excitation profile of the G mode indicates its position and full width at half maximum are linearly dependent on doping. Doping-enhanced electron-electron interactions dominate the lifetime of Raman scattering pathways, and reduce Raman interference. This paves the way for engineering quantum pathways in doped graphene, nanotubes and topological insulators., Comment: 18 pages, 4 figures

We demonstrate the possibility of engineering the optical properties of transition metal dichalcogenide heterobilayers when one of the constitutive layers has a Janus structure. This has important consequences for the charge separation efficiency. We investigate different MoS$_2$@Janus layer combinations using first-principles methods including electron-hole interactions (excitons) and exciton-phonon coupling. The direction of the intrinsic electric field from the Janus layer modifies the electronic band alignments and, consequently, the energy separation between interlayer exciton states -- which usually have a very low oscillator strength and hence are almost dark in absorption -- and bright in-plane excitons. We find that in-plane lattice vibrations strongly couple the two states, so that exciton-phonon scattering may be a viable generation mechanism for interlayer excitons upon light absorption. In particular, in the case of MoS$_2$@WSSe, the energy separation of the low-lying interlayer exciton from the in-plane exciton is resonant with the transverse optical phonon modes (40 meV). We thus identify this heterobilayer as a prime candidate for efficient electron-hole pair generation with efficient charge carrier separation.

We present a symmetry-based exhaustive approach to explore the structural and compositional richness of two-dimensional materials. We use a ``combinatorial engine'' that constructs potential compounds by occupying all possible Wyckoff positions for a certain space group with combinations of chemical elements. These combinations are restricted by imposing charge neutrality and the Pauling test for electronegativities. The structures are then pre-optimized with a specially crafted universal neural-network force-field, before a final step of geometry optimization using density-functional theory is performed. In this way we unveil an unprecedented variety of two-dimensional materials, covering the whole periodic table in more than 30 different stoichiometries of form A$_n$B$_m$ or A$_n$B$_m$C$_k$. Among the found structures we find examples that can be built by decorating nearly all Platonic and Archimedean tesselations as well as their dual Laves or Catalan tilings. We also obtain a rich, and unexpected, polymorphism for some specific compounds. We further accelerate the exploration of the chemical space of two-dimensional materials by employing machine-learning-accelerated prototype search, based on the structural types discovered in the exhaustive search. In total, we obtain around 6500 compounds, not present in previous available databases of 2D materials, with an energy of less than 250~meV/atom above the convex hull of thermodynamic stability.

Novelpy (v1.2) is an open-source Python package designed to compute bibliometrics indicators. The package aims to provide a tool to the scientometrics community that centralizes different measures of novelty and disruptiveness, enables their comparison and fosters reproducibility. This paper offers a comprehensive review of the different indicators available in Novelpy by formally describing these measures (both mathematically and graphically) and presenting their benefits and limitations. We then compare the different measures on a random sample of 1.5M articles drawn from Pubmed Knowledge Graph to demonstrate the module's capabilities. We encourage anyone interested to participate in the development of future versions.

We consider the counting problem of the number of \textit{leaf-labeled increasing trees}, where internal nodes may have an arbitrary number of descendants. The set of all such trees is a discrete representation of the genealogies obtained under certain population-genetical models such as multiple-merger coalescents. While the combinatorics of the binary trees among those are well understood, for the number of all trees only an approximate asymptotic formula is known. In this work, we validate this formula up to constant terms and compare the asymptotic behavior of the number of all leaf-labelled increasing trees to that of binary, ternary and quaternary trees.

We interpret solving the multi-vehicle routing problem as a team Markov game with partially observable costs. For a given set of customers to serve, the playing agents (vehicles) have the common goal to determine the team-optimal agent routes with minimal total cost. Each agent thereby observes only its own cost. Our multi-agent reinforcement learning approach, the so-called multi-agent Neural Rewriter, builds on the single-agent Neural Rewriter to solve the problem by iteratively rewriting solutions. Parallel agent action execution and partial observability require new rewriting rules for the game. We propose the introduction of a so-called pool in the system which serves as a collection point for unvisited nodes. It enables agents to act simultaneously and exchange nodes in a conflict-free manner. We realize limited disclosure of agent-specific costs by only sharing them during learning. During inference, each agents acts decentrally, solely based on its own cost. First empirical results on small problem sizes demonstrate that we reach a performance close to the employed OR-Tools benchmark which operates in the perfect cost information setting.

Uncertainty estimation bears the potential to make deep learning (DL) systems more reliable. Standard techniques for uncertainty estimation, however, come along with specific combinations of strengths and weaknesses, e.g., with respect to estimation quality, generalization abilities and computational complexity. To actually harness the potential of uncertainty quantification, estimators are required whose properties closely match the requirements of a given use case. In this work, we propose a framework that, firstly, structures and shapes these requirements, secondly, guides the selection of a suitable uncertainty estimation method and, thirdly, provides strategies to validate this choice and to uncover structural weaknesses. By contributing tailored uncertainty estimation in this sense, our framework helps to foster trustworthy DL systems. Moreover, it anticipates prospective machine learning regulations that require, e.g., in the EU, evidences for the technical appropriateness of machine learning systems. Our framework provides such evidences for system components modeling uncertainty.

Studying cell morphology changes in time is critical to understanding cell migration mechanisms. In this work, we present a deep learning-based workflow to segment cancer cells embedded in 3D collagen matrices and imaged with phase-contrast microscopy. Our approach uses transfer learning and recurrent convolutional long-short term memory units to exploit the temporal information from the past and provide a consistent segmentation result. Lastly, we propose a geometrical-characterization approach to studying cancer cell morphology. Our approach provides stable results in time, and it is robust to the different weight initialization or training data sampling. We introduce a new annotated dataset for 2D cell segmentation and tracking, and an open-source implementation to replicate the experiments or adapt them to new image processing problems.

Structural, electronic and thermodynamic properties of native defects in GaS and GaSe monolayers are investigated by means of accurate ab-initio calculations. Based on their charge transition levels we assess the influence of the studied defects on the electrical properties of the monolayers. Specifically, we show that native defects do not behave as shallow dopants and their presence cannot account for the experimentally observed intrinsic doping. In addition, we predict that native defects are efficient compensation and recombination centers. Besides pointing out their detrimental nature, we also calculate the corresponding finite temperature formation energies and provide a window of growth conditions able to reduce the concentration of all relevant native defects., Comment: 14 pages, 8 figures

We present a measurement of the cosmic-ray spectrum above 100\,PeV using the part of the surface detector of the Pierre Auger Observatory that has a spacing of 750~m. An inflection of the spectrum is observed, confirming the presence of the so-called \emph{second-knee} feature. The spectrum is then combined with that of the 1500\,m array to produce a single measurement of the flux, linking this spectral feature with the three additional breaks at the highest energies. The combined spectrum, with an energy scale set calorimetrically via fluorescence telescopes and using a single detector type, results in the most statistically and systematically precise measurement of spectral breaks yet obtained. These measurements are critical for furthering our understanding of the highest energy cosmic rays., Comment: Published version, 25 pages, 20 figures, 11 tables Updated April 2022 to fix typo in Table 10

We revisit the spatial ${\lambda}$-Fleming-Viot process introduced in [1]. Particularly, we are interested in the time $T_0$ to the most recent common ancestor for two lineages. We distinguish between the case where the process acts on the entire two-dimensional plane, and on a finite rectangle. Utilizing a differential equation linking $T_0$ with the physical distance between the lineages, we arrive at simple and reasonably accurate approximation schemes for both cases. Furthermore, our analysis enables us to address the question of whether the genealogical process of the model "comes down from infinity", which has been partly answered before in [2]., Comment: 31 pages, 8 figures

Many machine learning applications can benefit from simulated data for systematic validation - in particular if real-life data is difficult to obtain or annotate. However, since simulations are prone to domain shift w.r.t. real-life data, it is crucial to verify the transferability of the obtained results. We propose a novel framework consisting of a generative label-to-image synthesis model together with different transferability measures to inspect to what extent we can transfer testing results of semantic segmentation models from synthetic data to equivalent real-life data. With slight modifications, our approach is extendable to, e.g., general multi-class classification tasks. Grounded on the transferability analysis, our approach additionally allows for extensive testing by incorporating controlled simulations. We validate our approach empirically on a semantic segmentation task on driving scenes. Transferability is tested using correlation analysis of IoU and a learned discriminator. Although the latter can distinguish between real-life and synthetic tests, in the former we observe surprisingly strong correlations of 0.7 for both cars and pedestrians., Comment: The first two authors contributed equally. Accepted at the 4th Workshop on "Ensuring and Validating Safety for Automated Vehicles" (WS13), IV2021. Under IEEE Copyright

The use of deep neural networks (DNNs) in safety-critical applications like mobile health and autonomous driving is challenging due to numerous model-inherent shortcomings. These shortcomings are diverse and range from a lack of generalization over insufficient interpretability to problems with malicious inputs. Cyber-physical systems employing DNNs are therefore likely to suffer from safety concerns. In recent years, a zoo of state-of-the-art techniques aiming to address these safety concerns has emerged. This work provides a structured and broad overview of them. We first identify categories of insufficiencies to then describe research activities aiming at their detection, quantification, or mitigation. Our paper addresses both machine learning experts and safety engineers: The former ones might profit from the broad range of machine learning topics covered and discussions on limitations of recent methods. The latter ones might gain insights into the specifics of modern ML methods. We moreover hope that our contribution fuels discussions on desiderata for ML systems and strategies on how to propel existing approaches accordingly., Comment: 94 pages

Data-driven sensor interpretation in autonomous driving can lead to highly implausible predictions as can most of the time be verified with common-sense knowledge. However, learning common knowledge only from data is hard and approaches for knowledge integration are an active research area. We propose to use a partly human-designed, partly learned set of rules to describe relations between objects of a traffic scene on a high level of abstraction. In doing so, we improve and robustify existing deep neural networks consuming low-level sensor information. We present an initial study adapting the well-established Probabilistic Soft Logic (PSL) framework to validate and improve on the problem of semantic segmentation. We describe in detail how we integrate common knowledge into the segmentation pipeline using PSL and verify our approach in a set of experiments demonstrating the increase in robustness against several severe image distortions applied to the A2D2 autonomous driving data set., Comment: Published at SAIAD (Safe Artificial Intelligence for Automated Driving) workshop at CVPR2021

Artificial intelligence for autonomous driving must meet strict requirements on safety and robustness, which motivates the thorough validation of learned models. However, current validation approaches mostly require ground truth data and are thus both cost-intensive and limited in their applicability. We propose to overcome these limitations by a model agnostic validation using a-priori knowledge from street maps. In particular, we show how to validate semantic segmentation masks and demonstrate the potential of our approach using OpenStreetMap. We introduce validation metrics that indicate false positive or negative road segments. Besides the validation approach, we present a method to correct the vehicle's GPS position so that a more accurate localization can be used for the street-map based validation. Lastly, we present quantitative results on the Cityscapes dataset indicating that our validation approach can indeed uncover errors in semantic segmentation masks., Comment: Final version accepted at the International Conference on Pattern Recognition (ICPR). arXiv admin note: substantial text overlap with arXiv:2011.08008

Van der Waals layered materials with well-defined twist angles between the crystal lattices of individual layers have attracted increasing attention due to the emergence of unexpected material properties. As many properties critically depend on the exact twist angle and its spatial homogeneity, there is a need for a fast and non-invasive characterization technique of the local twist angle, to be applied preferably right after stacking. We demonstrate that confocal Raman spectroscopy can be utilized to spatially map the twist angle in stacked bilayer graphene for angles between 6.5{\deg} and 8{\deg} when using a green excitation laser. The twist angles can directly be extracted from the moir\'e superlattice-activated Raman scattering process of the transverse acoustic (TA) phonon mode. Furthermore, we show that the width of the TA Raman peak contains valuable information on spatial twist angle variations on length scales below the laser spot size of ~ 500 nm., Comment: 14 pages, 10 figures

Fake information poses one of the major threats for society in the 21st century. Identifying misinformation has become a key challenge due to the amount of fake news that is published daily. Yet, no approach is established that addresses the dynamics and versatility of fake news editorials. Instead of classifying content, we propose an evidence retrieval approach to handle fake news. The learning task is formulated as an unsupervised machine learning problem. For validation purpose, we provide the user with a set of news articles from reliable news sources supporting the hypothesis of the news article in query and the final decision is left to the user. Technically we propose a two-step process: (i) Aggregation-step: With information extracted from the given text we query for similar content from reliable news sources. (ii) Refining-step: We narrow the supporting evidence down by measuring the semantic distance of the text with the collection from step (i). The distance is calculated based on Word2Vec and the Word Mover's Distance. In our experiments, only content that is below a certain distance threshold is considered as supporting evidence. We find that our approach is agnostic to concept drifts, i.e. the machine learning task is independent of the hypotheses in a text. This makes it highly adaptable in times where fake news is as diverse as classical news is. Our pipeline offers the possibility for further analysis in the future, such as investigating bias and differences in news reporting.

The Pierre Auger Observatory, at present the largest cosmic-ray observatory ever built, is instrumented with a ground array of 1600 water-Cherenkov detectors, known as the Surface Detector (SD). The SD samples the secondary particle content (mostly photons, electrons, positrons and muons) of extensive air showers initiated by cosmic rays with energies ranging from $10^{17}~$eV up to more than $10^{20}~$eV. Measuring the independent contribution of the muon component to the total registered signal is crucial to enhance the capability of the Observatory to estimate the mass of the cosmic rays on an event-by-event basis. However, with the current design of the SD, it is difficult to straightforwardly separate the contributions of muons to the SD time traces from those of photons, electrons and positrons. In this paper, we present a method aimed at extracting the muon component of the time traces registered with each individual detector of the SD using Recurrent Neural Networks. We derive the performances of the method by training the neural network on simulations, in which the muon and the electromagnetic components of the traces are known. We conclude this work showing the performance of this method on experimental data of the Pierre Auger Observatory. We find that our predictions agree with the parameterizations obtained by the AGASA collaboration to describe the lateral distributions of the electromagnetic and muonic components of extensive air showers., Comment: Published version, 23 pages, 15 figures

Helium implantation in epitaxial thin films is a way to control the out-of-plane deformation independently from the in-plane strain controlled by epitaxy. In particular, implantation by means of a helium microscope allows for local implantation and patterning down to the nanometer resolution, which is of interest for device applications. We present here a study of bismuth ferrite (BiFeO3) films where strain was patterned locally by helium implantation. Our combined Raman, XRD and TEM study shows that the implantation causes an elongation of the BiFeO3 unit cell and ultimately a transition towards the so-called super-tetragonal polymorph via states with mixed phases. In addition, TEM reveals the onset of amorphization at a threshold dose that does not seem to impede the overall increase in tetragonality. The phase transition from the R-like to T-like BiFeO3 appears as first-order in character, with regions of phase coexistence and abrupt changes in lattice parameters.

We present the first measurement of the fluctuations in the number of muons in extensive air showers produced by ultrahigh energy cosmic rays. We find that the measured fluctuations are in good agreement with predictions from air shower simulations. This observation provides new insights into the origin of the previously reported deficit of muons in air shower simulations and constrains models of hadronic interactions at ultrahigh energies. Our measurement is compatible with the muon deficit originating from small deviations in the predictions from hadronic interaction models of particle production that accumulate as the showers develop., Comment: Published version, 11 pages, 3 figures, 1 table, Supplemental material appended

How has the science system reacted to the early stages of the COVID-19 pandemic? Here we compare the (growing) international network for coronavirus research with the broader international health science network. Our findings show that, before the outbreak, coronavirus research realized a relatively small and rather peculiar niche within the global health sciences. As a response to the pandemic, the international network for coronavirus research expanded rapidly along the hierarchical structure laid out by the global health science network. Thus, in face of the crisis, the global health science system proved to be structurally stable yet versatile in research. The observed versatility supports optimistic views on the role of science in meeting future challenges. However, the stability of the global core-periphery structure may be worrying, because it reduces learning opportunities and social capital of scientifically peripheral countries -- not only during this pandemic but also in its "normal" mode of operation.

FRAM (F/Photometric Robotic Atmospheric Monitor) is a robotic telescope operated at the Pierre Auger Observatory in Argentina for the purposes of atmospheric monitoring using stellar photometry. As a passive system which does not produce any light that could interfere with the observations of the fluorescence telescopes of the observatory, it complements the active monitoring systems that use lasers. We discuss the applications of stellar photometry for atmospheric monitoring at optical observatories in general and the particular modes of operation employed by the Auger FRAM. We describe in detail the technical aspects of FRAM, the hardware and software requirements for a successful operation of a robotic telescope for such a purpose and their implementation within the FRAM system., Comment: Published version, 27 pages, 8 figures

The Auger Muon Infill Ground Array (AMIGA) is part of the AugerPrime upgrade of the Pierre Auger Observatory. It consists of particle counters buried 2.3 m underground next to the water-Cherenkov stations that form the 23.5 km$^2$ large infilled array. The reduced distance between detectors in this denser area allows the lowering of the energy threshold for primary cosmic ray reconstruction down to about 10$^{17}$ eV. At the depth of 2.3 m the electromagnetic component of cosmic ray showers is almost entirely absorbed so that the buried scintillators provide an independent and direct measurement of the air showers muon content. This work describes the design and implementation of the AMIGA embedded system, which provides centralized control, data acquisition and environment monitoring to its detectors. The presented system was firstly tested in the engineering array phase ended in 2017, and lately selected as the final design to be installed in all new detectors of the production phase. The system was proven to be robust and reliable and has worked in a stable manner since its first deployment., Comment: Accepted for publication at JINST. Published version, 34 pages, 15 figures, 4 tables

Statistical models are inherently uncertain. Quantifying or at least upper-bounding their uncertainties is vital for safety-critical systems such as autonomous vehicles. While standard neural networks do not report this information, several approaches exist to integrate uncertainty estimates into them. Assessing the quality of these uncertainty estimates is not straightforward, as no direct ground truth labels are available. Instead, implicit statistical assessments are required. For regression, we propose to evaluate uncertainty realism -- a strict quality criterion -- with a Mahalanobis distance-based statistical test. An empirical evaluation reveals the need for uncertainty measures that are appropriate to upper-bound heavy-tailed empirical errors. Alongside, we transfer the variational U-Net classification architecture to standard supervised image-to-image tasks. We adopt it to the automotive domain and show that it significantly improves uncertainty realism compared to a plain encoder-decoder model., Comment: Accepted at the NeurIPS 2019 Workshop on Machine Learning for Autonomous Driving (ML4AD)

The atmospheric depth of the air shower maximum $X_{\mathrm{max}}$ is an observable commonly used for the determination of the nuclear mass composition of ultra-high energy cosmic rays. Direct measurements of $X_{\mathrm{max}}$ are performed using observations of the longitudinal shower development with fluorescence telescopes. At the same time, several methods have been proposed for an indirect estimation of $X_{\mathrm{max}}$ from the characteristics of the shower particles registered with surface detector arrays. In this paper, we present a deep neural network (DNN) for the estimation of $X_{\mathrm{max}}$. The reconstruction relies on the signals induced by shower particles in the ground based water-Cherenkov detectors of the Pierre Auger Observatory. The network architecture features recurrent long short-term memory layers to process the temporal structure of signals and hexagonal convolutions to exploit the symmetry of the surface detector array. We evaluate the performance of the network using air showers simulated with three different hadronic interaction models. Thereafter, we account for long-term detector effects and calibrate the reconstructed $X_{\mathrm{max}}$ using fluorescence measurements. Finally, we show that the event-by-event resolution in the reconstruction of the shower maximum improves with increasing shower energy and reaches less than $25~\mathrm{g/cm^{2}}$ at energies above $2\times 10^{19}~\mathrm{eV}$., Comment: Published version, 29 pages, 12 figures

We present a novel method to search for structures of coherently aligned patterns in ultra-high energy cosmic-ray arrival directions simultaneously across the entire sky. This method can be used to obtain information on the Galactic magnetic field, in particular the integrated component perpendicular to the line of sight, from cosmic-ray data only. Using a likelihood-ratio approach, neighboring cosmic rays are related by rotatable, elliptically shaped density distributions and the significance of their alignment with respect to circular distributions is evaluated. In this way, a vector field tangential to the celestial sphere is fitted which approximates the local deflections in cosmic magnetic fields if significant deflection structures are detected. The sensitivity of the method is evaluated on the basis of astrophysical simulations of the ultra-high energy cosmic-ray sky, where a discriminative power between isotropic and signal-induced scenarios is found., Comment: 14 pages, 15 figures

Quantification of uncertainty is one of the most promising approaches to establish safe machine learning. Despite its importance, it is far from being generally solved, especially for neural networks. One of the most commonly used approaches so far is Monte Carlo dropout, which is computationally cheap and easy to apply in practice. However, it can underestimate the uncertainty. We propose a new objective, referred to as second-moment loss (SML), to address this issue. While the full network is encouraged to model the mean, the dropout networks are explicitly used to optimize the model variance. We intensively study the performance of the new objective on various UCI regression datasets. Comparing to the state-of-the-art of deep ensembles, SML leads to comparable prediction accuracies and uncertainty estimates while only requiring a single model. Under distribution shift, we observe moderate improvements. As a side result, we introduce an intuitive Wasserstein distance-based uncertainty measure that is non-saturating and thus allows to resolve quality differences between any two uncertainty estimates., Comment: Accepted at the 3rd Symposium on Advances in Approximate Bayesian Inference (AABI), code is available on: https://github.com/fraunhofer-iais/second-moment-loss. arXiv admin note: substantial text overlap with arXiv:2012.12687

Despite of its importance for safe machine learning, uncertainty quantification for neural networks is far from being solved. State-of-the-art approaches to estimate neural uncertainties are often hybrid, combining parametric models with explicit or implicit (dropout-based) ensembling. We take another pathway and propose a novel approach to uncertainty quantification for regression tasks, Wasserstein dropout, that is purely non-parametric. Technically, it captures aleatoric uncertainty by means of dropout-based sub-network distributions. This is accomplished by a new objective which minimizes the Wasserstein distance between the label distribution and the model distribution. An extensive empirical analysis shows that Wasserstein dropout outperforms state-of-the-art methods, on vanilla test data as well as under distributional shift, in terms of producing more accurate and stable uncertainty estimates.

We propose DenseHMM - a modification of Hidden Markov Models (HMMs) that allows to learn dense representations of both the hidden states and the observables. Compared to the standard HMM, transition probabilities are not atomic but composed of these representations via kernelization. Our approach enables constraint-free and gradient-based optimization. We propose two optimization schemes that make use of this: a modification of the Baum-Welch algorithm and a direct co-occurrence optimization. The latter one is highly scalable and comes empirically without loss of performance compared to standard HMMs. We show that the non-linearity of the kernelization is crucial for the expressiveness of the representations. The properties of the DenseHMM like learned co-occurrences and log-likelihoods are studied empirically on synthetic and biomedical datasets., Comment: Accepted at LMRL workshop at NeurIPS 2020. Code is available on: https://github.com/fraunhofer-iais/dense-hmm

To obtain direct measurements of the muon content of extensive air showers with energy above $10^{16.5}$ eV, the Pierre Auger Observatory is currently being equipped with an underground muon detector (UMD), consisting of 219 10 $\mathrm{m^2}$-modules, each segmented into 64 scintillators coupled to silicon photomultipliers (SiPMs). Direct access to the shower muon content allows for the study of both of the composition of primary cosmic rays and of high-energy hadronic interactions in the forward direction. As the muon density can vary between tens of muons per m$^2$ close to the intersection of the shower axis with the ground to much less than one per m$^2$ when far away, the necessary broad dynamic range is achieved by the simultaneous implementation of two acquisition modes in the read-out electronics: the binary mode, tuned to count single muons, and the ADC mode, suited to measure a high number of them. In this work, we present the end-to-end calibration of the muon detector modules: first, the SiPMs are calibrated by means of the binary channel, and then, the ADC channel is calibrated using atmospheric muons, detected in parallel to the shower data acquisition. The laboratory and field measurements performed to develop the implementation of the full calibration chain of both binary and ADC channels are presented and discussed. The calibration procedure is reliable to work with the high amount of channels in the UMD, which will be operated continuously, in changing environmental conditions, for several years., Comment: Published version, 25 pages, 12 figures, 0 tables

Nanoparticles occur in various environments as a consequence of man-made processes, which raises concerns about their impact on the environment and human health. To allow for proper risk assessment, a precise and statistically relevant analysis of particle characteristics (such as e.g. size, shape and composition) is required that would greatly benefit from automated image analysis procedures. While deep learning shows impressive results in object detection tasks, its applicability is limited by the amount of representative, experimentally collected and manually annotated training data. Here, we present an elegant, flexible and versatile method to bypass this costly and tedious data acquisition process. We show that using a rendering software allows to generate realistic, synthetic training data to train a state-of-the art deep neural network. Using this approach, we derive a segmentation accuracy that is comparable to man-made annotations for toxicologically relevant metal-oxide nanoparticle ensembles which we chose as examples. Our study paves the way towards the use of deep learning for automated, high-throughput particle detection in a variety of imaging techniques such as microscopies and spectroscopies, for a wide variety of studies and applications, including the detection of plastic micro- and nanoparticles.

AMIGA (Auger Muons and Infill for the Ground Array) is an upgrade of the Pierre Auger Observatory to complement the study of ultra-high-energy cosmic rays (UHECR) by measuring the muon content of extensive air showers (EAS). It consists of an array of 61 water Cherenkov detectors on a denser spacing in combination with underground scintillation detectors used for muon density measurement. Each detector is composed of three scintillation modules, with 10 m$^2$ detection area per module, buried at 2.3 m depth, resulting in a total detection area of 30 m$^2$. Silicon photomultiplier sensors (SiPM) measure the amount of scintillation light generated by charged particles traversing the modules. In this paper, the design of the front-end electronics to process the signals of those SiPMs and test results from the laboratory and from the Pierre Auger Observatory are described. Compared to our previous prototype, the new electronics shows a higher performance, higher efficiency and lower power consumption, and it has a new acquisition system with increased dynamic range that allows measurements closer to the shower core. The new acquisition system is based on the measurement of the total charge signal that the muonic component of the cosmic ray shower generates in the detector., Comment: 40 pages, 33 figures

Artificial intelligence for autonomous driving must meet strict requirements on safety and robustness. We propose to validate machine learning models for self-driving vehicles not only with given ground truth labels, but also with additional a-priori knowledge. In particular, we suggest to validate the drivable area in semantic segmentation masks using given street map data. We present first results, which indicate that prediction errors can be uncovered by map-based validation.