Search

The control of electron spins in materials that are simultaneously ferromagnetic and insulating opens up a wealth of quantum phenomena in spin-based electronics. Thin films of europium oxide (EuO) are ideal for the generation and manipulation of spin-polarized states, but so far there are no experimental literature reports on the magnetic domain patterns for EuO. However, at these microscopic length scales, magnetic relaxation between the remanent and demagnetized states takes place in any spintronic device. This relaxation process involves displacements of magnetic domain walls and can therefore be strongly influenced by the film structure and thickness. Here we present an investigation of the temperature-dependent behavior of magnetic domains and hysteresis in bulk-like (25 nm) and ultrathin (3 nm) EuO films. Magneto-optical Kerr microscopy is used, a technique that is a valuable tool to explore microscopic features such as spin dynamics and magnetic domain walls. Significant Kerr rotation in EuO led to high-contrast magnetic domain images in thick films, facilitating observation of domain dynamics. The critical temperature (TC) and coercivity shows strong thickness-dependent variations. The analysis and comparison of hysteresis loops and domain imaging in EuO and EuO/Co reveal proximity effect-induced antiferromagnetic coupling of both layers. To elucidate the magnetization reversal dynamics in EuO, micromagnetic simulations using MuMax3 were performed below and above TC. This comprehensive approach aims to comprehend the impact of magnetism and magnetic proximity effect in EuO on the micromagnetic scale, potentially extending its magnetic ordering beyond TC.

The Archives of Photographic PLates for Astronomical USE (APPLAUSE) project is aimed at digitising astronomical photographic plates from three major German plate collections, making them accessible through integration into the International Virtual Observatory (IVO). Photographic plates and related materials (logbooks, envelopes, etc.) were scanned with commercial flatbed scanners. Astrometric and photometric calibrations were carried out with the developed PyPlate software, using Gaia EDR3 data as a reference. The APPLAUSE data publication complies with IVO standards. The latest data release contains images and metadata from 27 plate collections from the partner institutes in Hamburg, Bamberg, and Potsdam, along with digitised archives provided by Tautenburg, Tartu, and Vatican observatories. Altogether, over two billion calibrated measurements extracted from about 70,000 direct photographic plates can readily be used to create long-term light curves. For instance, we constructed the historic light curve of the enigmatic dipping star KIC 8462852. We found no evidence of previously assumed variations on timescales of decades in our light curve. Potential uses of APPLAUSE images for transient sources can be appreciated by following the development of the nova shell of GK Per (1901) over time and the change in brightness of two extragalactic supernovae. The database holds about 10,000 spectral plates. We made use of objective prism plates to follow the temporal changes of Nova DN Gem through 1912 and 1913, highlighting an outburst in early 1913., Comment: 19 pages,2 Tables, 19 figures, A&A accepted

In the framework of solid mechanics, the task of deriving material parameters from experimental data has recently re-emerged with the progress in full-field measurement capabilities and the renewed advances of machine learning. In this context, new methods such as the virtual fields method and physics-informed neural networks have been developed as alternatives to the already established least-squares and finite element-based approaches. Moreover, model discovery problems are starting to emerge and can also be addressed in a parameter estimation framework. These developments call for a new unified perspective, which is able to cover both traditional parameter estimation methods and novel approaches in which the state variables or the model structure itself are inferred as well. Adopting concepts discussed in the inverse problems community, we distinguish between all-at-once and reduced approaches. With this general framework, we are able to structure a large portion of the literature on parameter estimation in computational mechanics - and we can identify combinations that have not yet been addressed, two of which are proposed in this paper. We also discuss statistical approaches to quantify the uncertainty related to the estimated parameters, and we propose a novel two-step procedure for identification of complex material models based on both frequentist and Bayesian principles. Finally, we illustrate and compare several of the aforementioned methods with mechanical benchmarks based on synthetic and real data.

LiDAR-based 3D object detection has become an essential part of automated driving due to its ability to localize and classify objects precisely in 3D. However, object detectors face a critical challenge when dealing with unknown foreground objects, particularly those that were not present in their original training data. These out-of-distribution (OOD) objects can lead to misclassifications, posing a significant risk to the safety and reliability of automated vehicles. Currently, LiDAR-based OOD object detection has not been well studied. We address this problem by generating synthetic training data for OOD objects by perturbing known object categories. Our idea is that these synthetic OOD objects produce different responses in the feature map of an object detector compared to in-distribution (ID) objects. We then extract features using a pre-trained and fixed object detector and train a simple multilayer perceptron (MLP) to classify each detection as either ID or OOD. In addition, we propose a new evaluation protocol that allows the use of existing datasets without modifying the point cloud, ensuring a more authentic evaluation of real-world scenarios. The effectiveness of our method is validated through experiments on the newly proposed nuScenes OOD benchmark. The source code is available at https://github.com/uulm-mrm/mmood3d., Comment: Accepted for publication at the 2024 35th IEEE Intelligent Vehicles Symposium (IV 2024), June 2-5, 2024, in Jeju Island, Korea

Despite considerable strides in developing deep learning models for 3D medical image segmentation, the challenge of effectively generalizing across diverse image distributions persists. While domain generalization is acknowledged as vital for robust application in clinical settings, the challenges stemming from training with a limited Field of View (FOV) remain unaddressed. This limitation leads to false predictions when applied to body regions beyond the FOV of the training data. In response to this problem, we propose a novel loss function that penalizes predictions in implausible body regions, applicable in both single-dataset and multi-dataset training schemes. It is realized with a Body Part Regression model that generates axial slice positional scores. Through comprehensive evaluation using a test set featuring varying FOVs, our approach demonstrates remarkable improvements in generalization capabilities. It effectively mitigates false positive tumor predictions up to 85% and significantly enhances overall segmentation performance.

Within the framework of microscopic thermodynamics, correlations can play a crucial role for energy extraction. Our work sheds light on this connection by demonstrating that entanglement governs the amount of extractable energy in a controllable setting. We experimentally investigate a fundamental link between information, encoded in tunable non-classical correlations and quantified by quantum state tomography, and its utility as fuel for energy extraction. We realize an agent-demon protocol involving two trapped-ion qubits, and show that by implementing an appropriate feedback policy, the demon can optimize the energy extraction process, capitalizing on the correlations between the system's constituents. By quantifying both the concurrence of the two-qubit resource state and the energy extraction gain from applying the feedback policy, we corroborate the connection between information and energy, solidifying the role of non-classical correlations as a resource for thermodynamic processes., Comment: 5 pages, 4 figures

Many post-processing methods improve forecasts at individual locations but remove their correlation structure, which is crucial for predicting larger-scale events like total precipitation amount over areas such as river catchments that are relevant for weather warnings and flood predictions. We propose a method to reintroduce spatial correlation into a post-processed forecast using an R-vine copula fitted to historical observations. This method works similarly to related approaches like the Schaake shuffle and ensemble copula coupling, i.e., by rearranging predictions at individual locations and reintroducing spatial correlation while maintaining the post-processed marginal distribution. Here, the copula measures how well an arrangement compares with the historical distribution of precipitation. No close relationship is needed between the post-processed marginal distributions and the spatial correlation source. This is an advantage compared to Schaake shuffle and ensemble copula coupling, which rely on a ranking with no ties at each considered location in their source for spatial correlations. However, weather variables such as the precipitation amount, whose distribution has an atom at zero, have rankings with ties. To evaluate the proposed method, it is applied to a precipitation forecast produced by a combination model with two input forecasts that deliver calibrated marginal distributions but without spatial correlations. The obtained results indicate that the calibration of the combination model carries over to the output of the proposed model, i.e., the evaluation of area predictions shows a similar improvement in forecast quality as the predictions for individual locations. Additionally, the spatial correlation of the forecast is evaluated with the help of object-based metrics, for which the proposed model also shows an improvement compared to both input forecasts.

Proximity effects allow for the adjustment of magnetic properties in a physically elegant way. If two thin ferromagnetic (FM) films are brought into contact, electronic coupling alters their magnetic exchange interaction at their interface. For a low-TC rare-earth FM coupled to a 3d transition metal FM, even room temperature magnetism is within reach. In addition, magnetic proximity coupling is particularly promising for increasing the magnetic order of metastable materials such as europium monoxide (EuO) beyond their bulk TC, since neither the stoichiometry nor the insulating properties are modified. We investigate the magnetic proximity effect at Fe/EuO and Co/EuO interfaces using hard X-ray photoelectron spectroscopy. By exciting the FM layers with circularly polarized light, magnetic dichroism is observed in angular dependence on the photoemission geometry. In this way, the depth-dependence of the magnetic signal is determined element-specifically for the EuO and 3d FM parts of the bilayers. In connection with atomistic spin dynamics simulations, the thickness of EuO layer is found to be crucial, indicating that the observed antiferromagnetic proximity coupling is a short-ranged and genuine interface phenomenon. This fact turns the bilayer into a strong synthetic ferrimagnet. The increase in magnetic order in EuO occurs in a finite spatial range and is therefore particularly strong in the 2D limit-a counterintuitive but very useful phenomenon for spin-based device applications.

Systems of cooperation and interaction are usually studied in the context of real or complex vector spaces. Additional insight, however, is gained when such systems are represented in vector spaces with multiplicative structures, i.e., in algebras. Algebras, on the other hand, are conveniently viewed as polynomial algebras. In particular, basic interpretations of natural numbers yield natural polynomial algebras and offer a new unifying view on cooperation and interaction. For example, the concept of Galois transforms and zero-dividends of cooperative games is introduced as a nonlinear analogue of the classical Harsanyi dividends. Moreover, the polynomial model unifies various versions of Fourier transforms. Tensor products of polynomial spaces establish a unifying model with quantum theory and allow to study classical cooperative games as interaction activities in a quantum-theoretic context.

The release of nnU-Net marked a paradigm shift in 3D medical image segmentation, demonstrating that a properly configured U-Net architecture could still achieve state-of-the-art results. Despite this, the pursuit of novel architectures, and the respective claims of superior performance over the U-Net baseline, continued. In this study, we demonstrate that many of these recent claims fail to hold up when scrutinized for common validation shortcomings, such as the use of inadequate baselines, insufficient datasets, and neglected computational resources. By meticulously avoiding these pitfalls, we conduct a thorough and comprehensive benchmarking of current segmentation methods including CNN-based, Transformer-based, and Mamba-based approaches. In contrast to current beliefs, we find that the recipe for state-of-the-art performance is 1) employing CNN-based U-Net models, including ResNet and ConvNeXt variants, 2) using the nnU-Net framework, and 3) scaling models to modern hardware resources. These results indicate an ongoing innovation bias towards novel architectures in the field and underscore the need for more stringent validation standards in the quest for scientific progress., Comment: Accepted at MICCAI 2024

This paper introduces a code generator designed for node-level optimized, extreme-scalable, matrix-free finite element operators on hybrid tetrahedral grids. It optimizes the local evaluation of bilinear forms through various techniques including tabulation, relocation of loop invariants, and inter-element vectorization - implemented as transformations of an abstract syntax tree. A key contribution is the development, analysis, and generation of efficient loop patterns that leverage the local structure of the underlying tetrahedral grid. These significantly enhance cache locality and arithmetic intensity, mitigating bandwidth-pressure associated with compute-sparse, low-order operators. The paper demonstrates the generator's capabilities through a comprehensive educational cycle of performance analysis, bottleneck identification, and emission of dedicated optimizations. For three differential operators ($-\Delta$, $-\nabla \cdot (k(\mathbf{x})\, \nabla\,)$, $\alpha(\mathbf{x})\, \mathbf{curl}\ \mathbf{curl} + \beta(\mathbf{x}) $), we determine the set of most effective optimizations. Applied by the generator, they result in speed-ups of up to 58$\times$ compared to reference implementations. Detailed node-level performance analysis yields matrix-free operators with a throughput of 1.3 to 2.1 GDoF/s, achieving up to 62% peak performance on a 36-core Intel Ice Lake socket. Finally, the solution of the curl-curl problem with more than a trillion ($ 10^{12}$) degrees of freedom on 21504 processes in less than 50 seconds demonstrates the generated operators' performance and extreme-scalability as part of a full multigrid solver., Comment: 22 pages

We study the magnetism of a layered, spin-$\frac12$ organic-inorganic copper sulfate, which is a close realization of the star lattice antiferromagnet, one of the playgrounds of geometric frustration and resonating valence bond physics in two spatial dimensions. Our thermodynamic measurements show no ordering down to 0.1 K and a characteristic field-induced entropic shift, revealing the presence of an infinite number of competing states down to very low energy scales. The response to external magnetic fields shows, in addition, a peculiar anisotropy, reflected in the formation of a 1/3 magnetization plateau (stable up to full saturation around 105 T) and a paramagnetic, Curie-like susceptibility for one direction of the field (${\bf H}\parallel{\bf c}$), and a completely different response in other field directions. Our first-principles density functional theory calculations and exact diagonalizations show that these experimental puzzles are distinctive signatures of a strong interplay between geometric frustration and sizable Dzyaloshinskii-Moriya interactions, and the emergence of a continuous U(1) symmetry at low energy scales., Comment: 7 pages, 3 figures + Supplemental Material, accepted in Phys. Rev. B (Letter)

The 2HDM+$a$ model is one of the main models used in the interpretations of dark matter searches at the LHC. So far, all the 2HDM+$a$ benchmarks considered by the ATLAS and CMS experiments are limited to a type-II Yukawa sector, in which the Higgs bosons $A$, $H$, and $H^\pm$ are all constrained to be mass-degenerate and heavier than around $600 \, {\rm GeV}$. In this work, we present the first detailed study of 2HDM+$a$ models with a type-I Yukawa sector, which, for moderate values of $\tan\beta$, lift the constraints from flavour physics, allowing the extra Higgs bosons to be even lighter than the $125 \, {\rm GeV}$ Higgs boson discovered at the LHC. We discuss several benchmarks where the $A$, $H$, and $H^\pm$ states are not necessarily mass-degenerate and the signatures that arise in these models, some of which have not yet been explored at the LHC. We present the dominant channels in the studied benchmarks and the expected sensitivity in Run 2 data using truth-level analyses and discuss potential improvements in the experimental searches for Run 3., Comment: 47 pages, 14 figures; v2: modified version of the manuscript as published in JHEP

We study stationary points of the bending energy of curves $\gamma\colon[a,b]\to\mathbb{R}^n$ subject to constraints on the arc-length and total torsion while simultaneously allowing for a variable bending stiffness along the arc-length of the curve. Physically, this can be understood as a model for an elastic wire with isotropic cross-section of varying thickness. We derive the corresponding Euler-Lagrange equations for variations that are compactly supported away from the end points thus obtaining characterizations for elastic curves with variable bending stiffness. Moreover, we provide a collection of alternative characterizations, e.g., in terms of the curvature function. Adding to numerous known results relating elastic curves to dynamics, we establish connections between elastic curves with variable bending stiffness and damped pendulums and the flow of vortex filaments with finite thickness.

We propose, analyze, and test new robust iterative solvers for systems of linear algebraic equations arising from the space-time finite element discretization of reduced optimality systems defining the approximate solution of hyperbolic distributed, tracking-type optimal control problems with both the standard $L^2$ and the more general energy regularizations. In contrast to the usual time-stepping approach, we discretize the optimality system by space-time continuous piecewise-linear finite element basis functions which are defined on fully unstructured simplicial meshes. If we aim at the asymptotically best approximation of the given desired state $y_d$ by the computed finite element state $y_{\varrho h}$, then the optimal choice of the regularization parameter $\varrho$ is linked to the space-time finite element mesh-size $h$ by the relations $\varrho=h^4$ and $\varrho=h^2$ for the $L^2$ and the energy regularization, respectively. For this setting, we can construct robust (parallel) iterative solvers for the reduced finite element optimality systems. These results can be generalized to variable regularization parameters adapted to the local behavior of the mesh-size that can heavily change in the case of adaptive mesh refinements. The numerical results illustrate the theoretical findings firmly.

Accurately segmenting thin tubular structures, such as vessels, nerves, roads or concrete cracks, is a crucial task in computer vision. Standard deep learning-based segmentation loss functions, such as Dice or Cross-Entropy, focus on volumetric overlap, often at the expense of preserving structural connectivity or topology. This can lead to segmentation errors that adversely affect downstream tasks, including flow calculation, navigation, and structural inspection. Although current topology-focused losses mark an improvement, they introduce significant computational and memory overheads. This is particularly relevant for 3D data, rendering these losses infeasible for larger volumes as well as increasingly important multi-class segmentation problems. To mitigate this, we propose a novel Skeleton Recall Loss, which effectively addresses these challenges by circumventing intensive GPU-based calculations with inexpensive CPU operations. It demonstrates overall superior performance to current state-of-the-art approaches on five public datasets for topology-preserving segmentation, while substantially reducing computational overheads by more than 90%. In doing so, we introduce the first multi-class capable loss function for thin structure segmentation, excelling in both efficiency and efficacy for topology-preservation., Comment: Accepted at ECCV 2024

Lattice gauge theories (LGTs) provide valuable insights into problems in strongly correlated many-body systems. Confinement which arises when matter is coupled to gauge fields is just one of the open problems, where LGT formalism can explain the underlying mechanism. However, coupling gauge fields to dynamical charges complicates the theoretical and experimental treatment of the problem. Developing a simplified mean-field theory is thus one of the ways to gain new insights into these complicated systems. Here we develop a mean-field theory of a paradigmatic 1+1D $\mathbb{Z}_2$ lattice gauge theory with superconducting pairing term, the gauged Kitaev chain, by decoupling charge and $\mathbb{Z}_2$ fields while enforcing the Gauss law on the mean-field level. We first determine the phase diagram of the original model in the context of confinement, which allows us to identify the symmetry-protected topological transition in the Kitaev chain as a confinement transition. We then compute the phase diagram of the effective mean-field theory, which correctly captures the main features of the original LGT. This is furthermore confirmed by the Green's function results and a direct comparison of the ground state energy. This simple LGT can be implemented in state-of-the art cold atom experiments. We thus also consider string-length histograms and the electric field polarization, which are easily accessible quantities in experimental setups and show that they reliably capture the various phases., Comment: 26 pages, 16 figures

This work reviews goal-oriented a posteriori error control, adaptivity and solver control for finite element approximations to boundary and initial-boundary value problems for stationary and non-stationary partial differential equations, respectively. In particular, coupled field problems with different physics may require simultaneously the accurate evaluation of several quantities of interest, which is achieved with multi-goal oriented error control. Sensitivity measures are obtained by solving an adjoint problem. Error localization is achieved with the help of a partition-of-unity. We also review and extend theoretical results for efficiency and reliability by employing a saturation assumption. The resulting adaptive algorithms allow to balance discretization and non-linear iteration errors, and are demonstrated for four applications: Poisson's problem, non-linear elliptic boundary value problems, stationary incompressible Navier-Stokes equations, and regularized parabolic $p$-Laplace initial-boundary value problems. Therein, different finite element discretizations in two different software libraries are utilized, which are partially accompanied with open-source implementations on GitHub.

Classroom observation protocols standardize the assessment of teaching effectiveness and facilitate comprehension of classroom interactions. Whereas these protocols offer teachers specific feedback on their teaching practices, the manual coding by human raters is resource-intensive and often unreliable. This has sparked interest in developing AI-driven, cost-effective methods for automating such holistic coding. Our work explores a multimodal approach to automatically estimating encouragement and warmth in classrooms, a key component of the Global Teaching Insights (GTI) study's observation protocol. To this end, we employed facial and speech emotion recognition with sentiment analysis to extract interpretable features from video, audio, and transcript data. The prediction task involved both classification and regression methods. Additionally, in light of recent large language models' remarkable text annotation capabilities, we evaluated ChatGPT's zero-shot performance on this scoring task based on transcripts. We demonstrated our approach on the GTI dataset, comprising 367 16-minute video segments from 92 authentic lesson recordings. The inferences of GPT-4 and the best-trained model yielded correlations of r = .341 and r = .441 with human ratings, respectively. Combining estimates from both models through averaging, an ensemble approach achieved a correlation of r = .513, comparable to human inter-rater reliability. Our model explanation analysis indicated that text sentiment features were the primary contributors to the trained model's decisions. Moreover, GPT-4 could deliver logical and concrete reasoning as potential teacher guidelines. Our findings provide insights into using advanced, multimodal techniques for automated classroom observation, aiming to foster teacher training through frequent and valuable feedback., Comment: Accepted as a full paper by the 25th International Conference on Artificial Intelligence in Education (AIED 2024)

Large language models (LLMs) have gained widespread interest due to their ability to process human language and perform tasks on which they have not been explicitly trained. However, we possess only a limited systematic understanding of the chemical capabilities of LLMs, which would be required to improve models and mitigate potential harm. Here, we introduce "ChemBench," an automated framework for evaluating the chemical knowledge and reasoning abilities of state-of-the-art LLMs against the expertise of chemists. We curated more than 2,700 question-answer pairs, evaluated leading open- and closed-source LLMs, and found that the best models outperformed the best human chemists in our study on average. However, the models struggle with some basic tasks and provide overconfident predictions. These findings reveal LLMs' impressive chemical capabilities while emphasizing the need for further research to improve their safety and usefulness. They also suggest adapting chemistry education and show the value of benchmarking frameworks for evaluating LLMs in specific domains.