arXiv subjects

Tobias Weber

Publications and source records attributed to Tobias Weber.

At least 19 recordsLinked to original sources

arXiv2026

Magnetic Field Control of the Néel Vector and Magnon Visibility in Altermagnetic MnTe

Altermagnetic order gives rise to momentum-dependent spin splitting of electronic and magnonic excitations even in the absence of a net magnetization. Here, we investigate the magnetic field dependence of the spin-wave spectrum of altermagnetic $α$-MnTe using inelastic neutron scattering and linear spin-wave theory. An in-plane magnetic field continuously reorients the Néel vector by overcoming the weak crystalline anisotropy, while remaining small compared with the dominant exchange scale. We find that this reorientation leaves the magnon energies and line widths essentially unchanged, but strongly modifies the measured spectral intensity through the transverse-momentum projection. Our results demonstrate a clear separation between the soft orientational degree of freedom of the antiferromagnetic order and the robust exchange-dominated chiral magnon spectrum. This combination establishes $α$-MnTe as a platform for reconfigurable magnon coupling, in which external fields tune how excitations interact with polarized probes without substantially altering their frequency or coherence.

cond-mat.str-el↗

arXiv2026

In-context learning to predict critical transitions in dynamical systems

Critical transitions - abrupt, often irreversible changes in system dynamics - arise across human and natural systems, often with catastrophic consequences. Real-world observations of such shifts remain scarce, preventing the development of reliable early warning systems. Conventional statistical and spectral indicators, such as increasing variance, tend to fail under realistic conditions of limited data and correlated noise, whereas existing deep learning classifiers do not extrapolate beyond their training data distribution. In this work, we introduce TipPFN, an in-context learning (ICL) framework that uses a prior-data fitted network to infer a system's proximity to a critical transition. Trained on our novel synthetic data generator, which is based on canonical bifurcation scenarios coupled to diverse, randomized stochastic dynamics, TipPFN flexibly capitalizes on contexts of various sizes, complexity and dimensionalities. We demonstrate robust, state-of-the-art early detection of critical transitions in previously unseen tipping regimes, sim-to-real examples, and real-world observations in both ICL and zero-shot settings.

cs.LG↗

arXiv2026

Non-reciprocal spin excitations across the skyrmion-paramagnetic phase transition in MnSi

The magnetic excitations of the skyrmion lattice in MnSi comprise a multitude of individual modes, which are non-reciprocal and thereby propagate unidirectionally. We report inelastic neutron scattering experiments for temperatures near and above the skyrmion-paramagnetic phase transition in the chiral magnet MnSi tracking the evolution from the skyrmion lattice towards the high-temperature paramagnetic state. Within the resolution of the triple-axis measurements the excitations vary smoothly across the skyrmion-paramagnetic boundary, and, the quasi-elastic paramagnetic signal under applied field retains the non-reciprocal character seen in the skyrmion phase even far above the critical temperature. Using a resolution-convolution our results are consistent with linear spin-wave theory.

cond-mat.str-el↗

arXiv2025

Non-Reciprocal Zone Boundary Magnon Propagation in Cu$_2$OSeO$_3$

Inelastic neutron scattering in the chiral magnet Cu$_2$OSeO$_3$ reveals strong non-reciprocal effects on magnon propagation at the boundary of the nuclear Brillouin zone. The non-reciprocal response is strongest at a central position between the zone corner and edge mid-point. We explain these results using an effective linear spin-wave model. While directional effects in chiral magnets have so far only been known to exist at low momenta close to the center of the Brillouin zone, the present study shows that non-reciprocity persists at the highest possible reduced momenta. The observed magnons show very little damping within the limits of our experimental resolution, making them of great interest for the fundamental research on compact, high-frequency magnonic applications.

cond-mat.str-el↗

arXiv2025

Refined spin Hamiltonian on the Cairo pentagonal lattice of Bi2Fe4O9

The frustrated magnet Bi2Fe4O9 has been reported to exhibit complex spin dynamics coexisting with conventional spin wave excitations. The magnetic Fe3+ (S = 5/2) ions are arranged into a distorted two-dimensional Cairo pentagonal lattice with weak couplings between the layers, developing long-ranged non-collinear antiferromagnetic order below 245 K. In order to enable studies and modelling of the complex dynamics close to TN, we have reexamined the magnetic excitations across the complete energy scale (0 < E < 90 meV) at 10 K. We discover two distinct gaps, which can be explained by introducing, respectively, easy axis and easy plane anisotropy on the two unequivalent Fe-sites. We develop a refined spin Hamiltonian that accurately accounts for the dispersion of essentially all spin-wave branches across the full spectral range, except around 40 meV, where a splitting and dispersion are observed. We propose that this mode is derived from phonon hybridization. Polarisation analysis shows that the system has magnetic anisotropic fluctuations, consistent with our model. A continuum of scattering is observed above the spin wave branches and is found to principally be explained by an instrumental resolution effect. The full experimental mapping of the excitation spectrum and the refined spin Hamiltonian provides a foundation for future quantitative studies of spin waves coexisting with unconventional magnetic fluctuations in this frustrated magnet found at higher temperatures.

cond-mat.str-el↗

arXiv2025

laplax -- Laplace Approximations with JAX

The Laplace approximation provides a scalable and efficient means of quantifying weight-space uncertainty in deep neural networks, enabling the application of Bayesian tools such as predictive uncertainty and model selection via Occam's razor. In this work, we introduce laplax, a new open-source Python package for performing Laplace approximations with jax. Designed with a modular and purely functional architecture and minimal external dependencies, laplax offers a flexible and researcher-friendly framework for rapid prototyping and experimentation. Its goal is to facilitate research on Bayesian neural networks, uncertainty quantification for deep learning, and the development of improved Laplace approximation techniques.

cs.LG↗

arXiv2025

Kronecker-factored Approximate Curvature (KFAC) From Scratch

Kronecker-factored approximate curvature (KFAC) is arguably one of the most prominent curvature approximations in deep learning. Its applications range from optimization to Bayesian deep learning, training data attribution with influence functions, and model compression or merging. While the intuition behind KFAC is easy to understand, its implementation is tedious: It comes in many flavours, has common pitfalls when translating the math to code, and is challenging to test, which complicates ensuring a properly functioning implementation. Some of the authors themselves have dealt with these challenges and experienced the discomfort of not being able to fully test their code. Thanks to recent advances in understanding KFAC, we are now able to provide test cases and a recipe for a reliable KFAC implementation. This tutorial is meant as a ground-up introduction to KFAC. In contrast to the existing work, our focus lies on providing both math and code side-by-side and providing test cases based on the latest insights into KFAC that are scattered throughout the literature. We hope this tutorial provides a contemporary view of KFAC that allows beginners to gain a deeper understanding of this curvature approximation while lowering the barrier to its implementation, extension, and usage in practice.

cs.LG↗

arXiv2025

Block Graph Neural Networks for tumor heterogeneity prediction

Accurate tumor classification is essential for selecting effective treatments, but current methods have limitations. Standard tumor grading, which categorizes tumors based on cell differentiation, is not recommended as a stand-alone procedure, as some well-differentiated tumors can be malignant. Tumor heterogeneity assessment via single-cell sequencing offers profound insights but can be costly and may still require significant manual intervention. Many existing statistical machine learning methods for tumor data still require complex pre-processing of MRI and histopathological data. In this paper, we propose to build on a mathematical model that simulates tumor evolution (Ożański (2017)) and generate artificial datasets for tumor classification. Tumor heterogeneity is estimated using normalized entropy, with a threshold to classify tumors as having high or low heterogeneity. Our contributions are threefold: (1) the cut and graph generation processes from the artificial data, (2) the design of tumor features, and (3) the construction of Block Graph Neural Networks (BGNN), a Graph Neural Network-based approach to predict tumor heterogeneity. The experimental results reveal that the combination of the proposed features and models yields excellent results on artificially generated data ($89.67\%$ accuracy on the test data). In particular, in alignment with the emerging trends in AI-assisted grading and spatial transcriptomics, our results suggest that enriching traditional grading methods with birth (e.g., Ki-67 proliferation index) and death markers can improve heterogeneity prediction and enhance tumor classification.

cs.CV↗

arXiv2025

Deep Weight Factorization: Sparse Learning Through the Lens of Artificial Symmetries

Sparse regularization techniques are well-established in machine learning, yet their application in neural networks remains challenging due to the non-differentiability of penalties like the $L_1$ norm, which is incompatible with stochastic gradient descent. A promising alternative is shallow weight factorization, where weights are decomposed into two factors, allowing for smooth optimization of $L_1$-penalized neural networks by adding differentiable $L_2$ regularization to the factors. In this work, we introduce deep weight factorization, extending previous shallow approaches to more than two factors. We theoretically establish equivalence of our deep factorization with non-convex sparse regularization and analyze its impact on training dynamics and optimization. Due to the limitations posed by standard training practices, we propose a tailored initialization scheme and identify important learning rate requirements necessary for training factorized networks. We demonstrate the effectiveness of our deep weight factorization through experiments on various architectures and datasets, consistently outperforming its shallow counterpart and widely used pruning methods.

cs.LG↗

arXiv2025

Linearization Turns Neural Operators into Function-Valued Gaussian Processes

Neural operators generalize neural networks to learn mappings between function spaces from data. They are commonly used to learn solution operators of parametric partial differential equations (PDEs) or propagators of time-dependent PDEs. However, to make them useful in high-stakes simulation scenarios, their inherent predictive error must be quantified reliably. We introduce LUNO, a novel framework for approximate Bayesian uncertainty quantification in trained neural operators. Our approach leverages model linearization to push (Gaussian) weight-space uncertainty forward to the neural operator's predictions. We show that this can be interpreted as a probabilistic version of the concept of currying from functional programming, yielding a function-valued (Gaussian) random process belief. Our framework provides a practical yet theoretically sound way to apply existing Bayesian deep learning methods such as the linearized Laplace approximation to neural operators. Just as the underlying neural operator, our approach is resolution-agnostic by design. The method adds minimal prediction overhead, can be applied post-hoc without retraining the network, and scales to large models and datasets. We evaluate these aspects in a case study on Fourier neural operators.

cs.LG↗

arXiv2024

Phonon dispersion of quantum paraelectric SrTiO3 in electric fields

Here we report on an elastic and inelastic neutron scattering study addressing the effect of electric fields on quantum paraelectric SrTiO3. Our elastic scattering results find small changes as a function of field in a superlattice reflection that sample the octahedral rotations, which is indicative of only weak coupling of octahedral rotation and electric polarization. By collecting not only the change in gap, but also of the dispersion, we can better quantify the changes in the lattice dynamics. The findings are put in context to recent field DFT calculations predicting the E-field effect on the atomic motions of the lowest lying transverse optical (soft) mode. We find hints of non-linear coupling to the acoustic mode and to the phonon with polarization perpendicular to the E-field, which shows the non-linearity in the chemical potential that is also relevant when strongly driving SrTiO3 with E-field (THz) pulses.

cond-mat.mtrl-sci↗

arXiv2024

Post-hoc Orthogonalization for Mitigation of Protected Feature Bias in CXR Embeddings

Purpose: To analyze and remove protected feature effects in chest radiograph embeddings of deep learning models. Methods: An orthogonalization is utilized to remove the influence of protected features (e.g., age, sex, race) in CXR embeddings, ensuring feature-independent results. To validate the efficacy of the approach, we retrospectively study the MIMIC and CheXpert datasets using three pre-trained models, namely a supervised contrastive, a self-supervised contrastive, and a baseline classifier model. Our statistical analysis involves comparing the original versus the orthogonalized embeddings by estimating protected feature influences and evaluating the ability to predict race, age, or sex using the two types of embeddings. Results: Our experiments reveal a significant influence of protected features on predictions of pathologies. Applying orthogonalization removes these feature effects. Apart from removing any influence on pathology classification, while maintaining competitive predictive performance, orthogonalized embeddings further make it infeasible to directly predict protected attributes and mitigate subgroup disparities. Conclusion: The presented work demonstrates the successful application and evaluation of the orthogonalization technique in the domain of chest X-ray image classification.

cs.LG↗

arXiv2024

Generalizing Orthogonalization for Models with Non-Linearities

The complexity of black-box algorithms can lead to various challenges, including the introduction of biases. These biases present immediate risks in the algorithms' application. It was, for instance, shown that neural networks can deduce racial information solely from a patient's X-ray scan, a task beyond the capability of medical experts. If this fact is not known to the medical expert, automatic decision-making based on this algorithm could lead to prescribing a treatment (purely) based on racial information. While current methodologies allow for the "orthogonalization" or "normalization" of neural networks with respect to such information, existing approaches are grounded in linear models. Our paper advances the discourse by introducing corrections for non-linearities such as ReLU activations. Our approach also encompasses scalar and tensor-valued predictions, facilitating its integration into neural network architectures. Through extensive experiments, we validate our method's effectiveness in safeguarding sensitive data in generalized linear models, normalizing convolutional neural networks for metadata, and rectifying pre-existing embeddings for undesired attributes.

cs.LG↗

arXiv2024

Post-Training Network Compression for 3D Medical Image Segmentation: Reducing Computational Efforts via Tucker Decomposition

We address the computational barrier of deploying advanced deep learning segmentation models in clinical settings by studying the efficacy of network compression through tensor decomposition. We propose a post-training Tucker factorization that enables the decomposition of pre-existing models to reduce computational requirements without impeding segmentation accuracy. We applied Tucker decomposition to the convolutional kernels of the TotalSegmentator (TS) model, an nnU-Net model trained on a comprehensive dataset for automatic segmentation of 117 anatomical structures. Our approach reduced the floating-point operations (FLOPs) and memory required during inference, offering an adjustable trade-off between computational efficiency and segmentation quality. This study utilized the publicly available TS dataset, employing various downsampling factors to explore the relationship between model size, inference speed, and segmentation performance. The application of Tucker decomposition to the TS model substantially reduced the model parameters and FLOPs across various compression rates, with limited loss in segmentation accuracy. We removed up to 88% of the model's parameters with no significant performance changes in the majority of classes after fine-tuning. Practical benefits varied across different graphics processing unit (GPU) architectures, with more distinct speed-ups on less powerful hardware. Post-hoc network compression via Tucker decomposition presents a viable strategy for reducing the computational demand of medical image segmentation models without substantially sacrificing accuracy. This approach enables the broader adoption of advanced deep learning technologies in clinical practice, offering a way to navigate the constraints of hardware capabilities.

eess.IV↗

arXiv2023

Adversarial Anomaly Detection using Gaussian Priors and Nonlinear Anomaly Scores

Anomaly detection in imbalanced datasets is a frequent and crucial problem, especially in the medical domain where retrieving and labeling irregularities is often expensive. By combining the generative stability of a $β$-variational autoencoder (VAE) with the discriminative strengths of generative adversarial networks (GANs), we propose a novel model, $β$-VAEGAN. We investigate methods for composing anomaly scores based on the discriminative and reconstructive capabilities of our model. Existing work focuses on linear combinations of these components to determine if data is anomalous. We advance existing work by training a kernelized support vector machine (SVM) on the respective error components to also consider nonlinear relationships. This improves anomaly detection performance, while allowing faster optimization. Lastly, we use the deviations from the Gaussian prior of $β$-VAEGAN to form a novel anomaly score component. In comparison to state-of-the-art work, we improve the $F_1$ score during anomaly detection from 0.85 to 0.92 on the widely used MITBIH Arrhythmia Database.

cs.LG↗

arXiv2023

Constrained Probabilistic Mask Learning for Task-specific Undersampled MRI Reconstruction

Undersampling is a common method in Magnetic Resonance Imaging (MRI) to subsample the number of data points in k-space, reducing acquisition times at the cost of decreased image quality. A popular approach is to employ undersampling patterns following various strategies, e.g., variable density sampling or radial trajectories. In this work, we propose a method that directly learns the undersampling masks from data points, thereby also providing task- and domain-specific patterns. To solve the resulting discrete optimization problem, we propose a general optimization routine called ProM: A fully probabilistic, differentiable, versatile, and model-free framework for mask optimization that enforces acceleration factors through a convex constraint. Analyzing knee, brain, and cardiac MRI datasets with our method, we discover that different anatomic regions reveal distinct optimal undersampling masks, demonstrating the benefits of using custom masks, tailored for a downstream task. For example, ProM can create undersampling masks that maximize performance in downstream tasks like segmentation with networks trained on fully-sampled MRIs. Even with extreme acceleration factors, ProM yields reasonable performance while being more versatile than existing methods, paving the way for data-driven all-purpose mask generation.

eess.IV↗

arXiv2023

Automated wildlife image classification: An active learning tool for ecological applications

Wildlife camera trap images are being used extensively to investigate animal abundance, habitat associations, and behavior, which is complicated by the fact that experts must first classify the images manually. Artificial intelligence systems can take over this task but usually need a large number of already-labeled training images to achieve sufficient performance. This requirement necessitates human expert labor and poses a particular challenge for projects with few cameras or short durations. We propose a label-efficient learning strategy that enables researchers with small or medium-sized image databases to leverage the potential of modern machine learning, thus freeing crucial resources for subsequent analyses. Our methodological proposal is two-fold: (1) We improve current strategies of combining object detection and image classification by tuning the hyperparameters of both models. (2) We provide an active learning (AL) system that allows training deep learning models very efficiently in terms of required human-labeled training images. We supply a software package that enables researchers to use these methods directly and thereby ensure the broad applicability of the proposed framework in ecological practice. We show that our tuning strategy improves predictive performance. We demonstrate how the AL pipeline reduces the amount of pre-labeled data needed to achieve a specific predictive performance and that it is especially valuable for improving out-of-sample predictive performance. We conclude that the combination of tuning and AL increases predictive performance substantially. Furthermore, we argue that our work can broadly impact the community through the ready-to-use software package provided. Finally, the publication of our models tailored to European wildlife data enriches existing model bases mostly trained on data from Africa and North America.

cs.CV↗

arXiv2023

Neutron polarisation correction to triple-axis data with analytical derivations

Polarised neutron scattering is the method of choice to study magnetism in condensed matter. Polarised neutrons are typically very low in flux, and complex experimental configurations further reduce the count rate. Neutron polarisation corrections would therefore be needed. Here we analytically derive formulae of the corrected partial differential scattering cross-sections. The analytical method is designed for the longitudinal polarisation analysis, and the correction generally holds for time-independent polarised neutrons with a triple-axis spectrometer. We then apply the correction to recent results of our $P_x$ experiment on Y$_3$Fe$_5$O$_{12}$. Although there is a difficulty with the experimental determination of inefficiency parameters of neutron spin polarisers and flippers, the correction appears to work properly.

cond-mat.str-el↗