Search

6D object pose estimation suffers from reduced accuracy when applied to metallic objects. We set out to improve the state-of-the-art by addressing challenges such as reflections and specular highlights in industrial applications. Our novel BOP-compatible dataset, featuring a diverse set of metallic objects (cans, household, and industrial items) under various lighting and background conditions, provides additional geometric and visual cues. We demonstrate that these cues can be effectively leveraged to enhance overall performance. To illustrate the usefulness of the additional features, we improve upon the GDRNPP algorithm by introducing an additional keypoint prediction and material estimator head in order to improve spatial scene understanding. Evaluations on the new dataset show improved accuracy for metallic objects, supporting the hypothesis that additional geometric and visual cues can improve learning.

With the rapidly growing availability of machine-learned interatomic potential (MLIP) models for chemistry, much current research focuses on the development of generally applicable and ``foundational'' MLIPs. An important question in this context is whether, and how well, such models can transfer from one application domain to another. Here, we assess this transferability for an MLIP model at the interface of materials and molecular chemistry. Specifically, we study GO-MACE-23, a model designed for the extended covalent network of graphene oxide, and quantify its zero-shot performance for small, isolated molecules and chemical reactions outside its direct scope--in direct comparison with a state-of-the-art model which has been trained in-domain. Our work provides quantitative insight into the transfer and generalisation ability of graph-neural-network potentials and, more generally, makes a step towards the more widespread applicability of MLIPs in chemistry.

We propose a new class of multi-layer iterative schemes for solving sparse linear systems in saddle point structure. The new scheme consist of an iterative preconditioner that is based on the (approximate) nullspace method, combined with an iterative least squares approach and an iterative projection method. We present a theoretical analysis and demonstrate the effectiveness and robustness of the new scheme on sparse matrices from various applications.

Point defects in solid-state quantum systems are vital for enabling single-photon emission at specific wavelengths, making their precise identification essential for advancing applications in quantum technologies. However, pinpointing the microscopic origins of these defects remains a challenge. In this work, we propose Raman spectroscopy as a robust strategy for defect identification. Using density functional theory, we systematically characterize the Raman signatures of 100 defects in hexagonal boron nitride (hBN) spanning periodic groups III to VI, encompassing around 30,000 phonon modes. Our findings reveal that the local atomic environment plays a pivotal role in shaping the Raman lineshape, enabling the narrowing of potential defect candidates. Furthermore, we demonstrate that Raman spectroscopy can differentiate defects based on their spin and charge states as well as strain-induced variations, implying the versatility of this approach. Therefore, this study not only provides a comprehensive theoretical database of Raman spectra for hBN defects but also establishes a novel experiment framework for using tip-enhanced Raman spectroscopy to identify point defects. More broadly, our approach offers a universal method for defect identification in any quantum materials., Comment: 13 pages, 4 figures

In the age of AI-driven generative technologies, traditional biometric recognition systems face unprecedented challenges, particularly from sophisticated deepfake and face reenactment techniques. In this study, we propose a Two-Stream Spatial-Temporal Transformer Framework for person identification using upper body keypoints visible during online conversations, which we term conversational keypoints. Our framework processes both spatial relationships between keypoints and their temporal evolution through two specialized branches: a Spatial Transformer (STR) that learns distinctive structural patterns in keypoint configurations, and a Temporal Transformer (TTR) that captures sequential motion patterns. Using the state-of-the-art Sapiens pose estimator, we extract 133 keypoints (based on COCO-WholeBody format) representing facial features, head pose, and hand positions. The framework was evaluated on a dataset of 114 individuals engaged in natural conversations, achieving recognition accuracies of 80.12% for the spatial stream, 63.61% for the temporal stream. We then explored two fusion strategies: a shared loss function approach achieving 82.22% accuracy, and a feature-level fusion method that concatenates feature maps from both streams, significantly improving performance to 94.86%. By jointly modeling both static anatomical relationships and dynamic movement patterns, our approach learns comprehensive identity signatures that are more robust to spoofing than traditional appearance-based methods.

We introduce the THESAN-ZOOM project, a comprehensive suite of high-resolution zoom-in simulations of $14$ high-redshift ($z>3$) galaxies selected from the THESAN simulation volume. This sample encompasses a diverse range of halo masses, with $M_\mathrm{halo} \approx 10^8 - 10^{13}~\mathrm{M}_\odot$ at $z=3$. At the highest-resolution, the simulations achieve a baryonic mass of $142~\mathrm{M}_\odot$ and a gravitational softening length of $17~\mathrm{cpc}$. We employ a state-of-the-art multi-phase interstellar medium (ISM) model that self-consistently includes stellar feedback, radiation fields, dust physics, and low-temperature cooling through a non-equilibrium thermochemical network. Our unique framework incorporates the impact of patchy reionization by adopting the large-scale radiation field topology from the parent THESAN simulation box rather than assuming a spatially uniform UV background. In total, THESAN-ZOOM comprises $60$ simulations, including both fiducial runs and complementary variations designed to investigate the impact of numerical and physical parameters on galaxy properties. The fiducial simulation set reproduces a wealth of high-redshift observational data such as the stellar-to-halo-mass relation, the star-forming main sequence, the Kennicutt-Schmidt relation, and the mass-metallicity relation. While our simulations slightly overestimate the abundance of low-mass and low-luminosity galaxies they agree well with observed stellar and UV luminosity functions at the higher mass end. Moreover, the star-formation rate density closely matches the observational estimates from $z=3-14$. These results indicate that the simulations effectively reproduce many of the essential characteristics of high-redshift galaxies, providing a realistic framework to interpret the exciting new observations from JWST., Comment: 25 pages, 16 figures, Please visit https://thesan-project.com for more details, Comments welcome

It is the purpose of this paper to give a characterisation of circular light rays in a plasma on an axially symmetric and stationary spacetime. We restrict to the case of an unmagnetised, pressure-free electron-ion plasma and we assume that the plasma shares the symmetry of the spacetime. As a main tool we use two potentials, one for prograde and one for retrograde light rays, whose critical points are exactly the circular light rays in the plasma. In the case that the plasma density vanishes, the corresponding equipotential surfaces reduce to the relativistic Von Zeipel cylinders which have been discussed in many papers since the 1970s. In a plasma, the gradients of the potentials give the centrifugal and the Coriolis forces experienced by a light ray, where the plasma has an influence only on the centrifugal force. The introduction of these potentials allows us to generalise topological methods that have been successfully used for proving the existence or non-existence of circular vacuum light rays to the plasma case. The general results are illustrated with examples on Minkowski, Schwarzschild, Kerr and NUT spacetimes., Comment: 33 pages, 6 figures

Lung cancer is the second most common cancer and the leading cause of cancer-related deaths worldwide. Survival largely depends on tumor stage at diagnosis, and early detection with low-dose CT can significantly reduce mortality in high-risk patients. AI can improve the detection, measurement, and characterization of pulmonary nodules while reducing assessment time. However, the training data, functionality, and performance of available AI systems vary considerably, complicating software selection and regulatory evaluation. Manufacturers must specify intended use and provide test statistics, but they can choose their training and test data, limiting standardization and comparability. Under the EU AI Act, consistent quality assurance is required for AI-based nodule detection, measurement, and characterization. This position paper proposes systematic quality assurance grounded in a validated reference dataset, including real screening cases plus phantom data to verify volume and growth rate measurements. Regular updates shall reflect demographic shifts and technological advances, ensuring ongoing relevance. Consequently, ongoing AI quality assurance is vital. Regulatory challenges are also adressed. While the MDR and the EU AI Act set baseline requirements, they do not adequately address self-learning algorithms or their updates. A standardized, transparent quality assessment - based on sensitivity, specificity, and volumetric accuracy - enables an objective evaluation of each AI solution's strengths and weaknesses. Establishing clear testing criteria and systematically using updated reference data lay the groundwork for comparable performance metrics, informing tenders, guidelines, and recommendations., Comment: 12 pages incl. 2 figures, 2 charts, and references, summary in English (page 2), article in German (original title: Anforderungen an die Qualit\"atssicherung von KI-Modellen f\"ur die Lungenkrebs-Fr\"uherkennung)

Multimodal large language models (MLLMs) have demonstrated strong performance in understanding videos holistically, yet their ability to process streaming videos-videos are treated as a sequence of visual events-remains underexplored. Intuitively, leveraging past events as memory can enrich contextual and temporal understanding of the current event. In this paper, we show that leveraging memories as contexts helps MLLMs better understand video events. However, because such memories rely on predictions of preceding events, they may contain misinformation, leading to confabulation and degraded performance. To address this, we propose a confabulation-aware memory modification method that mitigates confabulated memory for memory-enhanced event understanding., Comment: Short paper (5 pages)

In industrial applications requiring real-time feedback, such as quality control and robotic manipulation, the demand for high-speed and accurate pose estimation remains critical. Despite advances improving speed and accuracy in pose estimation, finding a balance between computational efficiency and accuracy poses significant challenges in dynamic environments. Most current algorithms lack scalability in estimation time, especially for diverse datasets, and the state-of-the-art (SOTA) methods are often too slow. This study focuses on developing a fast and scalable set of pose estimators based on GDRNPP to meet or exceed current benchmarks in accuracy and robustness, particularly addressing the efficiency-accuracy trade-off essential in real-time scenarios. We propose the AMIS algorithm to tailor the utilized model according to an application-specific trade-off between inference time and accuracy. We further show the effectiveness of the AMIS-based model choice on four prominent benchmark datasets (LM-O, YCB-V, T-LESS, and ITODD).

Modern high-resolution simulations of the interstellar medium (ISM) have shown that key factors in governing star formation are the competing influences of radiative dissipation, pressure support driven by stellar feedback, and the relentless pull of gravity. Cosmological simulations of galaxy formation, such as IllustrisTNG or ASTRID, are however not able to resolve this physics in detail and therefore need to rely on approximate treatments. These have often taken the form of empirical subgrid models of the ISM expressed in terms of an effective equation of state (EOS) that relates the mean ISM pressure to the mean gas density. Here we seek to improve these heuristic models by directly fitting their key ingredients to results of the high-resolution TIGRESS simulations, which have shown that the dynamical equilibrium of the ISM can be understood in terms of a pressure-regulated, feedback modulated (PRFM) model for star formation. Here we explore a simple subgrid model that draws on the PRFM concept but uses only local quantities. It accurately reproduces PRFM for pure gas disks, while it predicts slightly less star formation than PRFM in the presence of an additional thin stellar disk. We compare the properties of this model with the older Springel and Hernquist and TNG prescriptions, and apply all three to isolated simulations of disk galaxies as well as to a set of high-resolution zoom-in simulations carried out with a novel 'multi-zoom' technique that we introduce in this study. The softer EOS implied by TIGRESS produces substantially thinner disk galaxies, which has important ramifications for disk stability and galaxy morphology. The total stellar mass of galaxies is however hardly modified at low redshift, reflecting the dominating influence of large-scale gaseous inflows and outflows to galaxies, which are not sensitive to the EOS itself, Comment: 22 pages, 21 figures, to be submitted to MNRAS. This is a Learning the Universe publication

Accretion is the dominant contribution to the cosmic massive black hole density in the Universe today. Yet, modelling it in cosmological simulations is challenging due to the dynamic range involved, as well as the theoretical uncertainties of the underlying mechanisms driving accretion from galactic to black hole horizon scales. We present a simple, flexible parametrization for gas inflows onto massive black holes in order to manage this uncertainty in large-volume cosmological simulations. This is done as part of the "Learning the Universe'' collaboration, which aims to jointly infer the initial conditions and physical processes governing the evolution of the Universe using a Bayesian forward-modelling approach. To allow such a forward-modelling, we update the prescription for accretion with a two-parameter free-fall based inflow estimate that allows for a radius-dependent inflow rate and add a simple model for unresolved accretion disks. We use uniform resolution cosmological hydrodynamical simulations and the IllustrisTNG framework to study the massive black hole population and its dependence on the introduced model parameters. Once the parameters of the accretion formula are chosen to result in a roughly similar redshift zero black hole mass density, the differences caused by the details in the accretion formula are moderate in the supermassive black hole regime, indicating that it is difficult to distinguish between accretion mechanisms based on luminous active galactic nuclei powered by supermassive black holes. Applying the same models to intermediate mass black holes at high redshift, however, reveals significantly different accretion rates in high redshift, moderate luminosity active galactic nuclei and different frequencies and mass distributions of intermediate mass black hole mergers for the same black hole formation model., Comment: 19 pages, 13 figures, submitted to A&A, comments welcome. This is a Learning the Universe publication

Visual instruction tuning refines pre-trained Multimodal Large Language Models (MLLMs) to enhance their real-world task performance. However, the rapid expansion of visual instruction datasets introduces significant data redundancy, leading to excessive computational costs. Existing data selection methods predominantly rely on proxy models or loss-based metrics, both of which impose substantial computational overheads due to the necessity of model inference and backpropagation. To address this challenge, we propose PRISM, a novel training-free approach for efficient multimodal data selection. Unlike existing methods, PRISM eliminates the reliance on proxy models, warm-up pretraining, and gradient-based optimization. Instead, it leverages Pearson correlation analysis to quantify the intrinsic visual encoding properties of MLLMs, computing a task-specific correlation score to identify high-value instances. This not only enbles data-efficient selection,but maintains the original performance. Empirical evaluations across multiple MLLMs demonstrate that PRISM reduces the overall time required for visual instruction tuning and data selection to just 30% of conventional methods, while surpassing fully fine-tuned models across eight multimodal and three language understanding benchmarks, achieving a 101.7% relative improvement in final performance.

We present a construction method for triharmonic maps to spheres. In particular, we show that for any $m\in\mathbb{N}$ with $m\geq 3$ there exists a triharmonic map from $\mathbb{R}^m\setminus\{0\}$ into a round sphere. In addition, we provide a construction method for proper $r$-harmonic maps between spheres based on a suitable deformation of eigenmaps.

We examine the light curves of a sample of novae, classifying them into single-peaked and multiple-peaked morphologies. Using accurate distances from Gaia, we determine the spatial distribution of these novae by computing their heights, $Z$, above the Galactic plane. We show that novae exhibiting a single peak in their light curves tend to concentrate near the Galactic plane, while those displaying multiple peaks are more homogeneously distributed, reaching heights up to 1000 pc above the plane. A KS test rejects the null hypothesis that the two distributions originate from the same population at a significance level corresponding to $4.2\sigma$., Comment: Accepted for publication in the Astrophysical Journal

The first hyperon was discovered about 70 years ago, but the nature of these particles, particularly with regard to multistrange hyperons, and many of their properties can still be considered to be literally strange. A dedicated and successful global spectroscopy program in the 1960s and 1970s using $K^-$ beams revealed many multistrange candidates, but the available evidence of their existence is statistically limited. For this reason, there is still much to learn about the systematics of the spectrum of excited hyperon states and what they have in common with their non-strange companions, or how they differ from the nucleon and $\Delta$ resonances. Recent years have also seen a great deal of progress in the field of charmed and bottom baryon spectroscopy. Unprecedented data from the Large Hadron Collider in particular indicate continued rapid progress in the field of bottom baryons. On the theoretical side, baryons with one heavy quark $Q$ and a light $qq$ system serve as an ideal laboratory for studying light $qq$ (diquark) correlations and the dynamics of the light quarks in the colour environment of a heavy quark. In this review, we discuss the status of doubly and triply strange $\Xi$ as well as $\Omega$ baryons, and the properties of all the known charmed and bottom states. The comparison of the two heavy sectors reveals many similarities as predicted by heavy-quark symmetries, together with differences in mass splittings easily understood by potential models. The multi-strange hyperons bridge the under-explored gap between the light- and the heavy-flavour baryons. How do the properties of a singly charmed $Q$-$qq$ system change with decreasing mass of the heavy quark in the transition to a doubly strange $q$-$QQ$ system with a heavier quark-quark system relative to one light quark?, Comment: Review article published by Reports on Progress in Physics, 87 pages, 24 figures

Computer simulations have long been key to understanding and designing phase-change materials (PCMs) for memory technologies. Machine learning is now increasingly being used to accelerate the modelling of PCMs, and yet it remains challenging to simultaneously reach the length and time scales required to simulate the operation of real-world PCM devices. Here, we show how ultra-fast machine-learned interatomic potentials, based on the atomic cluster expansion (ACE) framework, enable simulations of PCMs reflecting applications in devices with excellent scalability on high-performance computing platforms. We report full-cycle simulations -- including the time-consuming crystallisation process (from digital "zeroes" to "ones") -- thus representing the entire programming cycle for cross-point memory devices. We also showcase a simulation of full-cycle operations, relevant to neuromorphic computing, in a mushroom-type device geometry. Our work provides a springboard for the atomistic modelling of PCM-based memory and neuromorphic computing devices -- and, more widely, it illustrates the power of highly efficient ACE ML models for materials science and engineering.

Observations have established that the masses of supermassive black holes (SMBHs) correlate tightly with the stellar masses of their host galaxies, albeit with substantial scatter. The size of this scatter as a function of galaxy mass and redshift contains valuable information about the origin of SMBHs and the physical nature of their co-evolution with galaxies. In this work, we highlight this connection by studying the scatter in the $M_{\mathrm{BH}} - M_{\star}$ relation for massive galaxies in the Illustris, IllustrisTNG (TNG), and EAGLE cosmological simulations. We find that the scatter in TNG is significantly lower than in Illustris and EAGLE, reflecting their different BH feedback models. By performing various numerical experiments, we quantify different contributions to the scatter in the simulations, and also identify a suitably defined intrinsic scatter. The intrinsic scatter in Illustris and EAGLE is $\sim0.3$ dex at $z=0$, and is dominated by variations from BH accretion, whereas the smaller scatter of TNG is rather dominated by hierarchical merging, suggesting that the massive galaxies in TNG are more tightly quenched. Variations in the BH seed mass can contribute to the scatter of the $M_{\rm BH}-M_{\star}$ relation as well, but whether this still plays a role at $z=0$ depends on the feedback model. Simulations with disabled AGN feedback produce much higher scatter for low-mass galaxies than seen in our cosmological simulations, demonstrating the crucial influence of feedback for determining the co-evolution of SMBHs and their host galaxies in this regime. In contrast, an important factor in reducing the scatter for massive galaxies is hierarchical merging of mostly quenched systems. Based on our results, we expect that the scatter in the $M_{\mathrm{BH}} - M_{\star}$ relation at high redshift could be particularly powerful in providing clues to the origin of SMBHs., Comment: 16 pages, 13 figures, submit to MNRAS, comments welcome

This article explores the requirements for corpus compilation within the GiesKaNe project (University of Giessen and Kassel, Syntactic Basic Structures of New High German). The project is defined by three central characteristics: it is a reference corpus, a historical corpus, and a syntactically deeply annotated treebank. As a historical corpus, GiesKaNe aims to establish connections with both historical and contemporary corpora, ensuring its relevance across temporal and linguistic contexts. The compilation process strikes the balance between innovation and adherence to standards, addressing both internal project goals and the broader interests of the research community. The methodological complexity of such a project is managed through a complementary interplay of human expertise and machine-assisted processes. The article discusses foundational topics such as tokenization, normalization, sentence definition, tagging, parsing, and inter-annotator agreement, alongside advanced considerations. These include comparisons between grammatical models, annotation schemas, and established de facto annotation standards as well as the integration of human and machine collaboration. Notably, a novel method for machine-assisted classification of texts along the continuum of conceptual orality and literacy is proposed, offering new perspectives on text selection. Furthermore, the article introduces an approach to deriving de facto standard annotations from existing ones, mediating between standardization and innovation. In the course of describing the workflow the article demonstrates that even ambitious projects like GiesKaNe can be effectively implemented using existing research infrastructure, requiring no specialized annotation tools. Instead, it is shown that the workflow can be based on the strategic use of a simple spreadsheet and integrates the capabilities of the existing infrastructure.

Koch and Oesterreicher's model of "N\"ahe und Distanz" (N\"ahe = immediacy, conceptual orality; Distanz = distance, conceptual literacy) is constantly used in German linguistics. However, there is no statistical foundation for use in corpus linguistic analyzes, while it is increasingly moving into empirical corpus linguistics. Theoretically, it is stipulated, among other things, that written texts can be rated on a scale of conceptual orality and literacy by linguistic features. This article establishes such a scale based on PCA and combines it with automatic analysis. Two corpora of New High German serve as examples. When evaluating established features, a central finding is that features of conceptual orality and literacy must be distinguished in order to rank texts in a differentiated manner. The scale is also discussed with a view to its use in corpus compilation and as a guide for analyzes in larger corpora. With a theory-driven starting point and as a "tailored" dimension, the approach compared to Biber's Dimension 1 is particularly suitable for these supporting, controlling tasks.

The Standard Model of elementary particle physics is one of the most successful models of contemporary theoretical physics being in full agreement with experiments. However, its mathematical structure deserves further investigations both from a geometric and an analytic point of view. The aim of this manuscript is to provide a mathematically well-defined and self-contained description of the Standard Model in terms of gauge theory and differential geometry on globally hyperbolic manifolds. Within this setup we then prove the existence of a global solution for the Euler-Lagrange equations of the Standard Model (with the conformal Higgs potential) under the assumptions that the globally hyperbolic manifold is a four-dimensional spacetime of expanding type and small initial data. This is achieved by establishing a gauge-invariant energy estimate which is of independent mathematical interest., Comment: All comments are welcome!

Secure resource management (SRM) within a cloud computing environment is a critical yet infrequently studied research topic. This paper provides a comprehensive survey and comparative performance evaluation of potential cyber threat countermeasure strategies that address security challenges during cloud workload execution and resource management. Cybersecurity is explored specifically in the context of cloud resource management, with an emphasis on identifying the associated challenges. The cyber threat countermeasure methods are categorized into three classes: defensive strategies, mitigating strategies, and hybrid strategies. The existing countermeasure strategies belonging to each class are thoroughly discussed and compared. In addition to conceptual and theoretical analysis, the leading countermeasure strategies within these categories are implemented on a common platform and examined using two real-world virtual machine (VM) data traces. Based on this comprehensive study and performance evaluation, the paper discusses the trade-offs among these countermeasure strategies and their utility, providing imperative concluding remarks on the holistic study of cloud cyber threat countermeasures and secure resource management. Furthermore, the study suggests future methodologies that could effectively address the emerging challenges of secure cloud resource management., Comment: 16 Pages, 12 Figures, 6 Tables, in IEEE Transactions on Systems, Man, and Cybernetics: Systems, 2025

In this contribution, we give an overview of the ELPA library and ELSI interface, which are crucial elements for large-scale electronic structure calculations in FHI-aims. ELPA is a key solver library that provides efficient solutions for both standard and generalized eigenproblems, which are central to the Kohn-Sham formalism in density functional theory (DFT). It supports CPU and GPU architectures, with full support for NVIDIA and AMD GPUs, and ongoing development for Intel GPUs. Here we also report the results of recent optimizations, leading to significant improvements in GPU performance for the generalized eigenproblem. ELSI is an open-source software interface layer that creates a well-defined connection between "user" electronic structure codes and "solver" libraries for the Kohn-Sham problem, abstracting the step between Hamilton and overlap matrices (as input to ELSI and the respective solvers) and eigenvalues and eigenvectors or density matrix solutions (as output to be passed back to the "user" electronic structure code). In addition to ELPA, ELSI supports solvers including LAPACK and MAGMA, the PEXSI and NTPoly libraries (which bypass an explicit eigenvalue solution), and several others., Comment: Contribution to the upcoming Roadmap for Advancements of the FHI-aims Software Package

We identify a chain of galaxies along an almost straight line in the nearby Universe with a projected length of ~5 Mpc. The galaxies are distributed within projected distances of only 7-105 kpc from the axis of the identified filament. They have redshifts in a very small range of z=0.0361-0.0370 so that their radial velocities are consistent with galaxy proper motions. The filament galaxies are mainly star-forming and have stellar masses in a range of $\rm 10^{9.1}-10^{10.7}\,M_{\odot}$. We search for systems with similar geometrical properties in the full-sky mock galaxy catalogue of the MillenniumTNG simulations and find that although such straight filaments are unusual and rare, they are predicted by $\Lambda$CDM simulations (4% incidence). We study the cold HI gas in a 1.3 Mpc section of the filament through HI-21cm emission line observations and detect eleven HI sources, many more than expected from the HI mass function in a similar volume. They have HI masses $\rm 10^{8.5}-10^{9.5}\,M_{\odot}$ and are mostly within ~120 kpc projected distance from the filament axis. None of these HI sources has a confirmed optical counterpart. Their darkness together with their large HI-21cm line-widths indicate that they contain gas that might not yet be virialized. These clouds must be marking the peaks of the dark matter and HI distributions over large scales within the filament. The presence of such gas clouds around the filament spines is predicted by simulations, but this is the first time that the existence of such clouds in a filament is observationally confirmed., Comment: Published in ApJL. Main figures: Fig. 1 and Fig. 3

A main source of bias in transmission spectroscopy of exoplanet atmospheres is magnetic activity of the host star in the form of stellar spots, faculae or flares. However, the fact that main-sequence stars have a chromosphere and a corona, and that these optically thin layers are dominated by line emission may alter the global interpretation of the planetary spectrum, has largely been neglected. Using a JWST NIRISS/SOSS data set of hot Jupiter HAT-P-18 b, we show that even at near-IR and IR wavelengths, the presence of these layers leads to significant changes in the transmission spectrum of the planetary atmosphere. Accounting for these stellar outer layers thus improves the atmospheric fit of HAT-P-18 b, and increases its best-fit atmospheric temperature from 536 K to 736 K, a value much closer to the predicted equilibrium temperature of 852 K. Our analysis also decreases the best-fit abundance of CO2 by almost an order of magnitude. The approach provides a new window to the properties of chromospheres/corona in stars other than our Sun., Comment: 13 pages, 9 figures

We prove absence of ground states in the infrared-divergent spin boson model at large coupling. Our key argument reduces the proof to verifying long range order in the dual one-dimensional continuum Ising model, i.e., to showing that the respective two point function is lower bounded by a strictly positive constant. We can then use known results from percolation theory to establish long range order at large coupling. Combined with the known existence of ground states at small coupling, our result proves that the spin boson model undergoes a phase transition with respect to the coupling strength. We also present an expansion for the vacuum overlap of the spin boson ground state in terms of the Ising $n$-point functions, which implies that the phase transition is unique, i.e., that there is a critical coupling constant below which a ground state exists and above which none can exist., Comment: 17 pages

One of the primary objectives of the SPHEREx mission is to understand the origin of molecules such as H2O, CO2, and other volatile compounds at the early stages of planetary system formation. Because the vast majority of these compounds -- typically exceeding 95% -- exist in the solid phase rather than the gaseous phase in the systems of concern here, the observing strategy planned to characterize them is slightly unusual. Specifically, SPHEREx will target highly obscured sources throughout the Milky Way, and observe the species of concern in absorption against background illumination. SPHEREx spectrophotometry will yield ice column density measurements for millions of obscured Milky Way sources of all ages and types. By correlating those column densities with source ages, the SPHEREx mission will shed light on whether those molecules were formed in situ along with their nascent stellar systems, or whether instead they formed elsewhere and were introduced into those systems after their formation. To that end, this work describes version 7$.$1 of the SPHEREx Target List of Ice Sources (SPLICES) for the community. It contains about 8$.$6 million objects brighter than W2~12 Vega mag over much of the sky, principally within a broad strip running the length of the Milky Way midplane, but also within high-latitude molecular clouds and even the Magellanic Clouds., Comment: Published by ApJ. 21 pages, 6 figures. This article documents the original version of SPLICES (7.1). The current version as well as the complete catalog is publicly available along with release notes documenting all additions and changes at the NASA/IPAC Infrared Science Archive (IRSA) at this URL: https://irsa.ipac.caltech.edu/data/SPHEREx/SPLICES/

This paper represents one contribution to a larger Roadmap article reviewing the current status of the FHI-aims code. In this contribution, the implementation of density-functional perturbation theory in a numerical atom-centered framework is summarized. Guidelines on usage and links to tutorials are provided.

Reproducibility and replicability are critical pillars of empirical research, particularly in machine learning, where they depend not only on the availability of models, but also on the datasets used to train and evaluate those models. In this paper, we introduce the Construction Industry Steel Ordering List (CISOL) dataset, which was developed with a focus on transparency to ensure reproducibility, replicability, and extensibility. CISOL provides a valuable new research resource and highlights the importance of having diverse datasets, even in niche application domains such as table extraction in civil engineering. CISOL is unique in that it contains real-world civil engineering documents from industry, making it a distinctive contribution to the field. The dataset contains more than 120,000 annotated instances in over 800 document images, positioning it as a medium-sized dataset that provides a robust foundation for Table Structure Recognition (TSR) and Table Detection (TD) tasks. Benchmarking results show that CISOL achieves 67.22 mAP@0.5:0.95:0.05 using the YOLOv8 model, outperforming the TSR-specific TATR model. This highlights the effectiveness of CISOL as a benchmark for advancing TSR, especially in specialized domains., Comment: Accepted at WACV2025

This paper studies the effective convergence of iterative methods for solving convex minimization problems using block Gauss--Seidel algorithms. It investigates whether it is always possible to algorithmically terminate the iteration in such a way that the outcome of the iterative algorithm satisfies any predefined error bound. It is shown that the answer is generally negative. Specifically, it is shown that even if a computable continuous function which is convex in each variable possesses computable minimizers, a block Gauss--Seidel iterative method might not be able to effectively compute any of these minimizers. This means that it is impossible to algorithmically terminate the iteration such that a given performance guarantee is satisfied. The paper discusses two reasons for this behavior. First, it might happen that certain steps in the Gauss--Seidel iteration cannot be effectively implemented on a digital computer. Second, all computable minimizers of the problem may not be reachable by the Gauss--Seidel method. Simple and concrete examples for both behaviors are provided.

The masses of Population III stars are largely unconstrained since no simulations exist that take all relevant primordial star formation physics into account. We perform the first suite of radiation magnetohydrodynamics (RMHD) simulations of Population III star formation, with the POPSICLE project. Compared to control simulations that only include magnetic fields (MHD), protostellar ionizing and dissociating feedback, or neither, the RMHD simulation best resembles the MHD simulation during the earliest stages of collapse and star formation. In $5000\,\rm{yrs}$, the mass of the most massive star is $65\,\rm{M_{\odot}}$ in the RMHD simulation, compared to $120\,\rm{M_{\odot}}$ in simulations without magnetic fields. This difference arises because magnetic fields act against gravity, suppress mass transport, and reduce compressional heating. The maximum stellar mass of Population III stars is thus already limited by magnetic fields, even before accretion rates drop to allow significant protostellar radiative feedback. Following classical main sequence stellar evolution with MESA reveals that it is difficult to create Population III stars with masses larger than $600\,\rm{M_{\odot}}$ in typical dark matter minihaloes at $z \gtrsim 20$, with maximum stellar masses $\sim 100\,\rm{M_{\odot}}$ more likely due to expected negative feedback from both magnetic fields and stellar radiation. This work lays the first step in building a full physics-informed mass function of Population III stars., Comment: 5 pages + appendix, submitted to MNRAS

We introduce the AIDA-TNG project, a suite of cosmological magnetohydrodynamic simulations that simultaneously model galaxy formation and different variations of the underlying dark matter model. We consider the standard cold dark matter model and five variations, including three warm dark matter scenarios and two self-interacting models with constant or velocity-dependent cross-section. In each model, we simulate two cosmological boxes of 51.7 and 110.7 Mpc on a side, with the same initial conditions as TNG50 and TNG100, and combine the variations in the physics of dark matter with the fiducial IllustrisTNG galaxy formation model. The AIDA-TNG runs are thus ideal for studying the simultaneous effect of baryons and alternative dark matter models on observable properties of galaxies and large-scale structures. We resolve haloes in the range between $10^{8}$ and $4\times10^{14}\,$M$_{\odot}$ and scales down to the nominal resolution of 570 pc in the highest resolution runs. This work presents the first results on statistical quantities such as the halo mass function and the matter power spectrum; we quantify the modification in the number of haloes and the power on scales smaller than 1 Mpc, due to the combination of baryonic and dark matter physics. Despite being calibrated on cold dark matter, we find that the TNG galaxy formation model can produce a realistic galaxy population in all scenarios. The stellar and gas mass fraction, stellar mass function, black hole mass as a function of stellar mass and star formation rate density are very similar in all dark matter models, with some deviations only in the most extreme warm dark matter model. Finally, we also quantify changes in halo structure due to warm and self-interacting dark matter, which appear in the density profiles, concentration-mass relation and galaxy sizes., Comment: 20 pages, 18 figures, submitted to A&A. Comments welcome, get in touch to know more about opportunities to use the runs - https://gdespali.github.io/AIDA/

Detecting the first generation of stars, Population III (PopIII), has been a long-standing goal in astrophysics, yet they remain elusive even in the JWST era. Here we present a novel NIRCam-based selection method for PopIII galaxies, and carefully validate it through completeness and contamination simulations. We systematically search ~500 arcmin$^{2}$ across JWST legacy fields for PopIII candidates, including GLIMPSE which, assisted by gravitational lensing, has produced JWST's deepest NIRCam imaging thus far. We discover one promising PopIII galaxy candidate (GLIMPSE-16043) at $z=6.50^{+0.03}_{-0.24}$, a moderately lensed galaxy (mu=2.9) with an intrinsic UV magnitude of $M_{UV}$=-15.89. It exhibits key PopIII features: strong H$\alpha$ emission (rest-frame EW $2810\pm550$\AA); a Balmer jump; no dust (UV slope $\beta=-2.34\pm0.36$); and undetectable metal lines (e.g., [OIII]; [OIII]/H$\beta$<0.44) implying a gas-phase metallicity of Zgas/Zsun<0.5%. These properties indicate the presence of a nascent, metal-deficient young stellar population (<5Myr) with a stellar mass of $\simeq10^{5}M_{\odot}$. Intriguingly, this source deviates significantly from the extrapolated UV-metallicity relation derived from recent JWST observations at $z=4-10$, consistent with UV enhancement by a top-heavy PopIII initial mass function or the presence of an extremely metal-poor AGN. We also derive the first observational constraints on the PopIII UV luminosity function at z~6-7. The volume density of GLIMPSE-16043 ($\approx10^{-4}$ cMpc$^{-3}$) is in excellent agreement with theoretical predictions, independently reinforcing its plausibility. This study demonstrates the power of our novel NIRCam method to finally reveal distant galaxies even more pristine than the Milky Way's most metal-poor satellites, thereby promising to bring us closer to the first generation of stars than we have ever been before., Comment: 41 pages, 20 figures, and 10 Tables. Submitted to ApJ. Comments are welcome!

Biharmonic and conformal biharmonic maps are two fourth-order generalizations of the well-studied notion of harmonic maps in Riemannian geometry. In this article we consider maps into the Euclidean sphere and investigate a geometric algorithm that aims at rendering a given harmonic map either biharmonic or conformally biharmonic. For biharmonic maps we find that in the case of a closed domain the maximum principle imposes strong restrictions on our approach, whereas there is more flexibility when we have a non-compact domain and we highlight this difference by a number of examples. Concerning conformal biharmonic maps we show that our algorithm produces explicit critical points for maps between spheres. Moreover, it turns out that we do not get strong restrictions as we obtain for biharmonic maps, such that our algorithm might produce additional conformal biharmonic maps between spheres beyond the ones found in this article.

The audio visual benefit in speech perception, where congruent visual input enhances auditory processing, is well documented across age groups, particularly in challenging listening conditions and among individuals with varying hearing abilities. However, most studies rely on highly controlled laboratory environments with scripted stimuli. Here, we examine the audio visual benefit using unscripted, natural speech from untrained speakers within a virtual acoustic environment. Using electroencephalography (EEG) and cortical speech tracking, we assessed neural responses across audio visual, audio only, visual only, and masked lip conditions to isolate the role of lip movements. Additionally, we analysed individual differences in acoustic and visual features of the speakers, including pitch, jitter, and lip openness, to explore their influence on the audio visual speech tracking benefit. Results showed a significant audio visual enhancement in speech tracking with background noise, with the masked lip condition performing similarly to the audio-only condition, emphasizing the importance of lip movements in adverse listening situations. Our findings reveal the feasibility of cortical speech tracking with naturalistic stimuli and underscore the impact of individual speaker characteristics on audio-visual integration in real world listening contexts.

This roadmap consolidates recent advances while exploring emerging applications, reflecting the remarkable diversity of hardware platforms, neuromorphic concepts, and implementation philosophies reported in the field. It emphasizes the critical role of cross-disciplinary collaboration in this rapidly evolving field.

Efficient preference optimization algorithms such as Direct Preference Optimization (DPO) have become a popular approach in aligning large language models (LLMs) with human preferences. These algorithms implicitly treat the LLM as a reward model, and focus on training it to correct misranked preference pairs. However, recent work~\citep{chen2024preference} empirically finds that DPO training \textit{rarely improves these misranked preference pairs}, despite its gradient emphasizing on these cases. We introduce FocalPO, a DPO variant that instead \textit{down-weighs} misranked preference pairs and prioritizes enhancing the model's understanding of pairs that it can already rank correctly. Inspired by Focal Loss used in vision tasks, FocalPO achieves this by adding a modulating factor to dynamically scale DPO loss. Our experiment demonstrates that FocalPO surpasses DPO and its variants on popular benchmarks like Alpaca Eval 2.0 using Mistral-Base-7B and Llama-3-Instruct-8B. Additionally, we empirically reveals how FocalPO affects training on correct and incorrect sample groups, further underscoring its effectiveness.

Autonomous robotic grasping of uncertain objects in uncertain environments is an impactful open challenge for the industries of the future. One such industry is the recycling of Waste Electrical and Electronic Equipment (WEEE) materials, in which electric devices are disassembled and readied for the recovery of raw materials. Since devices may contain hazardous materials and their disassembly involves heavy manual labor, robotic disassembly is a promising venue. However, since devices may be damaged, dirty and unidentified, robotic disassembly is challenging since object models are unavailable or cannot be relied upon. This case study explores grasping strategies for industrial robotic disassembly of WEEE devices with uncertain vision data. We propose three grippers and appropriate tactile strategies for force-based manipulation that improves grasping robustness. For each proposed gripper, we develop corresponding strategies that can perform effectively in different grasping tasks and leverage the grippers design and unique strengths. Through experiments conducted in lab and factory settings for four different WEEE devices, we demonstrate how object uncertainty may be overcome by tactile sensing and compliant techniques, significantly increasing grasping success rates.

Video Question Answering (Video QA) is a challenging video understanding task that requires models to comprehend entire videos, identify the most relevant information based on contextual cues from a given question, and reason accurately to provide answers. Recent advancements in Multimodal Large Language Models (MLLMs) have transformed video QA by leveraging their exceptional commonsense reasoning capabilities. This progress is largely driven by the effective alignment between visual data and the language space of MLLMs. However, for video QA, an additional space-time alignment poses a considerable challenge for extracting question-relevant information across frames. In this work, we investigate diverse temporal modeling techniques to integrate with MLLMs, aiming to achieve question-guided temporal modeling that leverages pre-trained visual and textual alignment in MLLMs. We propose T-Former, a novel temporal modeling method that creates a question-guided temporal bridge between frame-wise visual perception and the reasoning capabilities of LLMs. Our evaluation across multiple video QA benchmarks demonstrates that T-Former competes favorably with existing temporal modeling approaches and aligns with recent advancements in video QA., Comment: WACV 2025

Extensive gamma-ray spectroscopy of very neutron-rich nuclei of isotopes between the Ni and Sn isotopic chains was facilitated by the high luminosity LH2 target system, MINOS. Results show a persistence of deformation when going beyond the N = 60 threshold of the transition between spherical to deformed ground states at N < 60 and N > 60, respectively. Close to 78Ni, a more detailed image of the N > 50 Zn isotopes shows an erosion of the N = 50 shell closure, with core-breaking effects needed from theoretical models to replicate observation. As well as the experimental results indicating collective effects, the projected generator coordinate method is discussed in detail within the context of the neutron-rich Ge isotopes., Comment: Contribution to review article collection for the Symposium "Direct reactions and spectroscopy with hydrogen targets: past 10 years at the RIBF and future prospects", 18 pages, 7 figures, to be submitted to PTEP

Machine learning has become ubiquitous in materials modelling and now routinely enables large-scale atomistic simulations with quantum-mechanical accuracy. However, developing machine-learned interatomic potentials requires high-quality training data, and the manual generation and curation of such data can be a major bottleneck. Here, we introduce an automated framework for the exploration and fitting of potential-energy surfaces, implemented in an openly available software package that we call autoplex (`automatic potential-landscape explorer'). We discuss design choices, particularly the interoperability with existing software architectures, and the ability for the end user to easily use the computational workflows provided. We show wide-ranging capability demonstrations: for the titanium-oxygen system, SiO2, crystalline and liquid water, as well as phase-change memory materials. More generally, our study illustrates how automation can speed up atomistic machine learning -- with a long-term vision of making it a genuine mainstream tool in physics, chemistry, and materials science.

The relationship between different dissipativity concepts for linear time-varying systems is studied, in particular between port-Hamiltonian systems, passive systems, and systems with nonnegative supply. It is shown that linear time-varying port-Hamiltonian systems are passive, have nonnegative supply rates, and solve (under different smoothness assumptions) Kalman-Yakubovich-Popov differential and integral inequalities. The converse relations are also studied in detail. In particular, sufficient conditions are presented to obtain a port-Hamiltonian representation starting from any of the other dissipativity concepts. Two applications are presented.

Nuclear magnetic resonance (NMR) is a powerful spectroscopic technique that is sensitive to the local atomic structure of matter. Computational predictions of NMR parameters can help to interpret experimental data and validate structural models, and machine learning (ML) has emerged as an efficient route to making such predictions. Here, we systematically study graph-neural-network approaches to representing and learning tensor quantities for solid-state NMR -- specifically, the anisotropic magnetic shielding and the electric field gradient. We assess how the numerical accuracy of different ML models translates into prediction quality for experimentally relevant NMR properties: chemical shifts, quadrupolar coupling constants, tensor orientations, and even static 1D spectra. We apply these ML models to a structurally diverse dataset of amorphous SiO$_2$ configurations, spanning a wide range of density and local order, to larger configurations beyond the reach of traditional first-principles methods, and to the dynamics of the $\alpha\unicode{x2013}\beta$ inversion in cristobalite. Our work marks a step toward streamlining ML-driven NMR predictions for both static and dynamic behavior of complex materials, and toward bridging the gap between first-principles modeling and real-world experimental data., Comment: 13 pages, 7 figures

Modern challenges of robustness, fairness, and decision-making in machine learning have led to the formulation of multi-distribution learning (MDL) frameworks in which a predictor is optimized across multiple distributions. We study the calibration properties of MDL to better understand how the predictor performs uniformly across the multiple distributions. Through classical results on decomposing proper scoring losses, we first derive the Bayes optimal rule for MDL, demonstrating that it maximizes the generalized entropy of the associated loss function. Our analysis reveals that while this approach ensures minimal worst-case loss, it can lead to non-uniform calibration errors across the multiple distributions and there is an inherent calibration-refinement trade-off, even at Bayes optimality. Our results highlight a critical limitation: despite the promise of MDL, one must use caution when designing predictors tailored to multiple distributions so as to minimize disparity.

Due to the increase of data volumes expected for the LHC Run 3 and Run 4, the ALICE Collaboration designed and deployed a new, energy efficient, computing model to run Online and Offline O$^2$ data processing within a single software framework. The ALICE O$^2$ Event Processing Nodes (EPN) project performs online data reconstruction using GPUs (Graphic Processing Units) instead of CPUs and applies an efficient, entropy-based, online data compression to cope with PbPb collision data at a 50 kHz hadronic interaction rate. Also, the O$^2$ EPN farm infrastructure features an energy efficient, environmentally friendly, adiabatic cooling system which allows for operational and capital cost savings.

Multimodal Large Language Models (MLLMs) have significantly advanced visual tasks by integrating visual representations into large language models (LLMs). The textual modality, inherited from LLMs, equips MLLMs with abilities like instruction following and in-context learning. In contrast, the visual modality enhances performance in downstream tasks by leveraging rich semantic content, spatial information, and grounding capabilities. These intrinsic modalities work synergistically across various visual tasks. Our research initially reveals a persistent imbalance between these modalities, with text often dominating output generation during visual instruction tuning. This imbalance occurs when using both full fine-tuning and parameter-efficient fine-tuning (PEFT) methods. We then found that re-balancing these modalities can significantly reduce the number of trainable parameters required, inspiring a direction for further optimizing visual instruction tuning. We introduce Modality Linear Representation-Steering (MoReS) to achieve the goal. MoReS effectively re-balances the intrinsic modalities throughout the model, where the key idea is to steer visual representations through linear transformations in the visual subspace across each model layer. To validate our solution, we composed LLaVA Steering, a suite of models integrated with the proposed MoReS method. Evaluation results show that the composed LLaVA Steering models require, on average, 500 times fewer trainable parameters than LoRA needs while still achieving comparable performance across three visual benchmarks and eight visual question-answering tasks. Last, we present the LLaVA Steering Factory, an in-house developed platform that enables researchers to quickly customize various MLLMs with component-based architecture for seamlessly integrating state-of-the-art models, and evaluate their intrinsic modality imbalance.

We analyze four super-Earth exoplanets, LHS 1140 b, K2-18 b, TOI-1452 b, and TOI-1468 c, which orbit M-dwarf stars in the habitable zone. Their relative proximity, within 40 parsecs, makes them prime candidates for follow-up observations and atmospheric and habitability studies. This paper aims to assess their internal structure and habitability, considering their tidal heating, atmospheric heating, and global transport. We model the interior structure of the planets by applying Bayesian inference to an exoplanet's interior model. A constant quality factor model is used to calculate the range of tidal heating, and a one-dimensional analytical model of tidally locked planets is used to assess their surface temperature distribution and habitability. Assuming no or only thin atmospheres, K2-18 b and TOI-1468 c are likely to be water worlds. However, TOI-1452 b and LHS 1140 b may have rocky surfaces. We find that tidal heating is not enough to raise the global mean surface temperature, but greenhouse heating can effectively do so. If the considered planets have retained thick atmospheres, K2-18 b, TOI-1468 c, and TOI-1452 b may, for significant atmospheric heating and heat transport factors, be too hot to sustain liquid water on their surface. However, the lower instellation of LHS 1140 b and the non-zero probability of it having a rocky surface give more space for habitable conditions on the planet., Comment: Accepted to ApJ

Modern robotic perception is highly dependent on neural networks. It is well known that neural network-based perception can be unreliable in real-world deployment, especially in difficult imaging conditions. Out-of-distribution detection is commonly proposed as a solution for ensuring reliability in real-world deployment. Previous work has shown that normalizing flow models can be used for out-of-distribution detection to improve reliability of robotic perception tasks. Specifically, camera parameters can be optimized with respect to the likelihood output from a normalizing flow, which allows a perception system to adapt to difficult vision scenarios. With this work we propose to use the absolute gradient values from a normalizing flow, which allows the perception system to optimize local regions rather than the whole image. By setting up a table top picking experiment with exceptionally difficult lighting conditions, we show that our method achieves a 60% higher success rate for an object detection task compared to previous methods.