Search

Omega Centauri ($\omega$ Cen) is the Milky Way's most massive globular cluster and is likely the stripped nucleus of an accreted dwarf galaxy. In this paper, we analyze $\omega$ Cen's kinematics using data from oMEGACat, a comprehensive catalog of $\omega$ Cen's central regions, including 1.4 million proper motion measurements and 300,000 spectroscopic radial velocities. Our velocity dispersion profiles and kinematic maps are consistent with previous work but improve on their resolution, precision, and spatial coverage. The cluster's 3D dispersion is isotropic in the core, with increasing radial anisotropy at larger radii. The 2D kinematic maps show an elongation of the velocity dispersion field comparable to the flattening observed photometrically. We find good agreement between proper motions and line-of-sight velocity dispersion and measure a kinematic distance of 5494$\pm$61 pc, the most precise kinematic distance to $\omega$ Cen available. The subset of data with precise metallicity measurements shows no correlation between metallicity and kinematics, supporting the picture of well-mixed stellar populations within the half-light radius of $\omega$ Cen. Finally, we study the degree of energy equipartition using a large range of stellar masses. We find partial energy equipartition in the center that decreases towards large radii. The spatial dependence of the radial energy equipartition is stronger than the tangential energy equipartition. Our kinematic observations can serve as a new reference for future dynamical modeling efforts that will help to further disentangle the complex mass distribution within $\omega$ Cen., Comment: 31 pages, 23 Figures, 4 Tables. Accepted for publication in ApJ. Data products available under: https://doi.org/10.5281/zenodo.14978551

Continual learning (CL) is the sub-field of machine learning concerned with accumulating knowledge in dynamic environments. So far, CL research has mainly focused on incremental classification tasks, where models learn to classify new categories while retaining knowledge of previously learned ones. Here, we argue that maintaining such a focus limits both theoretical development and practical applicability of CL methods. Through a detailed analysis of concrete examples - including multi-target classification, robotics with constrained output spaces, learning in continuous task domains, and higher-level concept memorization - we demonstrate how current CL approaches often fail when applied beyond standard classification. We identify three fundamental challenges: (C1) the nature of continuity in learning problems, (C2) the choice of appropriate spaces and metrics for measuring similarity, and (C3) the role of learning objectives beyond classification. For each challenge, we provide specific recommendations to help move the field forward, including formalizing temporal dynamics through distribution processes, developing principled approaches for continuous task spaces, and incorporating density estimation and generative objectives. In so doing, this position paper aims to broaden the scope of CL research while strengthening its theoretical foundations, making it more applicable to real-world problems.

One of the most popular methods for continual learning with deep neural networks is Elastic Weight Consolidation (EWC), which involves computing the Fisher Information. The exact way in which the Fisher Information is computed is however rarely described, and multiple different implementations for it can be found online. This blog post discusses and empirically compares several often-used implementations, which highlights that many currently reported results for EWC could likely be improved by changing the way the Fisher Information is computed., Comment: To appear in the blogpost track at ICLR 2025

This study explores the potential of graph neural networks (GNNs) to enhance semantic segmentation across diverse image modalities. We evaluate the effectiveness of a novel GNN-based U-Net architecture on three distinct datasets: PascalVOC, a standard benchmark for natural image segmentation, WoodScape, a challenging dataset of fisheye images commonly used in autonomous driving, introducing significant geometric distortions; and ISIC2016, a dataset of dermoscopic images for skin lesion segmentation. We compare our proposed UNet-GNN model against established convolutional neural networks (CNNs) based segmentation models, including U-Net and U-Net++, as well as the transformer-based SwinUNet. Unlike these methods, which primarily rely on local convolutional operations or global self-attention, GNNs explicitly model relationships between image regions by constructing and operating on a graph representation of the image features. This approach allows the model to capture long-range dependencies and complex spatial relationships, which we hypothesize will be particularly beneficial for handling geometric distortions present in fisheye imagery and capturing intricate boundaries in medical images. Our analysis demonstrates the versatility of GNNs in addressing diverse segmentation challenges and highlights their potential to improve segmentation accuracy in various applications, including autonomous driving and medical image analysis.

We report, for the first time, the detection of a sample of quenched and isolated dwarf galaxies (with 8.9 $<$ log(M$_{\rm \star}$/M$_{\rm \odot}$) $<$ 9.5) in the least dense regions of the cosmic web, including voids, filaments, and walls. These dwarfs have no neighbouring galaxy within 1.0~Mpc in projected distance. Based on the full spectral fitting of their central spectra using Sloan Digital Sky Survey data, these galaxies are gas-deprived, exhibit stellar mass assembly very similar to dwarfs in the central regions of galaxy clusters, and have experienced no significant star formation in the past 2 Gyr. Additionally, analysis of r-band images from the Dark Energy Camera Legacy Survey showed that these dwarf galaxies host a central Nuclear Star Cluster (NSC). Detecting quenched, isolated dwarf galaxies in cosmic voids indicates that environmental factors are not the sole drivers of their quenching. Internal mechanisms, such as feedback from in-situ star formation, also contributing to the NSC formation, black holes, or variations in conditions during their formation, offer potential explanations for star formation suppression in these galaxies. These findings highlight the need for a significant revision in our understanding of baryonic physics, particularly concerning the formation and evolution of low-mass galaxies., Comment: 9 pages, 5 figures, accepted as a Letter to the Editor in Astronomy & Astrophysics

Constraining the helium enhancement in stars is critical for understanding the formation mechanisms of multiple populations in star clusters. However, measuring helium variations for many stars within a cluster remains observationally challenging. We use Hubble Space Telescope photometry combined with MUSE spectroscopic data for over 7,200 red-giant branch stars in \omc\ to measure helium differences between distinct groups of stars as a function of metallicity separating the impact of helium enhancements from other abundance variations on the pseudo-color (chromosome) diagrams. Our results show that stars at all metallicities have subpopulations with significant helium enhancement ($\Delta Y_{min} \gtrsim$ 0.11). We find a rapid increase in helium enhancement from low metallicities ($\rm{[Fe/H] \simeq -2.05}$ to $\rm{[Fe/H] \simeq -1.92})$, with this enhancement leveling out at \deltay\ $= 0.154$ at higher metallicities. The fraction of helium-enhanced stars steadily increases with metallicity ranging from 10\% at $\rm{[Fe/H] \simeq -2.04}$ to over $90\%$ at $\rm{[Fe/H] \simeq -1.04}$. This study is the first to examine helium enhancement across the full range of metallicities in \omc{}, providing new insight into its formation history and additional constraints on enrichment mechanisms.

The innermost parts of the Milky Way (MW) are very difficult to observe due to the high extinction along the line of sight, especially close to the disc mid-plane. However, this region contains the most massive complex stellar component of the MW, the bulge, primarily composed of disc stars whose structure is (re-)shaped by the evolution of the bar. In this work, we extend the application of the orbit superposition method to explore the present-day 3D structure, orbital composition, chemical abundance trends and kinematics of the MW bulge. Thanks to our approach, we are able to transfer astrometry from Gaia and stellar parameters from APOGEE DR 17 to map the inner MW without obscuration by the survey footprint and selection function. We demonstrate that the MW bulge is made of two main populations originating from a metal-poor, high-{\alpha} thick disc and a metal-rich, low-{\alpha} thin disc, with a mass ratio of 4:3, seen as two major components in the MDF. Finer MDF structures hint at multiple sub-populations associated with different orbital families of the bulge, which, however, have broad MDFs themselves. Decomposition using 2D GMMs in [Fe/H] -[Mg/Fe] identifies five components including a population with ex-situ origin. Two dominant ones correspond to the thin and thick discs and two in between trace the transition between them. We show that no universal metallicity gradient value can characterise the MW bulge. The radial gradients closely trace the X-shaped bulge density structure, while the vertical gradient variations follow the boxy component. While having, on average, subsolar metallicity, the MW bulge populations are more metal-rich compared to the surrounding disc, in agreement with extragalactic observations and state-of-the-art simulations reinforcing its secular origin., Comment: 24 pages, 20 figures; submitted to A&A

The stellar disc is the dominant luminous component of the Milky Way (MW). Although our understanding of its structure is rapidly expanding due to advances in large-scale stellar surveys, our picture of the MW disc remains substantially obscured by selection functions and incomplete spatial coverage of observational data. In this work, we present the comprehensive chrono-chemo-kinematic structure of the MW disc, recovered using a novel orbit superposition approach combined with data from APOGEE DR 17. We detect periodic azimuthal metallicity variations within 6-8 kpc with an amplitude of 0.05-0.1 dex peaking along the bar major axis. The radial metallicity profile of the MW also varies with azimuth, displaying a pattern typical among other disc galaxies: a decline outside the solar radius and an almost flat profile in the inner region, attributed to the presence of old, metal-poor high-{\alpha} populations, which comprise about 40% of the total stellar mass. The geometrically defined thick disc and the high-{\alpha} populations have comparable masses, with differences in their stellar population content, which we quantify using the reconstructed 3D MW structure. The well-known [{\alpha}/Fe]-bimodality in the MW disc, once weighted by stellar mass, is less pronounced at a given metallicity for the whole galaxy but distinctly visible in a narrow range of galactic radii (5-9 kpc), explaining its relative lack of prominence in external galaxies and galaxy formation simulations. Analysing a more evident double age-abundance sequence, we construct a scenario for the MW disc formation, advocating for an inner/outer disc dichotomy genetically linked to the MW's evolutionary stages. In this picture, the extended solar vicinity is a transition zone that shares chemical properties of both the inner (old age-metallicity sequence) and outer discs (young age-metallicity sequence)., Comment: 36 pages, 34 figures; submitted to A&A

The processes driving the formation and evolution of late-type galaxies (LTGs) continue to be a debated subject in extragalactic astronomy. Investigating stellar kinematics, especially when combined with age estimates, provides crucial insights into the formation and subsequent development of galactic discs. Post-processing of exceptionally high-quality Integral Field Spectroscopy (IFS) data of NGC 4030 acquired with the Multi Unit Spectroscopic Explorer (MUSE), clearly reveals a striking grand design spiral pattern in the velocity dispersion map not previously detected in other galaxies. This pattern spatially correlates with HII regions, suggesting that stars currently being born exhibit lower velocity dispersion as compared to surrounding areas where star formation (SF) is less active. We examine the age-velocity relation (AVR) and propose that its configuration might be shaped by a combination of heating mechanisms, seemingly consistent with findings from recent high-resolution cosmological zoom-in simulations. The complex structure of the uncovered AVR of NGC 4030 support the hypothesis that stellar populations initially inherit the velocity dispersion {\sigma} of the progenitor cold molecular gas, which depends on formation time and galactocentric distance, subsequently experiencing kinematic heating by cumulative gravitational interactions during their lifetime. While advancing our understanding of the AVR, these findings offer a new framework for investigating disk heating mechanisms, and their role in the evolution of galactic disks., Comment: accepted for publication in A&A letters

Recent advances in AI -- including generative approaches -- have resulted in technology that can support humans in scientific discovery and decision support but may also disrupt democracies and target individuals. The responsible use of AI increasingly shows the need for human-AI teaming, necessitating effective interaction between humans and machines. A crucial yet often overlooked aspect of these interactions is the different ways in which humans and machines generalise. In cognitive science, human generalisation commonly involves abstraction and concept learning. In contrast, AI generalisation encompasses out-of-domain generalisation in machine learning, rule-based reasoning in symbolic AI, and abstraction in neuro-symbolic AI. In this perspective paper, we combine insights from AI and cognitive science to identify key commonalities and differences across three dimensions: notions of generalisation, methods for generalisation, and evaluation of generalisation. We map the different conceptualisations of generalisation in AI and cognitive science along these three dimensions and consider their role in human-AI teaming. This results in interdisciplinary challenges across AI and cognitive science that must be tackled to provide a foundation for effective and cognitively supported alignment in human-AI teaming scenarios.

We introduce a novel orbit superposition method designed to reconstruct the stellar density structure, kinematics, and chemical abundance distribution of the entire Milky Way by leveraging 6D phase-space information from its resolved stellar populations, limited by the spatial coverage of APOGEE DR17., Comment: 17 pages, 13 figures, submitted to A&A

Dark matter is one of the pillars of the current standard model of structure formation: it is assumed to constitute most of the matter in the Universe. However, it can so far only be probed indirectly through its gravitational effects, and its nature remains elusive. In this focus meeting, we discussed different methods used to estimate galaxies' visible and dark matter masses in the nearby and distant Universe. We reviewed successes of the standard model relying on cold dark matter, confronted observations with simulations, and highlighted inconsistencies between the two. We discussed how robust mass measurements can help plan, perform, and refine particle dark matter searches. We further exchanged about alternatives to cold dark matter, such as warm, self-interacting, and fuzzy dark matter, as well as modified gravity. Finally, we discussed prospects and strategies that could be implemented to reveal the nature of this crucial component of the Universe., Comment: Rationale and highlights of the Focus Meeting 9 "Measures of luminous and dark matter in galaxies across time" at the XXXIInd IAU General Assembly in Cape Town, August 2024. 12 pages

NGC 5139 ($\omega$ Cen), is the closest candidate of a Nuclear Star Cluster that has been stripped of its host galaxy in the Milky Way. Despite extensive studies through the last decades, many open questions about the cluster remain, including the properties of the binary population. In this study we use MUSE multi-epoch spectroscopy to identify binary systems in $\omega$ Cen. The observations span 8 years, with a total of 312 248 radial velocity measurements for 37 225 stars. Following the removal of known photometric variables, we identify 275 stars that show RV variations, corresponding to a discovery fraction of 1.4$\pm$0.1%. Using dedicated simulations, we find that our data is sensitive to 70$\pm$10% of the binaries expected in the sample, resulting in a completeness-corrected binary fraction of 2.1$\pm$0.4% in the central region of $\omega$ Cen. We find similar binary fractions for all stellar evolutionary stages covered by our data, the only notable exception being the blue straggler stars, which show an enhanced binary fraction. We also find no distinct correlation with distance from the cluster centre, indicating a limited amount of mass segregation within the half-light radius of $\omega$ Cen., Comment: 13 pages, 9 figures, 1 Table; Published by Monthly Notices of the Royal Astronomical Society (MNRAS)

The vertical evolution of galactic discs is governed by the sub-structures within them. We examine the diversity of kinematic sub-structure present in the first 12 galaxies observed from the GECKOS survey, a VLT/MUSE large programme providing a systematic study of 35 edge-on, Milky Way-mass disc galaxies. Employing the nGIST analysis pipeline, we derive the mean line-of-sight stellar velocity ($V_{\star}$), velocity dispersion ($\sigma_{\star}$), skew ($h_{3}$), and kurtosis ($h_{4}$) for the sample, and examine 2D maps and 1D line profiles. Visually, the majority of this sample (8/12) are found to possess boxy-peanut bulges and host the corresponding kinematic structure predicted for stellar bars viewed in projection. Four galaxies exhibit strong evidence for the presence of nuclear discs, including central $h_{3}$-$V_{\star}$ anti-correlations, `croissant'-shaped central depressions in $\sigma_{\star}$ maps, strong gradients in $h_{3}$, and positive $h_{4}$ plateaus over the expected nuclear disc extent. The strength of the $h_{3}$ feature corresponds to the size of the nuclear disc, measured from the $h_{3}$ turnover radius. We can explain the features within the kinematic maps of all sample galaxies via disc structure(s) alone. We do not find any need to invoke the existence of dispersion-dominated bulges. Obtaining the specialised data products for this paper and the broader GECKOS survey required significant development of existing integral field spectroscopic (IFS) analysis tools. Therefore, we also present the nGIST pipeline: a modern, sophisticated, and easy-to-use pipeline for the analysis of galaxy IFS data. We conclude that the variety of kinematic sub-structures seen in GECKOS galaxies requires a contemporary view of galaxy morphology, expanding on the traditional view of galaxy structure, and uniting the kinematic complexity observed in the Milky Way with the extragalactic., Comment: 32 pages (9 of which are appendix), 26 figures, submitted to A&A. Comments welcome!

Continuous fields (or bundles) of $C^*$-algebras form an important ingredient for describing emergent phenomena, such as phase transitions and spontaneous symmetry breaking. In this work, we consider the continuous $C^*$-bundle generated by increasing symmetric tensor powers of the complex $\ell\times\ell$ matrices $M_\ell(\mathbb{C})$, which can be interpreted as abstract description of mean-field theories defining the macroscopic limit of infinite quantum systems. Within this framework we discuss the principle of large deviations for the local Gibbs state in the high temperature regime and characterize the limit of the ensuing logarithmic generating function., Comment: 24 pages

We use Gaia DR3 to explore how well equilibrium dynamical models based on the Jeans equations and the Schwarzschild orbit superposition method are able to describe LMC's 5-dimensional phase-space distribution and line-of-sight (LOS) velocity distribution, respectively. In the latter model we incorporate a triaxial bar component and derive LMC's bar pattern speed. We fit Jeans dynamical models to all Gaia DR3 stars with proper motion and LOS velocity measurements found in the VMC VISTA survey of the LMC using a discrete maximum likelihood approach. These models are very efficient at discriminating genuine LMC member stars from Milky Way foreground stars and background galaxies. They constrain the shape, orientation, and enclosed mass of the galaxy under the assumption for axisymmetry. We use the Jeans model results as a stepping stone to more complex 2-component Schwarzschild models, which include an axisymmetric disc and a co-centric triaxial bar, which we fit to the LMC Gaia DR3 LOS velocity field, using a chi^2 minimisation approach. The Jeans models describe well the rotation and velocity dispersion of the LMC disc and we find an inclination angle 25.5 deg, line of nodes orientation 124 deg, and an intrinsic thickness of the disc 0.23 (minor to major axis ratio). However, bound to axisymmetry, these models fail to properly describe the kinematics in the central region of the galaxy, dominated by the bar. We use the derived disc orientation and the Gaia DR3 density image of the LMC to obtain the intrinsic shape of the bar. Using these two components as an input to our Schwarzschild models, we perform orbit integration and weighting in a rotating reference frame fixed to the bar, deriving an independent measurement of the LMC bar pattern speed 11+/-4 km/s/kpc. Both the Jeans and Schwarzschild models predict the same enclosed mass distribution within a radius of 6.2 kpc of ~1.4e10 Msun., Comment: Accepted for publication in A&A

Learning from Demonstration offers great potential for robots to learn to perform agricultural tasks, specifically selective harvesting. One of the challenges is that the target fruit can be oscillating while approaching. Grasping oscillating targets has two requirements: 1) close tracking of the target during the final approach for damage-free grasping, and 2) the complete path should be as short as possible for improved efficiency. We propose a new method called DualLQR. In this method, we use a finite horizon Linear Quadratic Regulator (LQR) on a moving target, without the need of refitting the LQR. To make this possible, we use a dual LQR set-up, with an LQR running in two separate reference frames. Through extensive simulation testing, it was found that the state-of-art method barely meets the required final accuracy without oscillations and drops below the required accuracy with an oscillating target. DualLQR, on the other hand, was found to be able to meet the required final accuracy even with high oscillations, while travelling the least distance. Further testing on a real-world apple grasping task showed that DualLQR was able to successfully grasp oscillating apples, with a success rate of 99%., Comment: Submitted to IAS-19

We present age estimates for over 8100 sub-giant branch (SGB) stars in Omega Centauri ($\omega$ Cen) to study its star formation history. Our large data set, which combines multi-wavelength HST photometry with MUSE metallicities, provides an unprecedented opportunity to measure individual stellar ages. We do this by fitting each star's photometry and metallicity with theoretical isochrones, that are embedded with an empirical [C+N+O]-[Fe/H] relation specifically for $\omega$ Cen. The bulk of the stars have ages between 13 and 10 Gyr, with the mean stellar age being 12.08$\pm$0.01 Gyrs and the median age uncertainty being 0.68 Gyrs. From these ages we construct the most complete age-metallicity relation (AMR) for $\omega$ Cen to date. We find that the mean age of stars decreases with increasing metallicity and find two distinct streams in the age-metallicity plane, hinting at different star formation pathways. We derive an intrinsic spread in the ages of 0.75$\pm$0.01 Gyr for the whole cluster, with the age spread showing a clear increase with metallicity. We verify the robustness of our age estimations by varying isochrone parameters and constraining our systematics. We find the C+N+O relation to be the most critical consideration for constraining the AMR. We also present the SGB chromosome map with age information. In the future, these stellar ages could be combined with chemical abundances to study age differences in subpopulations, and uncover the chemical evolution history of this massive nuclear star cluster., Comment: 23 pages, 11 figures

Schwarzschild dynamical models are now regularly employed in large surveys of galaxies in the local and distant Universe to derive information on galaxies' intrinsic properties such as their orbital structure and their (dark matter and stellar) mass distribution. Comparing the internal orbital structures and mass distributions of galaxies in the distant Universe with simulations is key to understanding what physical processes are responsible for shaping galaxy properties. However it is first crucial to understand whether observationally derived properties are directly comparable with intrinsic ones in simulations. To assess this, we build Schwarzschild dynamical models for MUSE-like IFS cubes (constructed to be like those obtained by the MAGPI survey) of 75 galaxies at z ~ 0.3 from the Eagle simulations. We compare the true particle-derived properties with the galaxies' model-derived properties. In general, we find that the models can recover the true galaxy properties qualitatively well, with the exception of the enclosed dark matter, where we find a median offset of 48%, which is due to the assumed NFW profile not being able to reproduce the dark matter distribution in the inner region of the galaxies. We then compare our model-derived properties with Schwarzschild models-derived properties of observed MAGPI galaxies and find good agreement between MAGPI and Eagle: the majority of our galaxies (57%) have non-oblate shapes within 1 effective radius. More triaxial galaxies show higher fractions of hot orbits in their inner regions and tend to be more radially anisotropic., Comment: 22 pages. Accepted for publication in MNRAS

In this study, we present the Graph Sub-Graph Network (GSN), a novel hybrid image classification model merging the strengths of Convolutional Neural Networks (CNNs) for feature extraction and Graph Neural Networks (GNNs) for structural modeling. GSN employs k-means clustering to group graph nodes into clusters, facilitating the creation of subgraphs. These subgraphs are then utilized to learn representative `atoms` for dictionary learning, enabling the identification of sparse, class-distinguishable features. This integrated approach is particularly relevant in domains like medical imaging, where discerning subtle feature differences is crucial for accurate classification. To evaluate the performance of our proposed GSN, we conducted experiments on benchmark datasets, including PascalVOC and HAM10000. Our results demonstrate the efficacy of our model in optimizing dictionary configurations across varied classes, which contributes to its effectiveness in medical classification tasks. This performance enhancement is primarily attributed to the integration of CNNs, GNNs, and graph learning techniques, which collectively improve the handling of datasets with limited labeled examples. Specifically, our experiments show that the model achieves a higher accuracy on benchmark datasets such as Pascal VOC and HAM10000 compared to conventional CNN approaches.

Star formation rates (SFRs), gas-phase metallicities, and stellar masses are crucial for studying galaxy evolution. The different relations resulting from these properties give insights into the complex interplay of gas inside galaxies and their evolutionary trajectory and current characteristics. We aim to characterize these relations at $z\sim 0.3$, corresponding to a 3-4 Gyr lookback time. We utilized optical integral field spectroscopy of 65 emission-line galaxies from the MAGPI survey at a redshift of $0.28

We study the classical and quantum KMS conditions within the context of spin lattice systems. Specifically, we define a strict deformation quantization (SDQ) for a $\mathbb{S}^2$-valued spin lattice system over $\mathbb{Z}^d$ generalizing the renown Berezin SDQ for a single sphere. This allows to promote a classical dynamics on the algebra of classical observables to a quantum dynamics on the algebra of quantum observables. We then compare the notion of classical and quantum thermal equilibrium by showing that any weak*-limit point of a sequence of quantum KMS states fulfils the classical KMS condition. In short, this proves that the semiclassical limit of quantum thermal states describes classical thermal equilibrium, strenghtening the physical interpretation of the classical KMS condition. Finally we provide two sufficient conditions ensuring uniqueness of classical and quantum KMS states: The latter are based on an version of the Kirkwood-Salzburg equations adapted to the system of interest. As a consequence we identify a mild condition which ensures uniqueness of classical KMS states and of quantum KMS states for the quantized dynamics for a common sufficiently high temperature., Comment: 45 pages

This work demonstrates that training autoregressive video diffusion models from a single, continuous video stream is not only possible but remarkably can also be competitive with standard offline training approaches given the same number of gradient steps. Our demonstration further reveals that this main result can be achieved using experience replay that only retains a subset of the preceding video stream. We also contribute three new single video generative modeling datasets suitable for evaluating lifelong video model learning: Lifelong Bouncing Balls, Lifelong 3D Maze, and Lifelong PLAICraft. Each dataset contains over a million consecutive frames from a synthetic environment of increasing complexity.

$\omega$ Centauri, the most massive globular cluster in the Milky Way, has long been suspected to be the stripped nucleus of a dwarf galaxy that fell into the Galaxy a long time ago. There is considerable evidence for this scenario including a large spread in metallicity and an unusually large number of distinct sub-populations seen in photometric studies. In this work, we use new MUSE spectroscopic and HST photometric catalogs to investigate the underlying metallicity distributions as well as the spatial variations of the populations within the cluster up to its half-light radius. Based on 11,050 member stars, the [M/H] distribution has a median of $ (-1.614 \pm 0.003)$ dex and a large spread of $\sim$ 1.37 dex reaching from $ -0.67$ dex to $ -2.04$ dex for 99.7 % of the stars. In addition, we show the chromosome map of the cluster, which separates the red giant branch stars into different sub-populations, and analyze the sub-populations of the metal-poorest component. Finally, we do not find any metallicity gradient within the half-light radius, and the different sub-populations are well mixed., Comment: 22 pages, 18 figures, and 3 tables. Accepted for publication in ApJ

Black holes have been found over a wide range of masses, from stellar remnants with masses of 5-150 solar masses (Msun), to those found at the centers of galaxies with $M>10^5$ Msun. However, only a few debated candidate black holes exist between 150 and $10^5$ Msun. Determining the population of these intermediate-mass black holes is an important step towards understanding supermassive black hole formation in the early universe. Several studies have claimed the detection of a central black hole in $\omega$ Centauri, the Milky Way's most massive globular cluster. However, these studies have been questioned due to the possible mass contribution of stellar mass black holes, their sensitivity to the cluster center, and the lack of fast-moving stars above the escape velocity. Here we report observations of seven fast-moving stars in the central 3 arcseconds (0.08 pc) of $\omega$ Centauri. The velocities of the fast-moving stars are significantly higher than the expected central escape velocity of the star cluster, so their presence can only be explained by being bound to a massive black hole. From the velocities alone, we can infer a firm lower limit of the black hole mass of $\sim$8,200 Msun, making this a compelling candidate for an intermediate-mass black hole in the local universe., Comment: 33 pages, 11 figures, and 2 tables. Published in Nature. This is the accepted author's version including the correction issued on Sep 17, 2024. The version of record is available from the Journal (open access)

Continual learning (CL) provides a framework for training models in ever-evolving environments. Although re-occurrence of previously seen objects or tasks is common in real-world problems, the concept of repetition in the data stream is not often considered in standard benchmarks for CL. Unlike with the rehearsal mechanism in buffer-based strategies, where sample repetition is controlled by the strategy, repetition in the data stream naturally stems from the environment. This report provides a summary of the CLVision challenge at CVPR 2023, which focused on the topic of repetition in class-incremental learning. The report initially outlines the challenge objective and then describes three solutions proposed by finalist teams that aim to effectively exploit the repetition in the stream to learn continually. The experimental results from the challenge highlight the effectiveness of ensemble-based solutions that employ multiple versions of similar modules, each trained on different but overlapping subsets of classes. This report underscores the transformative potential of taking a different perspective in CL by employing repetition in the data stream to foster innovative strategy design., Comment: Accepted version, to appear in Neural Networks; Challenge Report of the 4th Workshop on Continual Learning in Computer Vision at CVPR

Bars drive gas inflow. As the gas flows inwards, shocks and shear occur along the bar dust lanes. Such shocks and shear can affect the star formation and change the gas properties. For four barred galaxies, we present H{\alpha} velocity gradient maps that highlight bar-driven shocks and shear using data from the PHANGS-MUSE and PHANGS-ALMA surveys which allow us to study bar kinematics in unprecedented detail. Velocity gradients are enhanced along the bar dust lanes, where shocks and shear are shown to occur in numerical simulations. Velocity gradient maps also efficiently pick up expanding shells around HII regions. We put pseudo slits on the regions where velocity gradients are enhanced and find that H{\alpha} and CO velocities jump up to ~170 km/s, even after removing the effects of circular motions due to the galaxy rotation. Enhanced velocity gradients either coincide with the peak of CO intensity along the bar dust lanes or are slightly offset from CO intensity peaks, depending on the objects. Using the BPT diagnostic, we identify the source of ionization on each spaxel and find that star formation is inhibited in the high velocity gradient regions of the bar, and the majority of those regions are classified as LINER or composite. This implies that star formation is inhibited where bar-driven shear and shocks are strong. Our results are consistent with the results from the numerical simulations that show star formation is inhibited in the bar where shear force is strong., Comment: 26 pages, Accepted for publication in ApJ

Omega Centauri ($\omega$ Cen) is the most massive globular cluster of the Milky Way. It is thought to be the nucleus of an accreted dwarf galaxy because of its high mass and its complex stellar populations. To decipher its formation history and study its dynamics, we created the most comprehensive kinematic catalog for its inner region, by analyzing both archival and new Hubble Space Telescope (HST) data. Our catalog contains 1 395 781 proper-motion measurements out to the half-light radius of the cluster ($\sim$5.0') and down to $m_{F625W}\approx$25. The typical baseline for our proper-motion measurements is 20 years, leading to a median 1D proper motion precision of $\sim$11 $\mu$as yr$^{-1}$ for stars with $m_{F625W}\approx$18 mag, with even better precision ($\sim$6.6 $\mu$as yr$^{-1}$) achieved in the extensively observed centermost (r$<$1.5') region. In addition to our astrometric measurements, we also obtained precise HST photometry in seven filters spanning from the ultraviolet to the near-infrared. This allows detailed color-magnitude-diagram studies and to separate the multiple stellar populations of the cluster. In this work, we describe the data reduction used to obtain both the photometric and the proper-motion measurements. We also illustrate the creation and the content of our catalog, which is made publicly available. Finally, we present measurements of the plane-of-sky rotation of $\omega$ Cen in the previously unprobed inner few arcminutes and a precise measurement of the inclination $i = (43.9\pm1.3)^\circ$., Comment: 43 pages, 25 figures, 9 tables. Published by ApJ. The full catalog is publicly available under: https://doi.org/10.5281/zenodo.11104046

We studied the effects of cluster environments on galactic structures by using the TNG50 cosmological simulation and observed galaxies in the Fornax cluster. We focused on galaxies with stellar masses of $10^{8-12}M_{\odot}$ at z=0 that reside in Fornax-like clusters with total masses of $M_{200c} = 10^{13.4-14.3}M_{\odot}$. We characterized the stellar structures by decomposing each galaxy into a dynamically cold disk and a hot non-disk component, and studied the evolution of both the stellar and gaseous constituents. In TNG50, we find that the cold gas is quickly removed when a galaxy falls into a Fornax-mass cluster. About 87\% of the galaxies have lost $80\%$ of their star-forming gas at 4 billion years after infall, with the remaining gas concentrating in the inner regions of the galaxy. The radius of the star-forming gaseous disk decreases to half its original size at 4 billion years after infall for 66\% of the galaxies. As a result, star formation in the extended dynamically cold disk sharply decreases, even though a low level of SF persists at the center for a few additional gigayears. This leads to a tight correlation between the average stellar age in the dynamically cold disk and the infall time of galaxies. Furthermore, the luminosity fraction of the dynamically cold disk in ancient infallers is only about 1/3 of that in recent infallers, controlling for galaxy stellar mass. This quantitatively agrees with what is observed in early-type galaxies in the Fornax cluster. Gas removal stops the possible growth of the disk, with gas removed earlier in galaxies that fell in earlier, and hence the cold-disk fraction is correlated with the infall time. The stellar disk can be significantly disrupted by tidal forces after infall, through a long-term process that enhances the difference among cluster galaxies with different infall times., Comment: Accepted to A&A, 16 pages, 12 figures

Within the $\Lambda$CDM cosmology, dark matter haloes are expected to deviate from spherical symmetry. Constraining the halo shapes at large galactocentric distances is challenging due to the low density of luminous tracers. The well-studied early-type galaxy NGC 5128 (Centaurus A - CenA), has a large number of radial velocities for globular clusters (GCs) and planetary nebulae (PNe) of its extended stellar halo. In this work, we aim to determine the deviation from spherical symmetry of the dark matter halo of CenA at 5 $R_{\rm e}$ using its GCs as kinematic tracers. We used the largest photometric catalogue of GC candidates to accurately characterise the spatial distribution of the relaxed population and investigated the presence of non-relaxed structures in the kinematic catalogue of GCs using the relaxed point-symmetric velocity field as determined by the host's PNe population. We used anisotropic Jeans modelling under axisymmetric assumptions together with the Gaussian likelihood and GCs as discrete tracers. The gravitational potential is generated by flattened stellar and dark matter distributions. We leveraged different orbital properties of the blue and red GCs to model them separately. We find that discrete kinematics of the GCs are consistent with being drawn from an underlying relaxed velocity field determined from PNe. The best-fit parameters of the gravitational potential recovered from the blue and red GCs separately agree well and the joint results are: $M_{200} = 1.86^{1.61}_{-0.69}\times 10^{12}$ M$_\odot$, $M_\star/L_{\rm B} = 2.98^{+0.96}_{-0.78}$ and the flattening $q_{\rm DM} = 1.45^{+0.78}_{-0.53}$. Both GC populations show mild rotation, with red having a slightly stronger rotational signature and radially biased orbits, and blue GCs preferring negative velocity anisotropy. An oblate or a spherical dark matter halo of CenA is strongly disfavoured by our modelling., Comment: 22 pages, 17 figures, 8 tables, accepted for publication in A&A

This book chapter delves into the dynamics of continual learning, which is the process of incrementally learning from a non-stationary stream of data. Although continual learning is a natural skill for the human brain, it is very challenging for artificial neural networks. An important reason is that, when learning something new, these networks tend to quickly and drastically forget what they had learned before, a phenomenon known as catastrophic forgetting. Especially in the last decade, continual learning has become an extensively studied topic in deep learning. This book chapter reviews the insights that this field has generated., Comment: Preprint of a book chapter; 21 pages, 4 figures

In the standard cosmological model, the assembly of galaxies is primarily driven by the growth of their host dark matter halos. At the center of these halos, however, baryonic processes take over, leading to the plethora of observed galaxy properties. The coupling between baryonic and dark matter physics is central to our understanding of galaxies and yet, it remains a challenge for theoretical models and observations. Here, we demonstrate that measured ages, metallicities, stellar angular momentum, morphology and star formation rates, correlate with both stellar and halo mass. Using dynamical modeling, we find that at fixed stellar mass, CALIFA galaxies become younger, more metal-poor and rotationally supported, have higher star formation rates and later-type morphologies as their total mass increases, with independent stellar and total masses measurements. These results indicate that the formation of galaxies and thus their baryonic properties do not vary with stellar mass alone, with halo mass also playing an important role., Comment: 14 pages, 5 Figures. Accepted in Nature Astronomy

In this paper, we provide a survey on automotive surround-view fisheye optics, with an emphasis on the impact of optical artifacts on computer vision tasks in autonomous driving and ADAS. The automotive industry has advanced in applying state-of-the-art computer vision to enhance road safety and provide automated driving functionality. When using camera systems on vehicles, there is a particular need for a wide field of view to capture the entire vehicle's surroundings, in areas such as low-speed maneuvering, automated parking, and cocoon sensing. However, one crucial challenge in surround-view cameras is the strong optical aberrations of the fisheye camera, which is an area that has received little attention in the literature. Additionally, a comprehensive dataset is needed for testing safety-critical scenarios in vehicle automation. The industry has turned to simulation as a cost-effective strategy for creating synthetic datasets with surround-view camera imagery. We examine different simulation methods (such as model-driven and data-driven simulations) and discuss the simulators' ability (or lack thereof) to model real-world optical performance. Overall, this paper highlights the optical aberrations in automotive fisheye datasets, and the limitations of optical reality in simulated fisheye datasets, with a focus on computer vision in surround-view optical systems., Comment: 23 pages, 19 figures, 2 tables

Accretion disks surrounding stellar mass black holes (BHs) have been suggested as potential locations for the nucleosynthesis of light elements, which are our primary observational discriminant of multiple stellar populations within globular clusters. The population of enriched stars in globular clusters are enhanced in N14, Na23, and sometimes in Al27 and/or in K39. In this study, our aim is to investigate the feasibility of initiating nucleosynthesis for these four elements in BH accretion disks, considering various internal parameters such as the temperature of the gas and timescale of the accretion. To achieve this, we employed a 132-species reaction network. We used the slim disk model, suitable for the Super-Eddington mass accretion rate and for geometrically and optically thick disks. We explored the conditions related to the mass, mass accretion rate, viscosity, and radius of the BH-accretion disk system that would allow for the creation of N14, Na23, Al27, and K39 before the gas is accreted onto the central object. Our findings reveal that there is no region in the parameter space where the formation of Na23 can occur and only a very limited region where the formation of N14, Al27, and K39 is plausible. Specifically, this occurs for BHs with masses lower than 10 solar masses, with a preference toward even lower mass values and extremely low viscosity parameters ($\alpha <10^{-3}$). Such values are highly unlikely based on current observations of stellar mass BHs. However, such low mass BHs could actually exist in the early universe, as so-called primordial BHs. In conclusion, our study suggests that the nucleosynthesis within BH accretion disks of four elements of interest for the multiple stellar populations is improbable, but not impossible, using the slim disk model., Comment: 11 pages, 7 figures, 2 tables, Accepted by A&A

The kinematics and chemical composition of stellar populations of different ages provide crucial information about the evolution of a galaxy. We aim to provide such information for IC 1613, an isolated, gas-rich, star-forming dwarf galaxy in the Local Group. We present here the results of a new spectroscopic study performed with MUSE, an integral-field spectrograph on the Very Large Telescope. We extracted from the data cubes more than 2000 sources from which we separated stellar objects for further spectroscopic analysis. The quality of the data set allowed us to obtain accurate classifications and line-of-sight velocities for about 800 stars. Our sample includes not only Red Giant Branch (RGB) and Main Sequence (MS) stars, but also a number of probable Be and C stars. We also obtained reliable metallicities for about 300 RGB stars. The kinematic analysis revealed for the first time the presence of stellar rotation with high significance. We found general agreement with the rotation velocity of the neutral gas component. Examining the kinematics of stars as a function of broad age ranges, we find that the velocity dispersion increases as a function of age, with the behaviour being very clear in the outermost pointings, while the rotation-to-velocity dispersion support decreases. On timescales shorter than a Gyr, the stellar kinematics still follow very closely that of the neutral gas, while the two components decouple on longer timescales. The chemical analysis of the RGB stars revealed average properties comparable to other Local Group dwarf galaxies. We also provide a new estimation of the inclination angle using only independent stellar tracers. Our work provides the largest spectroscopic sample of an isolated LG dwarf galaxy. The results obtained seem to support the scenario in which the stars of a dwarf galaxy are born from a less turbulent gas over time., Comment: Accepted for publication in A&A. Changes from previous version include: a more structured text, updated figures and the addition of a discussion section (incl. a new figure). Previous results remain unchanged, but are now contextualised with respect to similar dwarf galaxies. A new estimate of the inclination angle has been added. References updated. 18 pages (10 figures, 5 tables) + appendices

$\omega$ Centauri is considered the most massive globular cluster of the Milky Way and likely the former nuclear star cluster of a galaxy accreted by the Milky Way. It is speculated to contain an intermediate-mass black hole (IMBH) from several dynamical models. However, uncertainties regarding the location of the cluster center or the retention of stellar remnants limit the robustness of the IMBH detections reported so far. In this paper, we derive and study the stellar kinematics from the highest-resolution spectroscopic data yet, using the Multi Unit Spectroscopic Explorer (MUSE) in the narrow field mode (NFM) and wide field mode (WFM). Our exceptional data near the center reveal for the first time that stars within the inner 20" ($\sim$0.5 pc) counter-rotate relative to the bulk rotation of the cluster. Using this dataset, we measure the rotation and line-of-sight velocity dispersion (LOSVD) profile out to 120$''$ with different centers proposed in the literature. We find that the velocity dispersion profiles using different centers match well with those previously published. Based on the counter--rotation, we determine a kinematic center and look for any signs of an IMBH using the high-velocity stars close to the center. We do not find any significant outliers $>$60 km/s within the central 20$''$, consistent with no IMBH being present at the center of $\omega$ Centauri. A detailed analysis of Jeans' modeling of the putative IMBH will be presented in the next paper of the series., Comment: Accepted for publication in MNRAS, 16 pages, 9 figures

The Modulation Transfer Function (MTF) is an important image quality metric typically used in the automotive domain. However, despite the fact that optical quality has an impact on the performance of computer vision in vehicle automation, for many public datasets, this metric is unknown. Additionally, wide field-of-view (FOV) cameras have become increasingly popular, particularly for low-speed vehicle automation applications. To investigate image quality in datasets, this paper proposes an adaptation of the Natural Scenes Spatial Frequency Response (NS-SFR) algorithm to suit cameras with a wide field-of-view., Comment: Accepted for publication in the Electronic Imagine Autonomous Vehicles and Machines (EI-AVM) Conference 2024

The ability of machine learning systems to learn continually is hindered by catastrophic forgetting, the tendency of neural networks to overwrite previously acquired knowledge when learning a new task. Existing methods mitigate this problem through regularization, parameter isolation, or rehearsal, but they are typically evaluated on benchmarks comprising only a handful of tasks. In contrast, humans are able to learn over long time horizons in dynamic, open-world environments, effortlessly memorizing unfamiliar objects and reliably recognizing them under various transformations. To make progress towards closing this gap, we introduce Infinite dSprites, a parsimonious tool for creating continual classification and disentanglement benchmarks of arbitrary length and with full control over generative factors. We show that over a sufficiently long time horizon, the performance of all major types of continual learning methods deteriorates on this simple benchmark. This result highlights an important and previously overlooked aspect of continual learning: given a finite modelling capacity and an arbitrarily long learning horizon, efficient learning requires memorizing class-specific information and accumulating knowledge about general mechanisms. In a simple setting with direct supervision on the generative factors, we show how learning class-agnostic transformations offers a way to circumvent catastrophic forgetting and improve classification accuracy over time. Our approach sets the stage for continual learning over hundreds of tasks with explicit control over memorization and forgetting, emphasizing open-set classification and one-shot generalization., Comment: 10 pages, 10 figures

The debate about the universality of the stellar initial mass function (IMF) revolves around two competing lines of evidence. While measurements in the Milky Way, an archetypal spiral galaxy, seem to support an invariant IMF, the observed properties of massive early-type galaxies (ETGs) favor an IMF somehow sensitive to the local star formation conditions. The fundamental methodological and physical differences between both approaches have hampered, however, a comprehensive understanding of IMF variations. We describe here an improved modelling scheme that allows for the first time consistent IMF measurements across stellar populations with different ages and complex star formation histories. Making use of the exquisite MUSE optical data from the TIMER survey and powered by the MILES stellar population models, we show the age, metallicity, [Mg/Fe], and IMF slope maps of the inner regions of NGC 3351, a spiral galaxy with a mass similar to that of the Milky Way. The measured IMF values in NGC3351 follow the expectations from a Milky Way-like IMF, although they simultaneously show systematic and spatially coherent variations, particularly for low-mass stars. In addition, our stellar population analysis reveals the presence of metal-poor and Mg-enhanced star-forming regions that appear to be predominantly enriched by the stellar ejecta of core-collapse supernovae. Our findings showcase therefore the potential of detailed studies of young stellar populations to better understand the early stages of galaxy evolution and, in particular, the origin of the observed IMF variations beyond and within the Milky Way., Comment: 16 pages, 11 figures. Accepted for publication in Astronomy and Astrophysics

We present a study on the star formation histories (SFHs) of galaxies covering the range $10^{4}$ < M$_{\star}$/M$_{\odot}$ < $10^{12}$, leveraging full spectral fitting algorithms. Our sample consists of 31 dwarf galaxies from the SAMI-Fornax Survey with stellar masses between $10^{7}$-$10^{9.5} M_{\odot}$, early-type galaxies from the ATLAS$^{3D}$ project with stellar masses between $10^{10}$-$10^{12} M_{\odot}$, and dwarf galaxies that are satellites of Andromeda and the Milky Way, with $10^{4}$ < M$_{\star}$/M$_{\odot}$ < $10^{8}$. We find that galaxies from $10^{7}$-$10^{8} M_{\odot}$ exhibit the smallest star formation rates (SFRs), while the SFR increase as we move down or up in mass. In this sense, we find that some $10^{5} M_{\odot}$ galaxies have cumulative SFHs that are comparable to those of $10^{12} M_{\odot}$ galaxies. Our study shows that the evolution of giant galaxies is primarily governed by their internal properties, with timescales that do not depend on their environmental location. In contrast, dwarf galaxies below $10^{8} M_{\odot}$ can be significantly affected in dense environments, such as the inner regions of a cluster, that severely quench the galaxies before the assembly of their 50% present-day mass. We find that, only dwarfs with stellar masses between $10^{7}$-$10^{9} M_{\odot}$ actively form stars nowadays, while less massive galaxies seem to remain unaffected by the environment due to the expulsion of most of their gas at an early stage in their evolution. Our study highlights and corroborates a critical threshold around $10^{8}-10^{9} M_{\odot}$ in galaxy evolution from previous studies, separating more massive galaxies minimally impacted by the environment from those less massive galaxies quenched by it., Comment: Accepted for publication in MNRAS

Diffusion models are powerful generative models that achieve state-of-the-art performance in image synthesis. However, training them demands substantial amounts of data and computational resources. Continual learning would allow for incrementally learning new tasks and accumulating knowledge, thus enabling the reuse of trained models for further learning. One potentially suitable continual learning approach is generative replay, where a copy of a generative model trained on previous tasks produces synthetic data that are interleaved with data from the current task. However, standard generative replay applied to diffusion models results in a catastrophic loss in denoising capabilities. In this paper, we propose generative distillation, an approach that distils the entire reverse process of a diffusion model. We demonstrate that our approach substantially improves the continual learning performance of generative replay with only a modest increase in the computational costs., Comment: To appear in the Proceedings of the Third Conference on Lifelong Learning Agents (CoLLAs), 2024

Continual learning is a subfield of machine learning, which aims to allow machine learning models to continuously learn on new data, by accumulating knowledge without forgetting what was learned in the past. In this work, we take a step back, and ask: "Why should one care about continual learning in the first place?". We set the stage by examining recent continual learning papers published at four major machine learning conferences, and show that memory-constrained settings dominate the field. Then, we discuss five open problems in machine learning, and even though they might seem unrelated to continual learning at first sight, we show that continual learning will inevitably be part of their solution. These problems are model editing, personalization and specialization, on-device learning, faster (re-)training and reinforcement learning. Finally, by comparing the desiderata from these unsolved problems and the current assumptions in continual learning, we highlight and discuss four future directions for continual learning research. We hope that this work offers an interesting perspective on the future of continual learning, while displaying its potential value and the paths we have to pursue in order to make it successful. This work is the result of the many discussions the authors had at the Dagstuhl seminar on Deep Continual Learning, in March 2023.

Deep learning models have demonstrated remarkable results for various computer vision tasks, including the realm of medical imaging. However, their application in the medical domain is limited due to the requirement for large amounts of training data, which can be both challenging and expensive to obtain. To mitigate this, pre-trained models have been fine-tuned on domain-specific data, but such an approach can suffer from inductive biases. Furthermore, deep learning models struggle to learn the relationship between spatially distant features and their importance, as convolution operations treat all pixels equally. Pioneering a novel solution to this challenge, we employ the Image Foresting Transform to optimally segment images into superpixels. These superpixels are subsequently transformed into graph-structured data, enabling the proficient extraction of features and modeling of relationships using Graph Neural Networks (GNNs). Our method harnesses an ensemble of three distinct GNN architectures to boost its robustness. In our evaluations targeting pneumonia classification, our methodology surpassed prevailing Deep Neural Networks (DNNs) in performance, all while drastically cutting down on the parameter count. This not only trims down the expenses tied to data but also accelerates training and minimizes bias. Consequently, our proposition offers a sturdy, economically viable, and scalable strategy for medical image classification, significantly diminishing dependency on extensive training data sets., Comment: Our code is available at https://github.com/aryan-at-ul/AICS_2023_submission

Recent years have seen considerable progress in the continual training of deep neural networks, predominantly thanks to approaches that add replay or regularization terms to the loss function to approximate the joint loss over all tasks so far. However, we show that even with a perfect approximation to the joint loss, these approaches still suffer from temporary but substantial forgetting when starting to train on a new task. Motivated by this 'stability gap', we propose that continual learning strategies should focus not only on the optimization objective, but also on the way this objective is optimized. While there is some continual learning work that alters the optimization trajectory (e.g., using gradient projection techniques), this line of research is positioned as alternative to improving the optimization objective, while we argue it should be complementary. In search of empirical support for our proposition, we perform a series of pre-registered experiments combining replay-approximated joint objectives with gradient projection-based optimization routines. However, this first experimental attempt fails to show clear and consistent benefits. Nevertheless, our conceptual arguments, as well as some of our empirical results, demonstrate the distinctive importance of the optimization trajectory in continual learning, thereby opening up a new direction for continual learning research., Comment: Full paper version of pre-registered report accepted at the 1st ContinualAI Unconference. The originally submitted pre-registered proposal can be found at arXiv:2311.04898v1

Machine learning has rapidly become a tool of choice for the astronomical community. It is being applied across a wide range of wavelengths and problems, from the classification of transients to neural network emulators of cosmological simulations, and is shifting paradigms about how we generate and report scientific results. At the same time, this class of method comes with its own set of best practices, challenges, and drawbacks, which, at present, are often reported on incompletely in the astrophysical literature. With this paper, we aim to provide a primer to the astronomical community, including authors, reviewers, and editors, on how to implement machine learning models and report their results in a way that ensures the accuracy of the results, reproducibility of the findings, and usefulness of the method., Comment: 14 pages, 3 figures; submitted to the Bulletin of the American Astronomical Society

The goal of this work is to probe the total mass distribution of early-type galaxies with globular clusters (GCs) as kinematic tracers, by constraining the parameters of the profile with a flexible modelling approach. To that end, we leverage the extended spatial distribution of GCs from the SLUGGS survey ($\langle R_{\rm GC,\ max} \rangle \sim 8R_{\rm e}$) in combination with discrete dynamical modelling. We use discrete Jeans anisotropic modelling in cylindrical coordinates to determine the velocity moments at the location of the GCs in our sample. We use a Bayesian framework to determine the best-fit parameters of the total mass density profile and orbital properties of the GC systems. We find that the orbital properties (anisotropy and rotation of the dispersion-dominated GC systems) minimally impact the measurements of the inner slope and enclosed mass, while a strong presence of dynamically-distinct subpopulations or low numbers of kinematic tracers can bias the results. Owing to the large spatial extent of the tracers our method is sensitive to the intrinsic inner slope of the total mass profile and we find $\overline{\alpha} = -1.88\pm 0.01$ for 12 galaxies with robust measurements. To compare our results with literature values we fit a single power-law profile to the resulting total mass density. In the radial range 0.1-4~$R_{\rm e}$ our measured slope has a value of $\langle \gamma_{\rm tot}\rangle = -2.22\pm0.14$ and is in good agreement with the literature., Comment: 17 pages, 13 Figures, 8 Tables, Accepted for publication in A&A

We employ a recently developed population-orbit superposition technique to simultaneously fit the stellar kinematic and age maps of 82 CALIFA spiral galaxies and obtain the ages of stars in different dynamical structures. We first evaluated the capabilities of this method on CALIFA-like mock data created from the Auriga simulations. The recovered mean ages of dynamically cold, warm, and hot components match the true values well, with an observational error of up to $20\%$ in the mock age maps. For CALIFA spiral galaxies, we find that the stellar ages of the cold, warm, and hot components all increase with the stellar mass of the galaxies, from $\overline{t_{\rm cold}}\sim2.2$ Gyr, $\overline{t_{\rm warm}}\sim2.3$ Gyr, and $\overline{t_{\rm hot}}\sim2.6$ Gyr for galaxies with stellar mass $M_*<10^{10}\,\rm M_{\odot}$, to $\overline{t_{\rm cold}}\sim4.0$ Gyr, $\overline{t_{\rm warm}}\sim5.1$ Gyr, and $\overline{t_{\rm hot}}\sim5.9$ Gyr for galaxies with $M_*>10^{11}\,\rm M_{\odot}$. About $80\%$ of the galaxies in our sample have $t_{\rm hot}>t_{\rm cold}$, and the mean values of $t_{\rm hot}-t_{\rm cold}$ also increase with stellar mass, from $0.7_{-0.2}^{+0.6}$ Gyr in low-mass galaxies ($10^{8.9}\,\rm M_{\odot}