arXiv subjects

Jo Bovy

Publications and source records attributed to Jo Bovy.

At least 19 recordsLinked to original sources

arXiv2026

The Twentieth Data Release of the Sloan Digital Sky Survey: First All-Sky BOSS Spectra, eROSITA-SDSS-V Mapper Coordinated Observations, and a Preview of the Local Volume Mapper

This paper presents the twentieth data release (DR20) from the Sloan Digital Sky Survey, the third data release of its fifth generation (SDSS-V). SDSS-V is a panoptic spectroscopy survey that is mapping the stars, gas, and galaxies through three scientific programs: the Milky Way Mapper (MWM), the Local Volume Mapper (LVM), and the Black Hole Mapper (BHM). DR20 presents the first optical (BOSS) SDSS-V spectra from southern hemisphere for the MWM and BHM surveys; new optical MWM and BHM data from the northern hemisphere are also available, for a total over 3 million spectra of 1.5 million stars and half a million galaxies and quasars, with galactic and extragalactic x-ray targets coordinate with eROSITA DR2. DR20 includes integral field spectroscopy maps from LVM of six targets and 169 tiles, spanning Galactic HII regions, planetary nebulae, and nearby galaxies. Additionally, eighteen value added catalogs are also released with DR20, based on SDSS-V MWM and BHM data, and we present a new LVM visualization tool including an RGB HiPS map as a value added product.

astro-ph.GA↗

arXiv2026

Bo\"otes III is a Tidally Disrupting Ultra-Faint Dwarf Galaxy on an Eccentric Polar Orbit

We present updated systemic properties of the ultra-faint dwarf galaxy Bo\"otes III from the Southern Stellar Stream Spectroscopic Survey (S$^5$). We identify 21 high-probability members and measure a velocity dispersion of $\sigma_{v} = 1.69^{+1.03}_{-0.85}$ km s$^{-1}$, about six times smaller than the previously reported $10.7 \pm 3.5$ km s$^{-1}$, and a mean metallicity of [Fe/H] $= -2.34 \pm 0.11$. The revised dispersion brings Bo\"otes III in line with other tidally disrupting dwarfs such as Antlia II and Crater II. Orbit integrations in a Milky Way (MW) + Large Magellanic Cloud (LMC) potential confirm a highly eccentric ($e \approx 0.8$), polar ($i \approx 89.5^\circ$) orbit with a recent pericentric passage $\sim 0.14$ Gyr ago at $r_{\rm peri} \approx 9.5$ kpc. Bo\"otes III is thus likely actively tidally disrupting, as its tidal radius at pericenter, $r_t \approx 164$ pc, is only $\sim 0.35$ of its half-light radius. The unusually low dispersion also implies that Bo\"otes III has either lost most of its dark matter to tides or hosts a cored inner density profile, making it a probe of the nature of dark matter. Simulated tidal streams are broadly consistent with the Styx stellar stream, though the predicted track and kinematics are sensitive to the MW halo mass, LMC mass, and solar velocity. Bo\"otes III overlaps the Typhon stream in integrals-of-motion space but has a much lower mean metallicity, suggesting the two are not the same system but may have had a common group infall origin. Sagittarius-stream contamination prevents a direct tidal-tail detection, so deep spectroscopic follow-up remains essential, both to confirm Styx as a genuine stream and to establish it as Bo\"otes III's tidal tail.

astro-ph.GA↗

arXiv2025

Semantic search for 100M+ galaxy images using AI-generated captions

Finding scientifically interesting phenomena through slow manual labeling campaigns severely limits our ability to explore the billions of galaxy images produced by telescopes. In this work, we develop a pipeline to create a semantic search engine from completely unlabeled image data. Our method leverages Vision-Language Models (VLMs) to generate descriptions for galaxy images, then contrastively aligns a pre-trained astronomy foundation model with these embedded descriptions to produce searchable embeddings at scale. We find that current VLMs provide descriptions that are sufficiently informative to train a semantic search model that outperforms direct image similarity search. Our model, AION-Search, achieves state-of-the-art zero-shot performance on finding rare phenomena despite training on randomly selected images with no deliberate curation for rare cases. Furthermore, we introduce a VLM-based re-ranking method that nearly doubles the recall for our most challenging targets in the top-100 results. For the first time, AION-Search enables flexible semantic search for over 100 million galaxy images, enabling discovery from previously infeasible searches, including the identification of 36 new extragalactic stellar stream candidates. More broadly, our work provides an approach for making large, unlabeled scientific image archives semantically searchable, expanding data exploration capabilities in fields from Earth observation to microscopy. The code, data, and app are publicly available at https://github.com/NolanKoblischke/AION-Search

astro-ph.IM↗

arXiv2025

Fuzzy dark matter dynamical friction: stalling of globular clusters induced by dynamical heatings

We present a new implementation of fuzzy dark matter (FDM) dynamical friction within the galpy framework, enabling orbital integrations of globular clusters (GCs) across a broad range of halo-to-GC mass ratios and boson masses. In this alternative DM scenario, dynamical friction is reduced or even suppressed by heating induced by FDM density granules. We further quantify the role of baryons and solitonic cores, natural consequences of FDM in galaxies, on the efficiency of orbital decay and the long-term survival of GCs. The most significant deviations from the cold DM (CDM) paradigm arise in the dwarf-galaxy regime, where FDM dynamical friction can stall the inspiral of GCs over a Hubble time, thereby preventing their sinking into galactic centers and halting the canonical galactic cannibalism of clusters. Importantly, our FDM-only friction model should be regarded as a conservative lower bound, since the inclusion of realistic FDM cores can only strengthen the survival of GCs through core stalling. This stalling mechanism not only preserves in-situ populations that would otherwise be erased in CDM, but also strongly suppresses the mixing of in-situ and ex-situ clusters, yielding a bimodal radial distribution of GCs. Our results show that the demographics of GC systems encode a distinct dynamical signature of FDM in dwarfs. These predictions open a new pathway to constrain the boson mass parameter with upcoming Euclid DR1 observations of extragalactic GCs, while simultaneously offering a natural explanation for the long-standing Fornax timing problem.

astro-ph.GA↗

arXiv2025

The orbital anisotropy profile of the Gaia-Sausage/Enceladus accretion remnant

The Gaia-Sausage/Enceladus (GS/E) accretion remnant is one of the most important stellar populations in the Milky Way halo. Recent simulation-based work has suggested that the anisotropy profiles of remnants like GS/E decline towards the center and may be well-fit by an Osipkov-Merritt-type distribution function (DF). We study the anisotropy profile of GS/E using a chemically-selected sample of stars from APOGEE DR17 and Gaia. We find that the anisotropy profile of GS/E is high and constant with $\beta \sim 0.9$ beyond 8 kpc, dropping to $\beta \sim 0.4$ at 2 kpc. We fit a two-component Osipkov-Merritt anisotropy profile to the GS/E data, finding that a superposition of profiles with scale radii $r_{\mathrm{a}}=2$ kpc and 547 kpc, with a mixture fraction $k_\mathrm{om} = 0.88$ provide a better fit to the data than a constant anisotropy profile. Using this new model, we re-assess the density profile and mass of the GS/E remnant from a previous work that assumed a contant anisotropy, finding an increase in the derived mass from $1.5\times 10^{8}~\mathrm{M}_{\odot}$ to $2.29 ^{+0.95}_{-0.63}\times 10^{8}~\mathrm{M}_{\odot}$. In general, the superposition Osipkov-Merritt DF more satisfactorily matches the kinematics of the GS/E remnant than the constant anisotropy DF, and in the future will form a more reliable basis for modelling the remnant. Additionally, these new constraints on the kinematics of GS/E near the Galactic center are an important measurement for any future theoretical or simulation-based investigation into its nature and origin.

astro-ph.GA↗

arXiv2025

Chemodynamics of Bo\"otesI with $S^{5}$: Revised Velocity Gradient, Dark Matter Density, and Galactic Chemical Evolution Constraints

We combine new spectroscopic observations of the ultra faint dwarf galaxy (UFD) Bo\"otes I (Boo I) from the Southern Stellar Stream Spectroscopic Survey ($S^{5}$) with $\sim$15 years of archival spectroscopic data to create the largest sample of stellar kinematics and metallicities to date in any Milky Way UFD. Our combined sample includes 148 members extending out to $\sim$7 half-light radii ($r_h$), including 24 newly confirmed members, 18 binary candidates, 15 RR Lyrae stars, and 92 [Fe/H] measurements. Using this larger and more spatially extended sample, we provide updated constraints on Boo I's systemic properties, including its radial population gradients. Properly accounting for perspective rotation effects in a UFD for the first time, we detect a $4\sigma$ line-of-sight velocity gradient of $1.2\pm0.3$ km s$^{-1}$ $r_h^{-1}$ aligned along Boo I's orbit and discuss its potential tidal origins. We also infer a metallicity gradient of $-0.10\pm0.02$ dex $r_h^{-1}$ in agreement with previous studies. Using an axisymmetric Jeans model, we provide updated constraints on Boo I's dark matter density profile, which weakly favor a cusped ($\gamma=1.0^{+0.5}_{-0.6}$) dark matter profile. Lastly, we re-analyze Boo I's metallicity distribution function with a one-zone galactic chemical evolution model and place new constraints on its rapid, inefficient star formation and strong galactic outflows.

astro-ph.GA↗

arXiv2025

Sloan Digital Sky Survey-V: Pioneering Panoptic Spectroscopy

The Sloan Digital Sky Survey-V (SDSS-V) is pioneering panoptic spectroscopy: it is the first all-sky, multi-epoch, optical-to-infrared spectroscopic survey. SDSS-V is mapping the sky with multi-object spectroscopy (MOS) at telescopes in both hemispheres (the 2.5-m Sloan Foundation Telescope at Apache Point Observatory and the 100-inch du Pont Telescope at Las Campanas Observatory), where 500 zonal robotic fiber positioners feed light from a wide-field focal plane to an optical (R$\sim 2000$, 500 fibers) and a near-infrared (R$\sim 22,000$, 300 fibers) spectrograph. In addition to these MOS capabilities, the survey is pioneering ultra wide-field ($\sim$ 4000~deg$^2$) integral field spectroscopy enabled by a new dedicated facility (LVM-I) at Las Campanas Observatory, where an integral field spectrograph (IFS) with 1801 lenslet-coupled fibers arranged in a 0.5 degree diameter hexagon feeds multiple R$\sim$4000 optical spectrographs that cover 3600-9800 angstroms. SDSS-V's hardware and multi-year survey strategy are designed to decode the chemo-dynamical history of the Milky Way Galaxy and tackle fundamental open issues in stellar physics in its Milky Way Mapper program, trace the growth physics of supermassive black holes in its Black Hole Mapper program, and understand the self-regulation mechanisms and the chemical enrichment of galactic ecosystems at the energy-injection scale in its Local Volume Mapper program. The survey is well-timed to multiply the scientific output from major all-sky space missions. The SDSS-V MOS programs began robotic operations in 2021; IFS observations began in 2023 with the completion of the LVM-I facility. SDSS-V builds upon decades of heritage of SDSS's pioneering advances in data analysis, collaboration spirit, infrastructure, and product deliverables in astronomy.

astro-ph.IM↗

arXiv2025

The Nineteenth Data Release of the Sloan Digital Sky Survey

Mapping the local and distant Universe is key to our understanding of it. For decades, the Sloan Digital Sky Survey (SDSS) has made a concerted effort to map millions of celestial objects to constrain the physical processes that govern our Universe. The most recent and fifth generation of SDSS (SDSS-V) is organized into three scientific ``mappers". Milky Way Mapper (MWM) that aims to chart the various components of the Milky Way and constrain its formation and assembly, Black Hole Mapper (BHM), which focuses on understanding supermassive black holes in distant galaxies across the Universe, and Local Volume Mapper (LVM), which uses integral field spectroscopy to map the ionized interstellar medium in the local group. This paper describes and outlines the scope and content for the nineteenth data release (DR19) of SDSS and the most substantial to date in SDSS-V. DR19 is the first to contain data from all three mappers. Additionally, we also describe nine value added catalogs (VACs) that enhance the science that can be conducted with the SDSS-V data. Finally, we discuss how to access SDSS DR19 and provide illustrative examples and tutorials.

astro-ph.GA↗

arXiv2025

The ejection and detectability of high- and hyper-velocity stars by compact object binaries in globular clusters

The dense cores of Milky Way globular clusters (GCs) play host to a variety of dynamical encounters between stellar objects, which can accelerate stars to velocities high enough to escape the GC. The most extreme examples of these encounters are interactions between single GC stars and binaries including at least one compact object. These interactions can result in ejection velocities of up to several hundred $\mathrm{km \ s^{-1}}$, approaching or even exceeding the escape velocity of the Galaxy itself. In order to study whether these interactions contribute to the Galactic population of hypervelocity stars (stars moving faster than the Galactic escape speed), we combine Monte Carlo $N$-body GC simulations, observations of Galactic GCs, and a particle spray code to generate realistic populations of stars which have escaped from Milky Way GCs following star + compact object binary (S+COB) interactions. We find that over the last 500 Myr, S+COB interactions have likely ejected $\sim$6300 stars from Galactic GCs, of which $839_{-67}^{+70}$ have present-day velocities exceeding $500 \; \mathrm{km \ s^{-1}}$. Using mock photometric observations, we find that $290_{-23}^{+28}$ ejected stars are detectable in Gaia Data Release 3, however, only $1_{-1}^{+2}$ stars faster than $500 \; \mathrm{km \ s^{-1}}$ are detectable. Even so, we show that observational prospects in the upcoming Legacy Survey of Space and Time are more optimistic, and future detected fast extratidal GC stars will serve as a useful probe of GC cores.

astro-ph.GA↗

arXiv2025

SDSS-V Milky Way Mapper (MWM): ASPCAP Stellar Parameters and Abundances in SDSS-V Data Release 19

The goal of this paper is to describe the science verification of Milky Way Mapper (MWM) APOGEE Stellar Parameter and Chemical Abundances Pipeline (ASPCAP) data products published in Data Release 19 (DR19) of the fifth phase of the Sloan Digital Sky Survey (SDSS-V). We compare MWM ASPCAP atmospheric parameters T$_{\rm eff}$, log g, 24 abundances of 21 elements (carbon, nitrogen, and oxygen have multiple sources for deriving their abundance values) and their uncertainties determined from Apache Point Observatory Galactic Evolution Experiment (APOGEE) spectrograph spectra with those of the literature and evaluate their accuracy and precision. We also test the zero-point calibration of the v$_{\rm rad}$ derived by the APOGEE Data Reduction Pipeline. This data release contains ASPCAP parameters for 964,989 stars, including all APOGEE-2 targets expanded with new observations of 336,511 stars from the Apache Point Observatory observed until 4 July 2023. Overall, the new T$_{\rm eff}$ values show excellent agreement with the IRFM scale, while the surface gravities exhibit slight systematic offsets compared to asteroseisimic gravities. The estimated precision of T$_{\rm eff}$ is between 50 and 70 K for giants and 70$-$100 K for dwarfs, while surface gravities are measured with a precision of 0.07$-$0.09 dex for giants. We achieve an estimated precision of 0.02$-$0.04 dex for multiple elements, including metallicity, $\alpha$, Mg, and Si, while the precision of at least 10 elements is better than 0.1 dex.

astro-ph.SR↗

arXiv2025

The impact of the Galactic bar and the Large Magellanic Cloud on hypervelocity star trajectories

Hypervelocity stars (HVSs) ejected from the Galactic Center (GC) at speeds faster than the Galactic escape velocity are useful tools to provide insight into the Milky Way's dark matter halo. However, most characterizations of HVS orbits assume static models of the Milky Way's gravitational potential. In this work, we assess the influence of the Galactic bar and the Large Magellanic Cloud (LMC) on HVS trajectories, comparing them with those from an axisymmetric potential. We simulate 28,000 HVSs ejected over the last 100 Myr and find that ignoring the bar and LMC can cause their apparent ejection location to drift by up to 100 pc. Applying two standard HVS potential fitting methods to our sample shows that they are unable to perform as designed when non-axisymmetric effects are neglected. We calculate the angle between HVS Galactocentric position and velocity and find the LMC and bar can induce a deflection angle of up to several degrees. Using mock Gaia Data Release 4 observations, however, we show that this deflection is too small in magnitude to be measured in the near future without significantly improved observational uncertainties, particularly in heliocentric distance. Our results emphasize the need to account for the bar and LMC in modeling the Galactic potential using HVSs as a tracer.

astro-ph.GA↗

arXiv2024

Dark Galactic subhalos and the Gaia snail

Gaia has revealed a clear signal of disequilibrium in the solar neighborhood in the form of a spiral (or snail) feature in the vertical phase-space distribution. We investigate the possibility that this structure emerges from ongoing perturbations by dark $\left(10^{6} M_{\odot} - 10^8 M_{\odot}\right)$ Galactic subhalos. We develop a probabilistic model for generating subhalo orbits based on a semi-analytic model of structure formation, and combine this framework with an approximate prescription for calculating the response of the disk to external perturbations. We also develop a phenomenological treatment for the diffusion of phase-space spirals caused by gravitational scattering between stars and giant molecular clouds, a process that erases the kinematic signatures of old ($t \gtrsim 0.6$ Gyr) events. Perturbations caused by dark subhalos are, on average, orders of magnitude weaker than those caused by luminous satellite galaxies, but the ubiquity of dark halos predicted by cold dark matter makes them a more probable source of strong perturbation to the dynamics of the solar neighborhood. Dark subhalos alone do not cause enough disturbance to explain the Gaia snail, but they excite fluctuations of $\sim 0.1-0.5 \ \rm{km} \ \rm{s^{-1}}$ in the mean vertical velocity of stars near the Galactic midplane that should persist to the present day. Subhalos also produce correlations between vertical frequency and orbital angle that could be mistaken as originating from a single past disturbance. Our results motivate investigation of the Milky Way's dark satellites by characterizing their kinematic signatures in phase-space spirals across the Galaxy.

astro-ph.GA↗

arXiv2024

SpectraFM: Tuning into Stellar Foundation Models

Machine learning models in astrophysics are often limited in scope and cannot adapt to data from new instruments or tasks. We introduce SpectraFM, a Transformer-based foundation model architecture that can be pre-trained on stellar spectra from any wavelength range and instrument. SpectraFM excels in generalization by combining flexibility with knowledge transfer from pre-training, allowing it to outperform traditional machine learning methods, especially in scenarios with limited training data. Our model is pre-trained on approximately 90k examples of synthetic spectra to predict the chemical abundances (Fe, Mg, O), temperature, and specific gravity of stars. We then fine-tune the model on real spectra to adapt it to observational data before fine-tuning it further on a restricted 100-star training set in a different wavelength range to predict iron abundance. Despite a small iron-rich training set of real spectra, transfer learning from the synthetic spectra pre-training enables the model to perform well on iron-poor stars. In contrast, a neural network trained from scratch fails at this task. We investigate the Transformer attention mechanism and find that the wavelengths receiving attention carry physical information about chemical composition. By leveraging the knowledge from pre-training and its ability to handle non-spectra inputs, SpectraFM reduces the need for large training datasets and enables cross-instrument and cross-domain research. Its adaptability makes it well-suited for tackling emerging challenges in astrophysics, like extracting insights from multi-modal datasets.

astro-ph.IM↗

arXiv2024

Estimating Probability Densities with Transformer and Denoising Diffusion

Transformers are often the go-to architecture to build foundation models that ingest a large amount of training data. But these models do not estimate the probability density distribution when trained on regression problems, yet obtaining full probabilistic outputs is crucial to many fields of science, where the probability distribution of the answer can be non-Gaussian and multimodal. In this work, we demonstrate that training a probabilistic model using a denoising diffusion head on top of the Transformer provides reasonable probability density estimation even for high-dimensional inputs. The combined Transformer+Denoising Diffusion model allows conditioning the output probability density on arbitrary combinations of inputs and it is thus a highly flexible density function emulator of all possible input/output combinations. We illustrate our Transformer+Denoising Diffusion model by training it on a large dataset of astronomical observations and measured labels of stars within our Galaxy and we apply it to a variety of inference tasks to show that the model can infer labels accurately with reasonable distributions.

cs.LG↗

arXiv2024

Distribution functions for the modelling of accretion remnants in Milky Way-like galaxies: insights from IllustrisTNG

We study accretion remnants around Milky Way analogs in the IllustrisTNG simulations to determine how well commonly used distribution functions (DFs) describe their phase-space distributions. We identify 30 Milky Way analogs and 116 remnants from mergers with stellar mass ratios greater than 1:20. Two-power density profiles, as well as rotating constant-anisotropy and Osipkov-Merritt DFs are fit to the remnants. We determine that the remnants are suitable for equilibrium modelling by assessing them in the context of the Jeans equation. Each of the models we consider are reasonably able to fit the stellar remnant energy and angular momentum distribution, as well as the magnitude and shape of velocity dispersion profiles. Case studies matched to two well-known merger remnants in the stellar halo-Gaia-Sausage/Enceladus (GS/E) and Sequoia-are explored in more depth. We find good evidence that remnants with high anisotropy $\beta$, such as GS/E, are better modelled with a superposition of two Osipkov-Merritt DFs than either a constant-anisotropy model or a single Osipkov-Merritt DF. We estimate an Osipkov-Merritt profile with scale radius between 2-4 kpc would be a good first-order representation of GS/E, and comment on existing observational evidence for this as well as studies which could demonstrate it. Overall, we find that DF-based models work well for describing the kinematics of large merger remnants. Our results will be an important reference for future studies which seek to constrain both the spatial and kinematic properties of merger remnants in the Milky Way stellar halo.

astro-ph.GA↗

arXiv2024

Scaling Laws for Galaxy Images

We present the first systematic investigation of supervised scaling laws outside of an ImageNet-like context - on images of galaxies. We use 840k galaxy images and over 100M annotations by Galaxy Zoo volunteers, comparable in scale to Imagenet-1K. We find that adding annotated galaxy images provides a power law improvement in performance across all architectures and all tasks, while adding trainable parameters is effective only for some (typically more subjectively challenging) tasks. We then compare the downstream performance of finetuned models pretrained on either ImageNet-12k alone vs. additionally pretrained on our galaxy images. We achieve an average relative error rate reduction of 31% across 5 downstream tasks of scientific interest. Our finetuned models are more label-efficient and, unlike their ImageNet-12k-pretrained equivalents, often achieve linear transfer performance equal to that of end-to-end finetuning. We find relatively modest additional downstream benefits from scaling model size, implying that scaling alone is not sufficient to address our domain gap, and suggest that practitioners with qualitatively different images might benefit more from in-domain adaption followed by targeted downstream labelling.

cs.CV↗

arXiv2023

Stream Members Only: Data-Driven Characterization of Stellar Streams with Mixture Density Networks

Stellar streams are sensitive probes of the Milky Way's gravitational potential. The mean track of a stream constrains global properties of the potential, while its fine-grained surface density constrains galactic substructure. A precise characterization of streams from potentially noisy data marks a crucial step in inferring galactic structure, including the dark matter, across orders of magnitude in mass scales. Here we present a new method for constructing a smooth probability density model of stellar streams using all of the available astrometric and photometric data. To characterize a stream's morphology and kinematics, we utilize mixture density networks to represent its on-sky track, width, stellar number density, and kinematic distribution. We model the photometry for each stream as a single-stellar population, with a distance track that is simultaneously estimated from the stream's inferred distance modulus (using photometry) and parallax distribution (using astrometry). We use normalizing flows to characterize the distribution of background stars. We apply the method to the stream GD-1, and the tidal tails of Palomar 5. For both streams we obtain a catalog of stellar membership probabilities that are made publicly available. Importantly, our model is capable of handling data with incomplete phase-space observations, making our method applicable to the growing census of Milky Way stellar streams. When applied to a population of streams, the resulting membership probabilities from our model form the required input to infer the Milky Way's dark matter distribution from the scale of the stellar halo down to subhalos.

astro-ph.GA↗

arXiv2023

Towards an astronomical foundation model for stars with a Transformer-based model

Rapid strides are currently being made in the field of artificial intelligence using Transformer-based models like Large Language Models (LLMs). The potential of these methods for creating a single, large, versatile model in astronomy has not yet been explored. In this work, we propose a framework for data-driven astronomy that uses the same core techniques and architecture as used by LLMs. Using a variety of observations and labels of stars as an example, we build a Transformer-based model and train it in a self-supervised manner with cross-survey data sets to perform a variety of inference tasks. In particular, we demonstrate that a $\textit{single}$ model can perform both discriminative and generative tasks even if the model was not trained or fine-tuned to do any specific task. For example, on the discriminative task of deriving stellar parameters from Gaia XP spectra, we achieve an accuracy of 47 K in $T_\mathrm{eff}$, 0.11 dex in $\log{g}$, and 0.07 dex in $[\mathrm{M/H}]$, outperforming an expert $\texttt{XGBoost}$ model in the same setting. But the same model can also generate XP spectra from stellar parameters, inpaint unobserved spectral regions, extract empirical stellar loci, and even determine the interstellar extinction curve. Our framework demonstrates that building and training a $\textit{single}$ foundation model without fine-tuning using data and parameters from multiple surveys to predict unmeasured observations and parameters is well within reach. Such "Large Astronomy Models" trained on large quantities of observational data will play a large role in the analysis of current and future large surveys.

astro-ph.IM↗