Search

Representation learning is a fundamental aspect of modern artificial intelligence, driving substantial improvements across diverse applications. While selfsupervised contrastive learning has led to significant advancements in fields like computer vision and natural language processing, its adaptation to tabular data presents unique challenges. Traditional approaches often prioritize optimizing model architecture and loss functions but may overlook the crucial task of constructing meaningful positive and negative sample pairs from various perspectives like feature interactions, instance-level patterns and batch-specific contexts. To address these challenges, we introduce TabDeco, a novel method that leverages attention-based encoding strategies across both rows and columns and employs contrastive learning framework to effectively disentangle feature representations at multiple levels, including features, instances and data batches. With the innovative feature decoupling hierarchies, TabDeco consistently surpasses existing deep learning methods and leading gradient boosting algorithms, including XG-Boost, CatBoost, and LightGBM, across various benchmark tasks, underscoring its effectiveness in advancing tabular data representation learning.

We investigate the robustness of Neural Ratio Estimators (NREs) and Neural Posterior Estimators (NPEs) to distributional shifts in the context of measuring the abundance of dark matter subhalos using strong gravitational lensing data. While these data-driven inference frameworks can be accurate on test data from the same distribution as the training sets, in real applications, it is expected that simulated training data and true observational data will differ in their distributions. We explore the behavior of a trained NRE and trained sequential NPEs to estimate the population-level parameters of dark matter subhalos from a large sample of images of strongly lensed galaxies with test data presenting distributional shifts within and beyond the bounds of the training distribution in the nuisance parameters (e.g., the background source morphology). While our results show that NREs and NPEs perform well when tested perfectly in distribution, they exhibit significant biases when confronted with slight deviations from the examples seen in the training distribution. This indicates the necessity for caution when applying NREs and NPEs to real astrophysical data, where high-dimensional underlying distributions are not perfectly known., Comment: 20 pages, 7 figures, 4 tables

Gravitational-wave (GW) parameter estimation typically assumes that instrumental noise is Gaussian and stationary. Obvious departures from this idealization are typically handled on a case-by-case basis, e.g., through bespoke procedures to ``clean'' non-Gaussian noise transients (glitches), as was famously the case for the GW170817 neutron-star binary. Although effective, manipulating the data in this way can introduce biases in the inference of key astrophysical properties, like binary precession, and compound in unpredictable ways when combining multiple observations; alternative procedures free of the same biases, like joint inference of noise and signal properties, have so far proved too computationally expensive to execute at scale. Here we take a different approach: rather than explicitly modeling individual non-Gaussianities to then apply the traditional GW likelihood, we seek to learn the true distribution of instrumental noise without presuming Gaussianity and stationarity in the first place. Assuming only noise additivity, we employ score-based diffusion models to learn an empirical noise distribution directly from detector data and then combine it with a deterministic waveform model to provide an unbiased estimate of the likelihood function. We validate the method by performing inference on a subset of GW parameters from 400 mock observations, containing real LIGO noise from either the Livingston or Hanford detectors. We show that the proposed method can recover the true parameters even in the presence of loud glitches, and that the inference is unbiased over a population of signals without applying any cleaning to the data. This work provides a promising avenue for extracting unbiased source properties in future GW observations over the coming decade., Comment: 10 pages, 3 figures

Correlations between galaxies and their supermassive black holes (SMBHs) have been observed, but the causal mechanisms remained unclear. The emerging field of causal inference now enables examining these relationships using observational data. This study, using advanced causal discovery techniques and a state-of-the-art dataset, reveals a causal link between galaxy properties and SMBH masses. In elliptical galaxies, bulge properties influence SMBH growth, while in spiral galaxies, SMBHs affect host galaxy properties, potentially through feedback in gas-rich environments. For spiral galaxies, SMBHs progressively quench star formation, whereas in elliptical galaxies, quenching is complete, and the causal connection has reversed. These findings support theoretical models of active galactic nuclei feedback regulating galaxy evolution and suggest further exploration of causal links in astrophysical and cosmological scaling relations., Comment: 34 pages, 20 figures, submitted for peer-review

Recent advances in machine learning algorithms have unlocked new insights in observational astronomy by allowing astronomers to probe new frontiers. In this article, we present a methodology to disentangle the intrinsic X-ray spectrum of galaxy clusters from the instrumental response function. Employing state-of-the-art modeling software and data mining techniques of the Chandra data archive, we construct a set of 100,000 mock Chandra spectra. We train a recurrent inference machine (RIM) to take in the instrumental response and mock observation and output the intrinsic X-ray spectrum. The RIM can recover the mock intrinsic spectrum below the 1-$\sigma$ error threshold; moreover, the RIM reconstruction of the mock observations are indistinguishable from the observations themselves. To further test the algorithm, we deconvolve extracted spectra from the central regions of the galaxy group NGC 1550, known to have a rich X-ray spectrum, and the massive galaxy clusters Abell 1795. Despite the RIM reconstructions consistently remaining below the 1-$\sigma$ noise level, the recovered intrinsic spectra did not align with modeled expectations. This discrepancy is likely attributable to the RIM's method of implicitly encoding prior information within the neural network. This approach holds promise for unlocking new possibilities in accurate spectral reconstructions and advancing our understanding of complex X-ray cosmic phenomena., Comment: Submitted to AJ

Cosmological hydrodynamical simulations, while the current state-of-the art methodology for generating theoretical predictions for the large scale structures of the Universe, are among the most expensive simulation tools, requiring upwards of 100 millions CPU hours per simulation. N-body simulations, which exclusively model dark matter and its purely gravitational interactions, represent a less resource-intensive alternative, however, they do not model galaxies, and as such cannot directly be compared to observations. In this study, we use conditional score-based models to learn a mapping from N-body to hydrodynamical simulations, specifically from dark matter density fields to the observable distribution of galaxies. We demonstrate that our model is capable of generating galaxy fields statistically consistent with hydrodynamical simulations at a fraction of the computational cost, and demonstrate our emulator is significantly more precise than traditional emulators over the scales 0.36 $h\ \text{Mpc}^{-1}$ $\leq$ k $\leq$ 3.88 $h\ \text{Mpc}^{-1}$., Comment: 7+4 pages, 3+1 figures, accepted at the ICML 2024 Workshop AI4Science

Bayesian inference for inverse problems hinges critically on the choice of priors. In the absence of specific prior information, population-level distributions can serve as effective priors for parameters of interest. With the advent of machine learning, the use of data-driven population-level distributions (encoded, e.g., in a trained deep neural network) as priors is emerging as an appealing alternative to simple parametric priors in a variety of inverse problems. However, in many astrophysical applications, it is often difficult or even impossible to acquire independent and identically distributed samples from the underlying data-generating process of interest to train these models. In these cases, corrupted data or a surrogate, e.g. a simulator, is often used to produce training samples, meaning that there is a risk of obtaining misspecified priors. This, in turn, can bias the inferred posteriors in ways that are difficult to quantify, which limits the potential applicability of these models in real-world scenarios. In this work, we propose addressing this issue by iteratively updating the population-level distributions by retraining the model with posterior samples from different sets of observations and showcase the potential of this method on the problem of background image reconstruction in strong gravitational lensing when score-based models are used as data-driven priors. We show that starting from a misspecified prior distribution, the updated distribution becomes progressively closer to the underlying population-level distribution, and the resulting posterior samples exhibit reduced bias after several updates., Comment: 17 pages, 15 figures, Submitted to The Astrophysical Journal

The first generation of ELT instruments includes an optical-infrared high-resolution spectrograph, indicated as ELT-HIRES and recently christened ANDES (ArmazoNes high Dispersion Echelle Spectrograph). ANDES consists of three fibre-fed spectrographs ([U]BV, RIZ, YJH) providing a spectral resolution of $\sim$100,000 with a minimum simultaneous wavelength coverage of 0.4-1.8 $\mu$m with the goal of extending it to 0.35-2.4 $\mu$m with the addition of a U arm to the BV spectrograph and a separate K band spectrograph. It operates both in seeing- and diffraction-limited conditions and the fibre feeding allows several, interchangeable observing modes including a single conjugated adaptive optics module and a small diffraction-limited integral field unit in the NIR. Modularity and fibre-feeding allow ANDES to be placed partly on the ELT Nasmyth platform and partly in the Coud\'e room. ANDES has a wide range of groundbreaking science cases spanning nearly all areas of research in astrophysics and even fundamental physics. Among the top science cases, there are the detection of biosignatures from exoplanet atmospheres, finding the fingerprints of the first generation of stars, tests on the stability of Nature's fundamental couplings, and the direct detection of the cosmic acceleration. The ANDES project is carried forward by a large international consortium, composed of 35 Institutes from 13 countries, forming a team of almost 300 scientists and engineers which include the majority of the scientific and technical expertise in the field that can be found in ESO member states., Comment: SPIE astronomical telescope and instrumentation 2024, in press

Gravitational lensing is the deflection of light rays due to the gravity of intervening masses. This phenomenon is observed in a variety of scales and configurations, involving any non-uniform mass such as planets, stars, galaxies, clusters of galaxies, and even the large scale structure of the universe. Strong lensing occurs when the distortions are significant and multiple images of the background source are observed. The lens objects must align on the sky of order ~1 arcsecond for galaxy-galaxy lensing, or 10's of arcseonds for cluster-galaxy lensing. As the discovery of lens systems has grown to the low thousands, these systems have become pivotal for precision measurements and addressing critical questions in astrophysics. Notably, they facilitate the measurement of the Universe's expansion rate, dark matter, supernovae, quasars, and the first stars among other topics. With future surveys expected to discover hundreds of thousands of lensing systems, the modelling and simulation of such systems must occur at orders of magnitude larger scale then ever before. Here we present `caustics`, a Python package designed to handle the extensive computational demands of modeling such a vast number of lensing systems., Comment: 13 pages, 3 figures, submitted to JOSS

Although the mass growth of supermassive black holes during the Epoch of Reionisation is expected to play a role in shaping the concurrent growth of their host-galaxies, observational evidence of feedback at z$\gtrsim$6 is still sparse. We perform the first multi-scale and multi-phase characterisation of black-hole driven outflows in the $z\sim6.6$ quasar J0923+0402 and assess how these winds impact the cold gas reservoir. We employ the SimBAL spectral synthesis to fit broad absorption line (BAL) features and find a powerful ionized outflow on $\lesssim210$ pc scale, with a kinetic power $\sim2-100$\% of the quasar luminosity. ALMA observations of [CII] emission allow us to study the morphology and kinematics of the cold gas. We detect high-velocity [CII] emission, likely associated with a cold neutral outflow at $\sim0.5-2$ kpc scale in the host-galaxy, and a bright extended [CII] halo with a size of $\sim15$ kpc. For the first time at such an early epoch, we accurately constrain the outflow energetics in both the ionized and the atomic neutral gas phases. We find such energetics to be consistent with expectations for an efficient feedback mechanism, and both ejective and preventative feedback modes are likely at play. The scales and energetics of the ionized and atomic outflows suggest that they might be associated with different quasar accretion episodes. The results of this work indicate that strong black hole feedback is occurring in quasars at $z\gtrsim6$ and is likely responsible for shaping the properties of the cold gas reservoir up to circum-galactic scales., Comment: Accepted for publication in ApJ

It is the position of the National Association of School Nurses (NASN) that regularly scheduled recess during the school day should be regarded as a childhood right that is necessary for the optimal health and educational growth of all students, and that recess should not be withheld for any student. The registered professional school nurse (hereinafter referred to as school nurse) bridges health and education and can apply leadership and collaborative skills to advocate for equitable policies that support scheduled recess and reject withholding recess (NASN, 2016, 2020). Safeguarding scheduled recess is important for promoting the physical, emotional, social, and cognitive development of all students. This statement provides the background and rationale for NASN's position. [This Position Statement was initially adopted in January 2019 and revised in January 2024.]

We propose a comprehensive sample-based method for assessing the quality of generative models. The proposed approach enables the estimation of the probability that two sets of samples are drawn from the same distribution, providing a statistically rigorous method for assessing the performance of a single generative model or the comparison of multiple competing models trained on the same dataset. This comparison can be conducted by dividing the space into non-overlapping regions and comparing the number of data samples in each region. The method only requires samples from the generative model and the test data. It is capable of functioning directly on high-dimensional data, obviating the need for dimensionality reduction. Significantly, the proposed method does not depend on assumptions regarding the density of the true distribution, and it does not rely on training or fitting any auxiliary models. Instead, it focuses on approximating the integral of the density (probability mass) across various sub-regions within the data space., Comment: 14 pages, 13 figures

We present a performant, general-purpose gradient-guided nested sampling algorithm, ${\tt GGNS}$, combining the state of the art in differentiable programming, Hamiltonian slice sampling, clustering, mode separation, dynamic nested sampling, and parallelization. This unique combination allows ${\tt GGNS}$ to scale well with dimensionality and perform competitively on a variety of synthetic and real-world problems. We also show the potential of combining nested sampling with generative flow networks to obtain large amounts of high-quality samples from the posterior distribution. This combination leads to faster mode discovery and more accurate estimates of the partition function., Comment: 10 pages, 5 figures. Code available at https://github.com/Pablo-Lemos/GGNS

Polarisation is a powerful remote-sensing tool to study the nature of particles scattering the starlight. It is widely used to characterise interplanetary dust particles in the Solar System and increasingly employed to investigate extrasolar dust in debris discs' systems. We aim to measure the scattering properties of the dust from the debris ring around HD 181327 at near-infrared wavelengths. We obtained high-contrast polarimetric images of HD 181327 in the H band with the SPHERE / IRDIS instrument on the Very Large Telescope (ESO). We complemented them with archival data from HST / NICMOS in the F110W filter reprocessed in the context of the Archival Legacy Investigations of Circumstellar Environments (ALICE) project. We developed a combined forward-modelling framework to simultaneously retrieve the scattering phase function in polarisation and intensity. We detected the debris disc around HD 181327 in polarised light and total intensity. We measured the scattering phase function and the degree of linear polarisation of the dust at 1.6 micron in the birth ring. The maximum polarisation is 23.6% +/- 2.6% and occurs between a scattering angle of 70 deg and 82 deg. We show that compact spherical particles made of a highly refractive and relatively absorbing material in a differential power-law size distribution of exponent $-3.5$ can simultaneously reproduce the polarimetric and total intensity scattering properties of the dust. This type of material cannot be obtained with a mixture of silicates, amorphous carbon, water ice, and porosity, and requires a more refracting component such as iron-bearing minerals. We reveal a striking analogy between the near-infrared polarisation of comets and that of HD 181327. The methodology developed here combining VLT/SPHERE and HST/NICMOS may be applicable in the future to combine the polarimetric capabilities of SPHERE with the sensitivity of JWST., Comment: 21 pages, 22 figures, accepted in Astronomy & Astrophysics

Symmetries have been leveraged to improve the generalization of neural networks through different mechanisms from data augmentation to equivariant architectures. However, despite their potential, their integration into neural solvers for partial differential equations (PDEs) remains largely unexplored. We explore the integration of PDE symmetries, known as Lie point symmetries, in a major family of neural solvers known as physics-informed neural networks (PINNs). We propose a loss function that informs the network about Lie point symmetries in the same way that PINN models try to enforce the underlying PDE through a loss function. Intuitively, our symmetry loss ensures that the infinitesimal generators of the Lie group conserve the PDE solutions. Effectively, this means that once the network learns a solution, it also learns the neighbouring solutions generated by Lie point symmetries. Empirical evaluations indicate that the inductive bias introduced by the Lie point symmetries of the PDEs greatly boosts the sample efficiency of PINNs., Comment: NeurIPS 2023

We consider normal almost contact structures on a Riemannian manifold and, through their associated sections of an ad-hoc twistor bundle, study their harmonicity, as sections or as maps. We rewrite these harmonicity equations in terms of the curvature tensor and find conditions relating the harmonicity of the almost contact metric and almost complex structures of the total and base spaces of the Morimoto fibration. We apply these results to homogeneous principal circle bundles over generalised flag manifolds, in particular Aloff-Wallach spaces, to mass-produce harmonic almost contact metric structures., Comment: 26 pages, this supersedes arXiv:1109.1964 [math.DG], with four authors and a slightly different title. It non only corrects arXiv:1109.1964 [math.DG] but also vastly improves it with a systematic approach to homogeneous principal circle bundles over generalised flag manifolds

We explore the use of a Neural Ratio Estimator (NRE) to determine the Hubble constant ($H_0$) in the context of time delay cosmography. Assuming a Singular Isothermal Ellipsoid (SIE) mass profile for the deflector, we simulate time delay measurements, image position measurements, and modeled lensing parameters. We train the NRE to output the posterior distribution of $H_0$ given the time delay measurements, the relative Fermat potentials (calculated from the modeled parameters and the measured image positions), the deflector redshift, and the source redshift. We compare the accuracy and precision of the NRE with traditional explicit likelihood methods in the limit where the latter is tractable and reliable, using Gaussian noise to emulate measurement uncertainties in the input parameters. The NRE posteriors track the ones from the conventional method and, while they show a slight tendency to overestimate uncertainties, they can be combined in a population inference without bias., Comment: 5+3 pages, 2+2 figures, Accepted (spotlight talk) for the Machine Learning for Astrophysics Workshop at the 40th International Conference on Machine Learning (ICML 2023)

We present AstroPhot, a fast, powerful, and user-friendly Python based astronomical image photometry solver. AstroPhot incorporates automatic differentiation and GPU (or parallel CPU) acceleration, powered by the machine learning library PyTorch. Everything: AstroPhot can fit models for sky, stars, galaxies, PSFs, and more in a principled Chi^2 forward optimization, recovering Bayesian posterior information and covariance of all parameters. Everywhere: AstroPhot can optimize forward models on CPU or GPU; across images that are large, multi-band, multi-epoch, rotated, dithered, and more. All at once: The models are optimized together, thus handling overlapping objects and including the covariance between parameters (including PSF and galaxy parameters). A number of optimization algorithms are available including Levenberg-Marquardt, Gradient descent, and No-U-Turn MCMC sampling. With an object-oriented user interface, AstroPhot makes it easy to quickly extract detailed information from complex astronomical data for individual images or large survey programs. This paper outlines novel features of the AstroPhot code and compares it to other popular astronomical image modeling software. AstroPhot is open-source, fully Python based, and freely accessible here: https://github.com/Autostronomy/AstroPhot, Comment: 18 Pages, 9 figures, published in MNRAS

Dark matter structures within strong gravitational lens galaxies and along their line of sight leave a gravitational imprint on the multiple images of lensed sources. Strong gravitational lensing provides, therefore, a key test of different dark matter models in a way that is independent of the baryonic content of matter structures on subgalactic scales. In this chapter, we describe how galaxy-scale strong gravitational lensing observations are sensitive to the physical nature of dark matter. We provide a historical perspective of the field, and review its current status. We discuss the challenges and advances in terms of data, treatment of systematic errors and theoretical predictions, that will enable one to deliver a stringent and robust test of different dark matter models in the near future. With the advent of the next generation of sky surveys, the number of known strong gravitational lens systems is expected to increase by several orders of magnitude. Coupled with high-resolution follow-up observations, these data will provide a key opportunity to constrain the properties of dark matter with strong gravitational lensing.

Reconstructing the initial conditions of the universe is a key problem in cosmology. Methods based on simulating the forward evolution of the universe have provided a way to infer initial conditions consistent with present-day observations. However, due to the high complexity of the inference problem, these methods either fail to sample a distribution of possible initial density fields or require significant approximations in the simulation model to be tractable, potentially leading to biased results. In this work, we propose the use of score-based generative models to sample realizations of the early universe given present-day observations. We infer the initial density field of full high-resolution dark matter N-body simulations from the present-day density field and verify the quality of produced samples compared to the ground truth based on summary statistics. The proposed method is capable of providing plausible realizations of the early universe density field from the initial conditions posterior distribution marginalized over cosmological parameters and can sample orders of magnitude faster than current state-of-the-art methods., Comment: 8 pages, 7 figures

Likelihood analysis is typically limited to normally distributed noise due to the difficulty of determining the probability density function of complex, high-dimensional, non-Gaussian, and anisotropic noise. This is a major limitation for precision measurements in many domains of science, including astrophysics, for example, for the analysis of the Cosmic Microwave Background, gravitational waves, gravitational lensing, and exoplanets. This work presents Score-based LIkelihood Characterization (SLIC), a framework that resolves this issue by building a data-driven noise model using a set of noise realizations from observations. We show that the approach produces unbiased and precise likelihoods even in the presence of highly non-Gaussian correlated and spatially varying noise. We use diffusion generative models to estimate the gradient of the probability density of noise with respect to data elements. In combination with the Jacobian of the physical model of the signal, we use Langevin sampling to produce independent samples from the unbiased likelihood. We demonstrate the effectiveness of the method using real data from the Hubble Space Telescope and James Webb Space Telescope., Comment: 8 pages, 4 figures

Parameter inference, i.e. inferring the posterior distribution of the parameters of a statistical model given some data, is a central problem to many scientific disciplines. Generative models can be used as an alternative to Markov Chain Monte Carlo methods for conducting posterior inference, both in likelihood-based and simulation-based problems. However, assessing the accuracy of posteriors encoded in generative models is not straightforward. In this paper, we introduce `Tests of Accuracy with Random Points' (TARP) coverage testing as a method to estimate coverage probabilities of generative posterior estimators. Our method differs from previously-existing coverage-based methods, which require posterior evaluations. We prove that our approach is necessary and sufficient to show that a posterior estimator is accurate. We demonstrate the method on a variety of synthetic examples, and show that TARP can be used to test the results of posterior inference analyses in high-dimensional spaces. We also show that our method can detect inaccurate inferences in cases where existing methods fail., Comment: 15 pages, Accepted at ICML 2023

Modeling strong gravitational lenses in order to quantify the distortions in the images of background sources and to reconstruct the mass density in the foreground lenses has been a difficult computational challenge. As the quality of gravitational lens images increases, the task of fully exploiting the information they contain becomes computationally and algorithmically more difficult. In this work, we use a neural network based on the Recurrent Inference Machine (RIM) to simultaneously reconstruct an undistorted image of the background source and the lens mass density distribution as pixelated maps. The method iteratively reconstructs the model parameters (the image of the source and a pixelated density map) by learning the process of optimizing the likelihood given the data using the physical model (a ray-tracing simulation), regularized by a prior implicitly learned by the neural network through its training data. When compared to more traditional parametric models, the proposed method is significantly more expressive and can reconstruct complex mass distributions, which we demonstrate by using realistic lensing galaxies taken from the IllustrisTNG cosmological hydrodynamic simulation., Comment: 13+7 pages, 13 figures; Accepted by The Astrophysical Journal. arXiv admin note: text overlap with arXiv:2207.01073

We use the Galaxy Morphology Posterior Estimation Network (GaMPEN) to estimate morphological parameters and associated uncertainties for $\sim 8$ million galaxies in the Hyper Suprime-Cam (HSC) Wide survey with $z \leq 0.75$ and $m \leq 23$. GaMPEN is a machine learning framework that estimates Bayesian posteriors for a galaxy's bulge-to-total light ratio ($L_B/L_T$), effective radius ($R_e$), and flux ($F$). By first training on simulations of galaxies and then applying transfer learning using real data, we trained GaMPEN with $<1\%$ of our dataset. This two-step process will be critical for applying machine learning algorithms to future large imaging surveys, such as the Rubin-Legacy Survey of Space and Time (LSST), the Nancy Grace Roman Space Telescope (NGRST), and Euclid. By comparing our results to those obtained using light-profile fitting, we demonstrate that GaMPEN's predicted posterior distributions are well-calibrated ($\lesssim 5\%$ deviation) and accurate. This represents a significant improvement over light profile fitting algorithms which underestimate uncertainties by as much as $\sim60\%$. For an overlapping sub-sample, we also compare the derived morphological parameters with values in two external catalogs and find that the results agree within the limits of uncertainties predicted by GaMPEN. This step also permits us to define an empirical relationship between the S\'ersic index and $L_B/L_T$ that can be used to convert between these two parameters. The catalog presented here represents a significant improvement in size ($\sim10 \times $), depth ($\sim4$ magnitudes), and uncertainty quantification over previous state-of-the-art bulge+disk decomposition catalogs. With this work, we also release GaMPEN's source code and trained models, which can be adapted to other datasets., Comment: 39 pages, 31 figures. Published in The Astrophysical Journal. We welcome comments and constructive criticism. Public Data Release at http://gampen.ghosharitra.com/

We report the application of implicit likelihood inference to the prediction of the macro-parameters of strong lensing systems with neural networks. This allows us to perform deep learning analysis of lensing systems within a well-defined Bayesian statistical framework to explicitly impose desired priors on lensing variables, to obtain accurate posteriors, and to guarantee convergence to the optimal posterior in the limit of perfect performance. We train neural networks to perform a regression task to produce point estimates of lensing parameters. We then interpret these estimates as compressed statistics in our inference setup and model their likelihood function using mixture density networks. We compare our results with those of approximate Bayesian neural networks, discuss their significance, and point to future directions. Based on a test set of 100,000 strong lensing simulations, our amortized model produces accurate posteriors for any arbitrary confidence interval, with a maximum percentage deviation of $1.4\%$ at $21.8\%$ confidence level, without the need for any added calibration procedure. In total, inferring 100,000 different posteriors takes a day on a single GPU, showing that the method scales well to the thousands of lenses expected to be discovered by upcoming sky surveys., Comment: Accepted for publication in The Astrophysical Journal, 17 pages, 11 figures

Inferring accurate posteriors for high-dimensional representations of the brightness of gravitationally-lensed sources is a major challenge, in part due to the difficulties of accurately quantifying the priors. Here, we report the use of a score-based model to encode the prior for the inference of undistorted images of background galaxies. This model is trained on a set of high-resolution images of undistorted galaxies. By adding the likelihood score to the prior score and using a reverse-time stochastic differential equation solver, we obtain samples from the posterior. Our method produces independent posterior samples and models the data almost down to the noise level. We show how the balance between the likelihood and the prior meet our expectations in an experiment with out-of-distribution data., Comment: 5+6 pages, 3 figures, Accepted (poster + contributed talk) for the Machine Learning and the Physical Sciences Workshop at the 36th conference on Neural Information Processing Systems (NeurIPS 2022); Corrected style file and added authors checklist