arXiv subjects

Francisco Villaescusa-Navarro

Publications and source records attributed to Francisco Villaescusa-Navarro.

At least 55 records · Page 3Linked to original sources

arXiv2025

Open Source Planning & Control System with Language Agents for Autonomous Scientific Discovery

We present a multi-agent system for automation of scientific research tasks, cmbagent (https://github.com/CMBAgents/cmbagent). The system is formed by about 30 Large Language Model (LLM) agents and implements a Planning & Control strategy to orchestrate the agentic workflow, with no human-in-the-loop at any point. Each agent specializes in a different task (performing retrieval on scientific papers and codebases, writing code, interpreting results, critiquing the output of other agents) and the system is able to execute code locally. We successfully apply cmbagent to carry out a PhD level cosmology task (the measurement of cosmological parameters using supernova data) and evaluate its performance on two benchmark sets, finding superior performance over state-of-the-art LLMs. The source code is available on GitHub, demonstration videos are also available, and the system is deployed on HuggingFace and will be available on the cloud.

cs.AI↗

arXiv2025

Evaluating Retrieval-Augmented Generation Agents for Autonomous Scientific Discovery in Astrophysics

We evaluate 9 Retrieval Augmented Generation (RAG) agent configurations on 105 Cosmology Question-Answer (QA) pairs that we built specifically for this purpose.The RAG configurations are manually evaluated by a human expert, that is, a total of 945 generated answers were assessed. We find that currently the best RAG agent configuration is with OpenAI embedding and generative model, yielding 91.4\% accuracy. Using our human evaluation results we calibrate LLM-as-a-Judge (LLMaaJ) system which can be used as a robust proxy for human evaluation. These results allow us to systematically select the best RAG agent configuration for multi-agent system for autonomous scientific discovery in astrophysics (e.g., cmbagent presented in a companion paper) and provide us with an LLMaaJ system that can be scaled to thousands of cosmology QA pairs. We make our QA dataset, human evaluation results, RAG pipelines, and LLMaaJ system publicly available for further use by the astrophysics community.

astro-ph.IM↗

arXiv2025

Boosting HI-Galaxy Cross-Clustering Signal through Higher-Order Cross-Correlations

After reionization, neutral hydrogen (HI) traces the large-scale structure (LSS) of the Universe, enabling HI intensity mapping (IM) to capture the LSS in 3D and constrain key cosmological parameters. We present a new framework utilizing higher-order cross-correlations to study HI clustering around galaxies, tested using real-space data from the IllustrisTNG300 simulation. This approach computes the joint distributions of $k$-nearest neighbor ($k$NN) optical galaxies and the HI brightness temperature field smoothed at relevant scales (the $k$NN-field framework), providing sensitivity to all higher-order cross-correlations, unlike two-point statistics. To simulate HI data from actual surveys, we add random thermal noise and apply a simple foreground cleaning model, filtering out Fourier modes of the brightness temperature field with $k_\parallel < k_{\rm min,\parallel}$. Under current levels of thermal noise and foreground cleaning, typical of a Canadian Hydrogen Intensity Mapping Experiment (CHIME)-like survey, the HI-galaxy cross-correlation signal in our simulations, using the $k$NN-field framework, is detectable at $>30σ$ across $r = [3,12] \, h^{-1}$Mpc. In contrast, the detectability of the standard two-point correlation function (2PCF) over the same scales depends strongly on the foreground filter: a sharp $k_\parallel$ filter can spuriously boost detection to $8σ$ due to position-space ringing, whereas a less sharp filter yields no detection. Nonetheless, we conclude that $k$NN-field cross-correlations are robustly detectable across a broad range of foreground filtering and thermal noise conditions, suggesting their potential for enhanced constraining power over 2PCFs.

astro-ph.CO↗

arXiv2025

Cosmology with Topological Deep Learning

The standard cosmological model with cold dark matter posits a hierarchical formation of structures. We introduce topological neural networks (TNNs), implemented as message-passing neural networks on higher-order structures, to effectively capture the topological information inherent in these hierarchies that traditional graph neural networks (GNNs) fail to account for. Our approach not only considers the vertices and edges that comprise a graph but also extends to higher-order cells such as tetrahedra, clusters, and hyperedges. This enables message-passing between these heterogeneous structures within a combinatorial complex. Furthermore, our TNNs are designed to conserve the $E(3)$-invariance, which refers to the symmetry arising from invariance against translations, reflections, and rotations. When applied to the Quijote suite, our TNNs achieve a significant reduction in the mean squared error. Compared to our GNNs, which lack higher-order message-passing, ClusterTNNs show improvements of up to 22% in $Ω_{\rm m}$ and 34% in $σ_8$ jointly, while the best FullTNN achieves an improvement of up to 60% in $σ_8$. In the context of the CAMELS suite, our models yield results comparable to the current GNN benchmark, albeit with a slight decrease in performance. We emphasize that our topology and symmetry-aware neural networks provide enhanced expressive power in modeling the large-scale structures of our universe.

astro-ph.CO↗

arXiv2025

Cosmology from Point Clouds with Dark Matter Halos from the Quijote Simulations

We train a novel deep learning architecture to perform likelihood-free inference on the value of the cosmological parameters from halo catalogs of the Quijote N-body simulations. Our model takes as input a halo catalog where each halo is characterized by its position, mass, and velocity modulus. By construction, our model is E(3) invariant and is designed to extract information hierarchically. Unlike graph neural networks, it does not require the transformation of the input halo (or galaxy) catalog into a graph. Given its simplicity, our model can process point clouds with large numbers of points. We discuss the advantages of this class of methods but also point out their limitations and potential ways to improve them for cosmological data.

astro-ph.CO↗

arXiv2025

Probing massive neutrinos and modified gravity with redshift-space morphologies and anisotropies of large-scale structure

Strong degeneracy exists between some modified gravity (MG) models and massive neutrinos because the enhanced structure growth produced by modified gravity can be suppressed due to the free-streaming massive neutrinos. Previous works showed this degeneracy can be broken with non-Gaussian or velocity information. Therefore in this work, we focus on the large-scale structure (LSS) in redshift space and investigate for the first time the possibility of using the non-Gaussian information and velocity information captured by the 3D scalar Minkowski functionals (MFs) and the 3D Minkowski tensors (MTs) to break this degeneracy. Based on the Quijote and Quijote-MG simulations, we find the imprints on redshift space LSS left by the Hu-Sawicki $f(R)$ gravity can be discriminated from those left by massive neutrinos with these statistics. With the Fisher information formalism, we first show how the MTs extract information with their perpendicular and parallel elements for both low- and high-density regions; then we compare constraints from the power spectrum monopole and MFs in real space with those in redshift space, and investigate how the constraining power is further improved with anisotropies captured by the quadrupole and hexadecapole of the power spectrum and the MTs; finally, we combine the power spectrum multipoles with MFs plus MTs and find the constraints from the power spectrum multipoles on $Ω_{\mathrm{m}}, h, σ_8$, $M_ν$, and $f_{R_0}$ can be improved, because they are complemented with non-Gaussian information, by a factor of 3.4, 3.0, 3.3, 3.3, and 1.9 on small scales ($k_{\rm{max}}=0.5~h\rm{Mpc}^{-1},\ R_G=5~h^{-1}\rm{Mpc}$), and 2.8, 2.2, 3.4, 3.4, and 1.5 on large scales ($k_{\rm{max}}=0.25~h\rm{Mpc}^{-1},\ R_G=10~h^{-1}\rm{Mpc}$).

astro-ph.CO↗

arXiv2025

Correcting for interloper contamination in the power spectrum with neural networks

Modern slitless spectroscopic surveys, such as Euclid and the Roman Space Telescope, collect vast numbers of galaxy spectra but suffer from low signal-to-noise ratios. This often leads to incorrect redshift assignments when relying on a single emission line, due to noise spikes or contamination from non-target emission lines, commonly referred to as redshift interlopers. We propose a machine learning approach to correct the impact of interlopers at the level of measured summary statistics, focusing on the power spectrum monopole and line interlopers as a proof of concept. To model interloper effects, we use halo catalogs from the Quijote simulations as proxies for galaxies, displacing a fraction of halos by the distance corresponding to the redshift offset between target and interloper galaxies. This yields contaminated catalogs with varying interloper fractions across a wide range of cosmologies from the Quijote suite. We train a neural network on the power spectrum monopole, alone or combined with the bispectrum monopole, from contaminated mocks to estimate the interloper fraction and reconstruct the cleaned power spectrum. We evaluate performance in two settings: one with fixed cosmology and another where cosmological parameters vary under broad priors. In the fixed case, the network recovers the interloper fraction and corrects the power spectrum to better than 1% accuracy. When cosmology varies, performance degrades, but adding bispectrum information significantly improves results, reducing the interloper fraction error by 40-60%. We also study the method's performance as a function of the size of the training set and find that optimal strategies depend on the correlation between target and interloper samples: bispectrum information aids performance when target and interloper galaxies are uncorrelated, while tighter priors are more effective when the two are strongly correlated.

astro-ph.CO↗

arXiv2025

On the effects of parameters on galaxy properties in CAMELS and the predictability of $Ω_{\rm m}$

Recent analyses of cosmological hydrodynamic simulations from CAMELS have shown that machine learning models can predict the parameter describing the total matter content of the universe, $Ω_{\rm m}$, from the features of a single galaxy. We investigate the statistical properties of two of these simulation suites, IllustrisTNG and ASTRID, confirming that $Ω_{\rm m}$ induces a strong displacement on the distribution of galaxy features. We also observe that most other parameters have little to no effect on the distribution, except for the stellar-feedback parameter $A_{SN1}$, which introduces some near-degeneracies that can be broken with specific features. These two properties explain the predictability of $Ω_{\rm m}$. We use Optimal Transport to further measure the effect of parameters on the distribution of galaxy properties, which is found to be consistent with physical expectations. However, we observe discrepancies between the two simulation suites, both in the effect of $Ω_{\rm m}$ on the galaxy properties and in the distributions themselves at identical parameter values. Thus, although $Ω_{\rm m}$'s signature can be easily detected within a given simulation suite using just a single galaxy, applying this result to real observational data may prove significantly more challenging.

astro-ph.CO↗

arXiv2025

Towards Robustness Across Cosmological Simulation Models TNG, SIMBA, ASTRID, and EAGLE

The rapid advancement of large-scale cosmological simulations has opened new avenues for cosmological and astrophysical research. However, the increasing diversity among cosmological simulation models presents a challenge to the robustness. In this work, we develop the Model-Insensitive ESTimator (MIEST), a machine that can robustly estimate the cosmological parameters, $Ω_m$ and $σ_8$, from neural hydrogen maps of simulation models in the CAMELS project$-$TNG, SIMBA, ASTRID, and EAGLE. An estimator is considered robust if it possesses a consistent predictive power across all simulations, including those used during the training phase. We train our machine using multiple simulation models and ensure that it only extracts common features between the models while disregarding the model-specific features. This allows us to develop a novel model that is capable of accurately estimating parameters across a range of simulation models, without being biased towards any particular model. Upon the investigation of the latent space$-$a set of summary statistics, we find that the implementation of robustness leads to the blending of latent variables across different models, demonstrating the removal of model-specific features. In comparison to a standard machine lacking robustness, the average performance of MIEST on the unseen simulations during the training phase has been improved by $\sim17$% for $Ω_m$ and $\sim 38$% for $σ_8$. By using a machine learning approach that can extract robust, yet physical features, we hope to improve our understanding of galaxy formation and evolution in a (subgrid) model-insensitive manner, and ultimately, gain insight into the underlying physical processes responsible for robustness. This is a Learning the Universe publication.

astro-ph.CO↗

arXiv2025

Learning the Universe: $3\ h^{-1}{\rm Gpc}$ Tests of a Field Level $N$-body Simulation Emulator

We apply and test a field-level emulator for non-linear cosmic structure formation in a volume matching next-generation surveys. Inferring the cosmological parameters and initial conditions from which the particular galaxy distribution of our Universe was seeded can be achieved by comparing simulated data to observational data. Previous work has focused on building accelerated forward models that efficiently mimic these simulations. One of these accelerated forward models uses machine learning to apply a non-linear correction to the linear $z=0$ Zeldovich approximation (ZA) fields, closely matching the cosmological statistics in the $N$-body simulation. This emulator was trained and tested at $(h^{-1}{\rm Gpc})^3$ volumes, although cosmological inference requires significantly larger volumes. We test this emulator at $(3\ h^{-1}{\rm Gpc})^3$ by comparing emulator outputs to $N$-body simulations for eight unique cosmologies. We consider several summary statistics, applied to both the raw particle fields and the dark matter (DM) haloes. We find that the power spectrum, bispectrum and wavelet statistics of the raw particle fields agree with the $N$-body simulations within ${\sim} 5 \%$ at most scales. For the haloes, we find a similar agreement between the emulator and the $N$-body for power spectrum and bispectrum, though a comparison of the stacked profiles of haloes shows that the emulator has slight errors in the positions of particles in the highly non-linear interior of the halo. At these large $(3\ h^{-1}{\rm Gpc})^3$ volumes, the emulator can create $z=0$ particle fields in a thousandth of the time required for $N$-body simulations and will be a useful tool for large-scale cosmological inference. This is a Learning the Universe publication.

astro-ph.CO↗

arXiv2025

X-raying CAMELS: Constraining Baryonic Feedback in the Circum-Galactic Medium with the CAMELS simulations and eRASS X-ray Observations

The circumgalactic medium (CGM) around massive galaxies plays a crucial role in regulating star formation and feedback. Using the CAMELS simulation suite, we develop emulators for the X-ray surface brightness profile and the X-ray luminosity--stellar mass scaling relation to investigate how stellar and AGN feedback shape the X-ray properties of the hot CGM. Our analysis shows that at CGM scales ($10^{12} \lesssim M_{\rm halo}/M_\odot \lesssim 10^{13}$, $10\lesssim r/{\rm kpc} \lesssim 400$), stellar feedback more significantly impacts the X-ray properties than AGN feedback within the parameters studied. Comparing the emulators to recent eROSITA All-Sky Survey observations, it was found that stronger feedback than currently implemented in the IllustrisTNG, SIMBA, and Astrid simulations is required to match observed CGM properties. However, adopting these enhanced feedback parameters causes deviations in the stellar-mass-halo-mass relations from observational constraints below the group mass scale. This tension suggests possible unaccounted systematics in X-ray CGM observations or inadequacies in the feedback models of cosmological simulations.

astro-ph.GA↗

arXiv2025

Field-level simulation-based inference with galaxy catalogs: the impact of systematic effects

It has been recently shown that a powerful way to constrain cosmological parameters from galaxy redshift surveys is to train graph neural networks to perform field-level likelihood-free inference without imposing cuts on scale. In particular, de Santi et al. (2023) developed models that could accurately infer the value of $Ω_{\rm m}$ from catalogs that only contain the positions and radial velocities of galaxies that are robust to uncertainties in astrophysics and subgrid models. However, observations are affected by many effects, including 1) masking, 2) uncertainties in peculiar velocities and radial distances, and 3) different galaxy selections. Moreover, observations only allow us to measure redshift, intertwining galaxies' radial positions and velocities. In this paper we train and test our models on galaxy catalogs, created from thousands of state-of-the-art hydrodynamic simulations run with different codes from the CAMELS project, that incorporate these observational effects. We find that, although the presence of these effects degrades the precision and accuracy of the models, and increases the fraction of catalogs where the model breaks down, the fraction of galaxy catalogs where the model performs well is over 90 %, demonstrating the potential of these models to constrain cosmological parameters even when applied to real data.

astro-ph.CO↗

arXiv2025

Quantifying Baryonic Feedback on Warm-Hot Circumgalactic Medium in CAMELS Simulations

The baryonic physics shaping galaxy formation and evolution are complex, spanning a vast range of scales and making them challenging to model. Cosmological simulations rely on subgrid models that produce significantly different predictions. Understanding how models of stellar and active galactic nuclei (AGN) feedback affect baryon behavior across different halo masses and redshifts is essential. Using the SIMBA and IllustrisTNG suites from the Cosmology and Astrophysics with MachinE Learning Simulations (CAMELS) project, we explore the effect of parameters governing the subgrid implementation of stellar and AGN feedback. We find that while IllustrisTNG shows higher cumulative feedback energy across all halos, SIMBA demonstrates a greater spread of baryons, quantified by the closure radius and circumgalactic medium (CGM) gas fraction. This suggests that feedback in SIMBA couples more effectively to baryons and drives them more efficiently within the host halo. There is evidence that different feedback modes are highly interrelated in these subgrid models. Parameters controlling stellar feedback efficiency significantly impact AGN feedback, as seen in the suppression of black hole mass growth and delayed activation of AGN feedback to higher mass halos with increasing stellar feedback efficiency in both simulations. Additionally, AGN feedback efficiency parameters affect the CGM gas fraction at low halo masses in SIMBA, hinting at complex, non-linear interactions between AGN and SNe feedback modes. Overall, we demonstrate that stellar and AGN feedback are intimately interwoven, especially at low redshift, due to subgrid implementation, resulting in halo property effects that might initially seem counterintuitive.

astro-ph.GA↗

arXiv2024

The constraining power of the Marked Power Spectrum: an analytical study

The marked power spectrum - a two-point correlation function of a transformed density field - has emerged as a promising tool for extracting cosmological information from the large-scale structure of the Universe. In this work, we present the first comprehensive analytical study of the marked power spectrum's sensitivity to primordial non-Gaussianity (PNG) of the non-local type. We extend previous effective field theory frameworks to incorporate PNG, developing a complete theoretical model that we validate against the Quijote simulation suite. Through a systematic Fisher analysis, we compare the constraining power of the marked power spectrum against traditional approaches combining the power spectrum and bispectrum (P+B). We explore different choices of mark parameters to evaluate their impact on parameter constraints, particularly focusing on equilateral and orthogonal PNG as well as neutrino masses. Our analysis shows that while marking up underdense regions yields optimal constraints in the low shot-noise regime, the marked power spectrum's performance for discrete tracers with BOSS-like number densities does not surpass that of P+B analysis at mildly non-linear scales ($k \lesssim 0.25 \,h/\text{Mpc}$). However, the marked approach offers several practical advantages, including simpler estimation procedures and potentially more manageable systematic effects. Our theoretical framework reveals how the marked power spectrum incorporates higher-order correlation information through terms resembling tree-level bispectra and power spectrum convolutions. This work establishes a robust foundation for applying marked statistics to future large-volume surveys.

astro-ph.CO↗

arXiv2024

Cosmological multifield emulator

We demonstrate the use of deep network to learn the distribution of data from state-of-the-art hydrodynamic simulations of the CAMELS project. To this end, we train a generative adversarial network to generate images composed of three different channels that represent gas density (Mgas), neutral hydrogen density (HI), and magnetic field amplitudes (B). We consider an unconstrained model and another scenario where the model is conditioned on the matter density $Ω_{\rm m}$ and the amplitude of density fluctuations $σ_{8}$. We find that the generated images exhibit great quality which is on a par with that of data, visually. Quantitatively, we find that our model generates maps whose statistical properties, quantified by probability distribution function of pixel values and auto-power spectra, agree reasonably well with those of the real maps. Moreover, the cross-correlations between fields in all maps produced by the emulator are in good agreement with those of the real images, which indicates that our model generates instances whose maps in all three channels describe the same physical region. Furthermore, a CNN regressor, which has been trained to extract $Ω_{\rm m}$ and $σ_{8}$ from CAMELS multifield dataset, recovers the cosmology from the maps generated by our conditional model, achieving $R^{2}$ = 0.96 and 0.83 corresponding to $Ω_{\rm m}$ and $σ_{8}$ respectively. This further demonstrates the great capability of the model to mimic CAMELS data. Our model can be useful for generating data that are required to analyze the information from upcoming multi-wavelength cosmological surveys.

astro-ph.CO↗

arXiv2024

Cosmological and Astrophysical Parameter Inference from Stacked Galaxy Cluster Profiles Using CAMELS-zoomGZ

We present a study on the inference of cosmological and astrophysical parameters using stacked galaxy cluster profiles. Utilizing the CAMELS-zoomGZ simulations, we explore how various cluster properties--such as X-ray surface brightness, gas density, temperature, metallicity, and Compton-y profiles--can be used to predict parameters within the 28-dimensional parameter space of the IllustrisTNG model. Through neural networks, we achieve a high correlation coefficient of 0.97 or above for all cosmological parameters, including $Ω_{\rm m}$, $H_0$, and $σ_8$, and over 0.90 for the remaining astrophysical parameters, showcasing the effectiveness of these profiles for parameter inference. We investigate the impact of different radial cuts, with bins ranging from $0.1R_{200c}$ to $0.7R_{200c}$, to simulate current observational constraints. Additionally, we perform a noise sensitivity analysis, adding up to 40\% Gaussian noise (corresponding to signal-to-noise ratios as low as 2.5), revealing that key parameters such as $Ω_{\rm m}$, $H_0$, and the IMF slope remain robust even under extreme noise conditions. We also compare the performance of full radial profiles against integrated quantities, finding that profiles generally lead to more accurate parameter inferences. Our results demonstrate that stacked galaxy cluster profiles contain crucial information on both astrophysical processes within groups and clusters and the underlying cosmology of the universe. This underscores their significance for interpreting the complex data expected from next-generation surveys and reveals, for the first time, their potential as a powerful tool for parameter inference.

astro-ph.CO↗

arXiv2024

Constraining Cosmology with Simulation-based inference and Optical Galaxy Cluster Abundance

We test the robustness of simulation-based inference (SBI) in the context of cosmological parameter estimation from galaxy cluster counts and masses in simulated optical datasets. We construct ``simulations'' using analytical models for the galaxy cluster halo mass function (HMF) and for the observed richness (number of observed member galaxies) to train and test the SBI method. We compare the SBI parameter posterior samples to those from an MCMC analysis that uses the same analytical models to construct predictions of the observed data vector. The two methods exhibit comparable performance, with reliable constraints derived for the primary cosmological parameters, ($Ω_m$ and $σ_8$), and richness-mass relation parameters. We also perform out-of-domain tests with observables constructed from galaxy cluster-sized halos in the Quijote simulations. Again, the SBI and MCMC results have comparable posteriors, with similar uncertainties and biases. Unsurprisingly, upon evaluating the SBI method on thousands of simulated data vectors that span the parameter space, SBI exhibits worsened posterior calibration metrics in the out-of-domain application. We note that such calibration tests with MCMC is less computationally feasible and highlight the potential use of SBI to stress-test limitations of analytical models, such as in the use for constructing models for inference with MCMC.

astro-ph.CO↗

arXiv2024

CHARM: Creating Halos with Auto-Regressive Multi-stage networks

To maximize the amount of information extracted from cosmological datasets, simulations that accurately represent these observations are necessary. However, traditional simulations that evolve particles under gravity by estimating particle-particle interactions (N-body simulations) are computationally expensive and prohibitive to scale to the large volumes and resolutions necessary for the upcoming datasets. Moreover, modeling the distribution of galaxies typically involves identifying virialized dark matter halos, which is also a time- and memory-consuming process for large N-body simulations, further exacerbating the computational cost. In this study, we introduce CHARM, a novel method for creating mock halo catalogs by matching the spatial, mass, and velocity statistics of halos directly from the large-scale distribution of the dark matter density field. We develop multi-stage neural spline flow-based networks to learn this mapping at redshift z=0.5 directly with computationally cheaper low-resolution particle mesh simulations instead of relying on the high-resolution N-body simulations. We show that the mock halo catalogs and painted galaxy catalogs have the same statistical properties as obtained from $N$-body simulations in both real space and redshift space. Finally, we use these mock catalogs for cosmological inference using redshift-space galaxy power spectrum, bispectrum, and wavelet-based statistics using simulation-based inference, performing the first inference with accelerated forward model simulations and finding unbiased cosmological constraints with well-calibrated posteriors. The code was developed as part of the Simons Collaboration on Learning the Universe and is publicly available at \url{https://github.com/shivampcosmo/CHARM}.

astro-ph.CO↗