arXiv subjects

Benjamin Wandelt

Publications and source records attributed to Benjamin Wandelt.

At least 19 recordsLinked to original sources

arXiv2026

AI Scientists as Engines of Discovery: A Case for Development within Reformed Institutions

Agentic artificial intelligence (AI) systems are beginning to assist, accelerate, and partially automate scientific discovery, performing tasks that span literature synthesis, code generation, data analysis, hypothesis proposal, and model criticism. We argue that this transition is qualitative rather than incremental, and that suitably designed multi-agent systems may evolve from passive computational tools into ``AI scientists'' that can expand the hypothesis-generating and verification capacity of science. Such systems must be developed and deployed within a scientific ecosystem fit for purpose: institutions must be redesigned for verification, accountability, interpretability, and dual-use safety. We sketch how multi-agent architectures, illustrated by the prototype framework \textit{Denario}, accelerate the discovery cycle and traverse model spaces beyond human reach; examine what this implies for authorship, peer review, and the enduring role of human scientists; and close with recommendations for governing AI as an epistemic actor rather than a mere instrument.

cs.AI↗

arXiv2026

Field-Level Inference of Primordial Non-Gaussianity with the Quijote Simulation Suite

Local primordial non-Gaussianity, parameterised as $f_{\rm NL}^{\rm local}$, will be stringently constrained using state-of-the-art methods applied to next-generation galaxy redshift survey data. In this paper, in preparation for the upcoming data sets, we demonstrate for the first time the joint field-level inference of $f_{\rm NL}^{\rm local}$, nuisance parameters, and the initial conditions in realistic halo catalogues, ones which are generated through full dark-matter-only $N$-body simulations. The field-level inference algorithm optimally constrains $f_{\rm NL}^{\rm local}$ through a Bayesian forward-modelling approach at the field level, which outperforms traditional methods by leveraging the full statistical power of the data at the scales considered. In addition, we assess its performance under various design choices in the forward model, including tests of the structure formation model and resolution. We demonstrate the robustness of our approach by applying it to a subset of the \textit{Quijote} simulation suite, performing the inference at scales down to $k_{\rm max} \approx 0.1 h \rm{Mpc}^{-1}$. Compared with a power spectrum and bispectrum estimator, we find a $\sim1.3$ improvement in $\sigma(f_{\rm NL}^{\rm local})$ when applying \borg{}, while marginalising over the initial conditions and bias parameters. From the small-scale information sensitivity tests, we show that the constraints on $f_{\rm NL}^{\rm local}$ improve as we increase the resolution of the inference. These findings underscore the transformative potential of field-level inference to leverage the information available in ongoing surveys such as \textit{Euclid}, providing accurate insights into the physics of cosmic inflation and the number of fields driving it.

astro-ph.CO↗

arXiv2025

PatchNet: A hierarchical approach for neural field-level inference from Quijote Simulations

\textit{What is the cosmological information content of a cubic Gigaparsec of dark matter? } Extracting cosmological information from the non-linear matter distribution has high potential to tighten parameter constraints in the era of next-generation surveys such as Euclid, DESI, and the Vera Rubin Observatory. Traditional approaches relying on summary statistics like the power spectrum and bispectrum, though analytically tractable, fail to capture the full non-Gaussian and non-linear structure of the density field. Simulation-Based Inference (SBI) provides a powerful alternative by learning directly from forward-modeled simulations. In this work, we apply SBI to the \textit{Quijote} dark matter simulations and introduce a hierarchical method that integrates small-scale information from field sub-volumes or \textit{patches} with large-scale statistics such as power spectrum and bispectrum. This hybrid strategy is efficient both computationally and in terms of the amount of training data required. It overcomes the memory limitations associated with full-field training. We show that our approach enhances Fisher information relative to analytical summaries and matches that of a very different approach (wavelet-based statistics), providing evidence that we are estimating the full information content of the dark matter density field at the resolution of $\sim 7.8~\mathrm{Mpc}/h$.

astro-ph.CO↗

arXiv2025

The Future of Artificial Intelligence and the Mathematical and Physical Sciences (AI+MPS)

This community paper developed out of the NSF Workshop on the Future of Artificial Intelligence (AI) and the Mathematical and Physics Sciences (MPS), which was held in March 2025 with the goal of understanding how the MPS domains (Astronomy, Chemistry, Materials Research, Mathematical Sciences, and Physics) can best capitalize on, and contribute to, the future of AI. We present here a summary and snapshot of the MPS community's perspective, as of Spring/Summer 2025, in a rapidly developing field. The link between AI and MPS is becoming increasingly inextricable; now is a crucial moment to strengthen the link between AI and Science by pursuing a strategy that proactively and thoughtfully leverages the potential of AI for scientific discovery and optimizes opportunities to impact the development of AI by applying concepts from fundamental science. To achieve this, we propose activities and strategic priorities that: (1) enable AI+MPS research in both directions; (2) build up an interdisciplinary community of AI+MPS researchers; and (3) foster education and workforce development in AI for MPS researchers and students. We conclude with a summary of suggested priorities for funding agencies, educational institutions, and individual researchers to help position the MPS community to be a leader in, and take full advantage of, the transformative potential of AI+MPS.

cs.AI↗

arXiv2025

Improving gravitational wave search sensitivity with TIER: Trigger Inference using Extended strain Representation

We introduce a machine learning (ML) framework called $\texttt{TIER}$ for improving the sensitivity of gravitational wave search pipelines. Typically, search pipelines only use a small region of strain data in the vicinity of a candidate signal to construct the detection statistic. However, extended strain data ($\sim 10$ s) in the candidate's vicinity can also carry valuable complementary information. We show that this information can be efficiently captured by ML classifier models trained on sparse summary representation/features of the extended data. Our framework is easy to train and can be used with already existing candidates from any search pipeline, and without requiring expensive injection campaigns. Furthermore, the output of our model can be easily integrated into the detection statistic of a search pipeline. Using $\texttt{TIER}$ on triggers from the $\texttt{IAS-HM}$ pipeline, we find up to $\sim 20\%$ improvement in sensitive volume time in LIGO-Virgo-Kagra O3 data, with improvements concentrated in regions of high masses and unequal mass ratios. Applying our framework increases the significance of several near-threshold gravitational-wave candidates, especially in the pair-instability mass gap and intermediate-mass black hole (IMBH) ranges.

gr-qc↗

arXiv2025

How many simulations do we need for simulation-based inference in cosmology?

How many simulations do we need to train machine learning methods to extract information available from summary statistics of the cosmological density field? Neural methods have shown the potential to extract non-linear information available from cosmological data. Success depends critically on having sufficient simulations for training the networks and appropriate network architectures. In the first detailed convergence study of neural network training for cosmological inference, we show that currently available simulation suites, such as the Quijote Latin Hypercube(LH) with 2000 simulations, do not provide sufficient training data for a generic neural network to reach the optimal regime, even for the dark matter power spectrum, and in an idealized case. We discover an empirical neural scaling law that predicts how much information a neural network can extract from a highly informative summary statistic, the dark matter power spectrum, as a function of the number of simulations used to train the network, for a wide range of architectures and hyperparameters. We combine this result with the Cramer-Rao information bound to forecast the number of training simulations needed for near-optimal information extraction. To verify our method we created the largest publicly released simulation data set in cosmology, the Big Sobol Sequence(BSQ), consisting of 32,768 $\Lambda$CDM n-body simulations uniformly covering the $\Lambda$CDM parameter space. Our method enables efficient planning of simulation campaigns for machine learning applications in cosmology, while the BSQ dataset provides an unprecedented resource for studying the convergence behavior of neural networks in cosmological parameter inference. Our results suggest that new large simulation suites or new training approaches will be necessary to achieve information-optimal parameter inference from non-linear simulations.

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, $\Omega_m$ and $\sigma_8$, 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 $\Omega_m$ and $\sim 38$% for $\sigma_8$. 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↗

arXiv2024

RTFAST-Spectra: Emulation of X-ray reverberation mapping for active galactic nuclei

Bayesian analysis has begun to be more widely adopted in X-ray spectroscopy, but it has largely been constrained to relatively simple physical models due to limitations in X-ray modelling software and computation time. As a result, Bayesian analysis of numerical models with high physics complexity have remained out of reach. This is a challenge, for example when modelling the X-ray emission of accreting black hole X-ray binaries, where the slow model computations severely limit explorations of parameter space and may bias the inference of astrophysical parameters. Here, we present RTFAST-Spectra: a neural network emulator that acts as a drop in replacement for the spectral portion of the black hole X-ray reverberation model RTDIST. This is the first emulator for the reltrans model suite and the first emulator for a state-of-the-art x-ray reflection model incorporating relativistic effects with 17 physically meaningful model parameters. We use Principal Component Analysis to create a light-weight neural network that is able to preserve correlations between complex atomic lines and simple continuum, enabling consistent modelling of key parameters of scientific interest. We achieve a $\mathcal{O}(10^2)$ times speed up over the original model in the most conservative conditions with $\mathcal{O}(1\%)$ precision over all 17 free parameters in the original numerical model, taking full posterior fits from months to hours. We employ Markov Chain Monte Carlo sampling to show how we can better explore the posteriors of model parameters in simulated data and discuss the complexities in interpreting the model when fitting real data.

astro-ph.HE↗

arXiv2024

On the Significance of Covariance for Constraining Theoretical Models From Galaxy Observables

In this study, we investigate the impact of covariance within uncertainties on the inference of cosmological and astrophysical parameters, specifically focusing on galaxy stellar mass functions derived from the CAMELS simulation suite. Utilizing both Fisher analysis and Implicit Likelihood Inference (ILI), we explore how different covariance structures, including simple toy models and physics-motivated uncertainties, affect posterior distributions and parameter variances. Our methodology utilizes forward modeling via emulators that are trained on CAMELS simulations to produce stellar mass functions based on input parameters, subsequently incorporating Gaussian noise as defined by covariance matrices. We examine both toy model covariance matrices and physically motivated covariance matrices derived from observational factors like the stellar Initial Mass Function (IMF) and photometric aperture size. Our results demonstrate that covariance terms significantly influence parameter inference, often leading to tighter constraints or revealing complex, multimodal posterior distributions. These findings underscore the necessity of accounting for covariance when interpreting astrophysical observations, especially in fields where accurate parameter estimation is critical for model validation and hypothesis testing.

astro-ph.CO↗

arXiv2024

LtU-ILI: An All-in-One Framework for Implicit Inference in Astrophysics and Cosmology

This paper presents the Learning the Universe Implicit Likelihood Inference (LtU-ILI) pipeline, a codebase for rapid, user-friendly, and cutting-edge machine learning (ML) inference in astrophysics and cosmology. The pipeline includes software for implementing various neural architectures, training schemata, priors, and density estimators in a manner easily adaptable to any research workflow. It includes comprehensive validation metrics to assess posterior estimate coverage, enhancing the reliability of inferred results. Additionally, the pipeline is easily parallelizable and is designed for efficient exploration of modeling hyperparameters. To demonstrate its capabilities, we present real applications across a range of astrophysics and cosmology problems, such as: estimating galaxy cluster masses from X-ray photometry; inferring cosmology from matter power spectra and halo point clouds; characterizing progenitors in gravitational wave signals; capturing physical dust parameters from galaxy colors and luminosities; and establishing properties of semi-analytic models of galaxy formation. We also include exhaustive benchmarking and comparisons of all implemented methods as well as discussions about the challenges and pitfalls of ML inference in astronomical sciences. All code and examples are made publicly available at https://github.com/maho3/ltu-ili.

astro-ph.IM↗

arXiv2023

Optimal simulation-based Bayesian decisions

We present a framework for the efficient computation of optimal Bayesian decisions under intractable likelihoods, by learning a surrogate model for the expected utility (or its distribution) as a function of the action and data spaces. We leverage recent advances in simulation-based inference and Bayesian optimization to develop active learning schemes to choose where in parameter and action spaces to simulate. This allows us to learn the optimal action in as few simulations as possible. The resulting framework is extremely simulation efficient, typically requiring fewer model calls than the associated posterior inference task alone, and a factor of $100-1000$ more efficient than Monte-Carlo based methods. Our framework opens up new capabilities for performing Bayesian decision making, particularly in the previously challenging regime where likelihoods are intractable, and simulations expensive.

stat.ML↗

arXiv2023

Sensitivity Analysis of Simulation-Based Inference for Galaxy Clustering

Simulation-based inference (SBI) is a promising approach to leverage high fidelity cosmological simulations and extract information from the non-Gaussian, non-linear scales that cannot be modeled analytically. However, scaling SBI to the next generation of cosmological surveys faces the computational challenge of requiring a large number of accurate simulations over a wide range of cosmologies, while simultaneously encompassing large cosmological volumes at high resolution. This challenge can potentially be mitigated by balancing the accuracy and computational cost for different components of the the forward model while ensuring robust inference. To guide our steps in this, we perform a sensitivity analysis of SBI for galaxy clustering on various components of the cosmological simulations: gravity model, halo-finder and the galaxy-halo distribution models (halo-occupation distribution, HOD). We infer the $σ_8$ and $Ω_m$ using galaxy power spectrum multipoles and the bispectrum monopole assuming a galaxy number density expected from the luminous red galaxies observed using the Dark Energy Spectroscopy Instrument (DESI). We find that SBI is insensitive to changing gravity model between $N$-body simulations and particle mesh (PM) simulations. However, changing the halo-finder from friends-of-friends (FoF) to Rockstar can lead to biased estimate of $σ_8$ based on the bispectrum. For galaxy models, training SBI on more complex HOD leads to consistent inference for less complex HOD models, but SBI trained on simpler HOD models fails when applied to analyze data from a more complex HOD model. Based on our results, we discuss the outlook on cosmological simulations with a focus on applying SBI approaches to future galaxy surveys.

astro-ph.CO↗

arXiv2023

Neutrino mass constraint from an Implicit Likelihood Analysis of BOSS voids

Cosmic voids identified in the spatial distribution of galaxies provide complementary information to two-point statistics. In particular, constraints on the neutrino mass sum, $\sum m_ν$, promise to benefit from the inclusion of void statistics. We perform inference on the CMASS NGC sample of SDSS-III/BOSS with the aim of constraining $\sum m_ν$. We utilize the void size function, the void galaxy cross power spectrum, and the galaxy auto power spectrum. To extract constraints from these summary statistics we use a simulation-based approach, specifically implicit likelihood inference. We populate approximate gravity-only, particle neutrino cosmological simulations with an expressive halo occupation distribution model. With a conservative scale cut of kmax=0.15 h/Mpc and a Planck-inspired LCDM prior, we find upper bounds on $\sum m_ν$ of 0.43 and 0.35 eV from the galaxy auto power spectrum and the full data vector, respectively (95% credible interval). We observe hints that the void statistics may be most effective at constraining $\sum m_ν$ from below. We also substantiate the usual assumption that the void size function is Poisson distributed.

astro-ph.CO↗

arXiv2023

Information-Ordered Bottlenecks for Adaptive Semantic Compression

We present the information-ordered bottleneck (IOB), a neural layer designed to adaptively compress data into latent variables ordered by likelihood maximization. Without retraining, IOB nodes can be truncated at any bottleneck width, capturing the most crucial information in the first latent variables. Unifying several previous approaches, we show that IOBs achieve near-optimal compression for a given encoding architecture and can assign ordering to latent signals in a manner that is semantically meaningful. IOBs demonstrate a remarkable ability to compress embeddings of image and text data, leveraging the performance of SOTA architectures such as CNNs, transformers, and diffusion models. Moreover, we introduce a novel theory for estimating global intrinsic dimensionality with IOBs and show that they recover SOTA dimensionality estimates for complex synthetic data. Furthermore, we showcase the utility of these models for exploratory analysis through applications on heterogeneous datasets, enabling computer-aided discovery of dataset complexity.

cs.LG↗

arXiv2023

Posterior Sampling of the Initial Conditions of the Universe from Non-linear Large Scale Structures using Score-Based Generative Models

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.

astro-ph.CO↗

arXiv2022

A Framework for Obtaining Accurate Posteriors of Strong Gravitational Lensing Parameters with Flexible Priors and Implicit Likelihoods using Density Estimation

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.

astro-ph.IM↗

arXiv2022

Calibrating cosmological simulations with implicit likelihood inference using galaxy growth observables

In a novel approach employing implicit likelihood inference (ILI), also known as likelihood-free inference, we calibrate the parameters of cosmological hydrodynamic simulations against observations, which has previously been unfeasible due to the high computational cost of these simulations. For computational efficiency, we train neural networks as emulators on ~1000 cosmological simulations from the CAMELS project to estimate simulated observables, taking as input the cosmological and astrophysical parameters, and use these emulators as surrogates to the cosmological simulations. Using the cosmic star formation rate density (SFRD) and, separately, stellar mass functions (SMFs) at different redshifts, we perform ILI on selected cosmological and astrophysical parameters (Omega_m, sigma_8, stellar wind feedback, and kinetic black hole feedback) and obtain full 6-dimensional posterior distributions. In the performance test, the ILI from the emulated SFRD (SMFs) can recover the target observables with a relative error of 0.17% (0.4%). We find that degeneracies exist between the parameters inferred from the emulated SFRD, confirmed with new full cosmological simulations. We also find that the SMFs can break the degeneracy in the SFRD, which indicates that the SMFs provide complementary constraints for the parameters. Further, we find that the parameter combination inferred from an observationally-inferred SFRD reproduces the target observed SFRD very well, whereas, in the case of the SMFs, the inferred and observed SMFs show significant discrepancies that indicate potential limitations of the current galaxy formation modeling and calibration framework, and/or systematic differences and inconsistencies between observations of the stellar mass function.

astro-ph.CO↗

arXiv2022

HIFlow: Generating Diverse HI Maps and Inferring Cosmology while Marginalizing over Astrophysics using Normalizing Flows

A wealth of cosmological and astrophysical information is expected from many ongoing and upcoming large-scale surveys. It is crucial to prepare for these surveys now and develop tools that can efficiently extract most information. We present HIFlow: a fast generative model of the neutral hydrogen (HI) maps that is conditioned only on cosmology ($Ω_{m}$ and $σ_{8}$) and designed using a class of normalizing flow models, the Masked Autoregressive Flow (MAF). HIFlow is trained on the state-of-the-art simulations from the Cosmology and Astrophysics with MachinE Learning Simulations (CAMELS) project. HIFlow has the ability to generate realistic diverse maps without explicitly incorporating the expected 2D maps structure into the flow as an inductive bias. We find that HIFlow is able to reproduce the CAMELS average and standard deviation HI power spectrum (Pk) within a factor of $\lesssim$ 2, scoring a very high $R^{2} > 90\%$. By inverting the flow, HIFlow provides a tractable high-dimensional likelihood for efficient parameter inference. We show that the conditional HIFlow on cosmology is successfully able to marginalize over astrophysics at the field level, regardless of the stellar and AGN feedback strengths. This new tool represents a first step toward a more powerful parameter inference, maximizing the scientific return of future HI surveys, and opening a new avenue to minimize the loss of complex information due to data compression down to summary statistics.

astro-ph.CO↗