Modelling the galaxy–halo connection with machine learning

Delgado, Anna; Wadekar, Digvijay; Hadzhiyska, Boryana; Bose, Sownak; Hernquist, Lars; Ho, Shirley

doi:10.1093/mnras/stac1951

Cited by 34 publications

(14 citation statements)

References 66 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For example, Kepler's laws were an empirical model crafted by hand to fit patterns in the orbits of planets, and ultimately inspired Newton to discover a formula for gravitational force which could explain Kepler's laws (see Hawking, 2004, for a review of this history). Astronomy has seen increasing use of symbolic regression to find new empirical models in an automated way (e.g., Graham et al, 2013;Cranmer et al, 2020;Wadekar et al, 2020;Delgado et al, 2021;Cranmer et al, 2021;Shao et al, 2021).…”

Section: Symbolic Regressionmentioning

confidence: 99%

Automated discovery of interpretable gravitational-wave population models

Wong¹,

Cranmer²

2022

Preprint

View full text Add to dashboard Cite

We present an automatic approach to discover analytic population models for gravitational-wave (GW) events from data. As more gravitationalwave (GW) events are detected, flexible models such as Gaussian Mixture Models have become more important in fitting the distribution of GW properties due to their expressivity. However, flexible models come with many parameters that lack physical motivation, making interpreting the implication of these models challenging. In this work, we demonstrate symbolic regression can complement flexible models by distilling the posterior predictive distribution of such flexible models into interpretable analytic expressions. We recover common GW population models such as a power-law-plus-Gaussian, and find a new empirical population model which combines accuracy and simplicity. This demonstrates a strategy to automatically discover interpretable population models in the ever-growing GW catalog, which can potentially be applied to other astrophysical phenomena.

show abstract

Section: Symbolic Regressionmentioning

confidence: 99%

Automated discovery of interpretable gravitational-wave population models

Wong¹,

Cranmer²

2022

Preprint

View full text Add to dashboard Cite

show abstract

“…Moreover, they operate by mechanisms regarding the evolution and formation of those, whose details are not yet fully understood [ 198 ]. Thus, various approaches have emerged for their investigation, such as searching of expressions capable to unify properties of galaxy clusters to their masses [ 197 ], studying the galaxy-halo connection [ 199 ], modelling the assembly bias [ 129 ] and estimating the total mass of a subhalo [ 198 ].…”

Section: Application In Science and Technologymentioning

confidence: 99%

Artificial Intelligence in Physical Sciences: Symbolic Regression Trends and Perspectives

Angelis

Sofos

Karakasidis

2023

Arch Computat Methods Eng

View full text Add to dashboard Cite

Symbolic regression (SR) is a machine learning-based regression method based on genetic programming principles that integrates techniques and processes from heterogeneous scientific fields and is capable of providing analytical equations purely from data. This remarkable characteristic diminishes the need to incorporate prior knowledge about the investigated system. SR can spot profound and elucidate ambiguous relations that can be generalizable, applicable, explainable and span over most scientific, technological, economical, and social principles. In this review, current state of the art is documented, technical and physical characteristics of SR are presented, the available programming techniques are investigated, fields of application are explored, and future perspectives are discussed.

show abstract

“…The advantage of RF models over other methods (for example Neural Network based algorithms) is in the interpretability of the output models, as they indicate the relative importance of the input variables in reaching a prediction. In galaxy formation, RFs (and related tree-based methods) have been widely used for regression problems using both simulation and observational data, for example in predicting the properties of large scale structure (Lucie-Smith et al 2018;Lovell et al 2022;Li et al 2022) and the properties of galaxies and haloes (Ucci et al 2017;Nadler et al 2018;Rafieferantsoa et al 2018;Cohn & Battaglia 2020;Moews et al 2021;Mucesh et al 2021;Delgado et al 2022;McGibbon & Khochfar 2022).…”

Section: Random Forest Regressionmentioning

confidence: 99%

Mapping Circumgalactic Medium Observations to Theory Using Machine Learning

Appleby¹,

Davé²,

Sorini³

et al. 2023

Preprint

View full text Add to dashboard Cite

We present a random forest framework for predicting circumgalactic medium (CGM) physical conditions from quasar absorption line observables, trained on a sample of Voigt profile-fit synthetic absorbers from the Simba cosmological simulation. Traditionally, extracting physical conditions from CGM absorber observations involves simplifying assumptions such as uniform single-phase clouds, but by using a cosmological simulation we bypass such assumptions to better capture the complex relationship between CGM observables and underlying gas conditions. We train random forest models on synthetic spectra for H i and selected metal lines around galaxies across a range of star formation rates, stellar masses, and impact parameters, to predict absorber overdensities, temperatures, and metallicities. The models reproduce the true values from Simba well, with transverse standard deviations of 0.2 − 0.3 dex in overdensity, 0.14 − 0.2 dex in temperature, and 0.16 − 0.2 dex in metallicity predicted from metal lines (not H i), across all ions. Examining the feature importance, the random forest indicates that the overdensity is most informed by the absorber column density, the temperature is driven by the line width, and the metallicity is most sensitive to the specific star formation rate. Alternatively examining feature importance by removing one observable at a time, the overdensity and metallicity appear to be more driven by the impact parameter. We introduce a normalising transform approach in order to ensure the scatter in the true physical conditions is accurately spanned by the network. The trained models are available online.

show abstract

Modelling the galaxy–halo connection with machine learning

Cited by 34 publications

References 66 publications

Automated discovery of interpretable gravitational-wave population models

Automated discovery of interpretable gravitational-wave population models

Artificial Intelligence in Physical Sciences: Symbolic Regression Trends and Perspectives

Mapping Circumgalactic Medium Observations to Theory Using Machine Learning

Contact Info

Product

Resources

About