QuasarNet: A new research platform for the data-driven investigation of black holes

Abstract

We present QuasarNet, a novel research platform that enables deployment of data-driven modeling techniques for the investigation of the properties of super-massive black holes. Black hole data setsobservations and simulations -have grown rapidly in the last decade in both complexity and abundance. However, our computational environments and tool sets have not matured commensurately to exhaust opportunities for discovery with these data. Our pilot study presented here is motivated by one of the fundamental open questions in understanding black hole formation and assembly across cosmic time -the nature of the black hole host galaxy and parent dark matter halo connection. To explore this, we combine and co-locate large, observational data sets of quasars, the high-redshift luminous population of accreting black holes, at z ≥ 3 alongside simulated data spanning the same cosmic epochs in QuasarNet. We demonstrate the extraction of the properties of observed quasars and their putative dark matter parent halos that permit studying their association and correspondence. In this paper, we describe the design, implementation, and operation of the publicly queryable QuasarNet database and provide examples of query types and visualizations that can be used to explore the data. Starting with data collated in QuasarNet, which will serve as training sets, we plan to utilize machine learning algorithms to predict properties of the as yet undetected, less luminous quasar population. To that ultimate goal, here we present the first key step in building the BH-galaxy-halo connection that underpins the formation and evolution of supermassive black holes. All the compiled data, codes and associated Jupyter notebooks are available on the public Google Kaggle Platform.

GaMPEN: A Machine-learning Framework for Estimating Bayesian Posteriors of Galaxy Morphological Parameters

“…Over the last decade, machine learning has been increasingly employed by astronomers for a wide variety of tasks-from identifying exoplanets to studying black holes (e.g., Hoyle 2016; Kim & Brunner 2017;Shallue & Vanderburg 2018;Sharma et al 2020;Natarajan et al 2021). Especially, convolutional neural networks (CNNs) 15 have revolutionized the field of image processing and have become increasingly popular for determining galaxy morphology (e.g., Dieleman et al 2015;Huertas-Company et al 2015;Tuccillo et al 2018;Hausen & Robertson 2020;Walmsley et al 2020;Cheng et al 2021;Vega-Ferrero et al 2021;Tarsitano et al 2022).…”

Modeling the Central Supermassive Black Hole Mass of Quasars via the LSTM Approach

“…As a statistical measurement of the combined distribution of BH mass through redshifts, the BH mass function encodes the mass growth history. Similar to the QLF, which reflects the accretion history, the BH mass function is a statistical measurement of the distribution of quasar luminosities through redshift (Natarajan et al 2021). Using neural networks, redshifts are also accurately obtained (Busca & Balland 2018).…”

“…The main quasar population data source is the NASA Extragalactic Database (NED), which contains quasars retrieved from several independent optical surveys, principally the magnitude-limited SDSS. There is no comparison between quasars from SDSS and those from other surveys when it comes to spectra and photometry (Natarajan et al 2021).…”

Morphological Parameters and Associated Uncertainties for 8 Million Galaxies in the Hyper Suprime-Cam Wide Survey

“…Over the past decade, machine learning (ML) has been increasingly used for a wide variety of tasks-from identifying exoplanets to studying black holes (e.g., Hoyle 2016; Kim & Brunner 2017;Shallue & Vanderburg 2018;Sharma et al 2020;Natarajan et al 2021). Unsurprisingly, these algorithms have become increasingly popular for determining galaxy morphology as well (e.g., Dieleman et al 2015;Huertas-Company et al 2015;Tuccillo et al 2018;Ghosh et al 2020;Hausen & Robertson 2020;Walmsley et al 2020;Cheng et al 2021;Vega-Ferrero et al 2021;Tarsitano et al 2022).…”