On the detection of stellar wakes in the Milky Way: a deep learning approach
Sven Põder, Joosep Pata, María Benito, Isaac Alonso Asensio, Claudio Dalla Vecchia · 2024
Read On the detection of stellar wakes in the Milky Way: a deep learning approach by Sven Põder, Joosep Pata, María Benito, Isaac Alonso Asensio, Claudio Dalla Vecchia (2024) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2412-02749.txt.
Introduction
Sven Põder and co-authors train deep learning to spot dark matter clumps by their wake in stars in a 2024 paper. Small dark subhalos inside the Milky Way are hard to see directly. As one plows through a star field, it leaves a wake that betrays its mass.
They simulated clumps moving through star and dark matter backgrounds matched to one distance from the galactic center. Networks learned to tell whether a clump was present and to guess its mass. Detection held up down to fairly small clumps and transferred to other distances. Star overdensity and velocity spread were the most telling features for the search.
Catalog record
| CREATOR | Sven Põder, Joosep Pata, María Benito, Isaac Alonso Asensio, Claudio Dalla Vecchia |
|---|---|
| YEAR | 2024 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 61 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2412-02749
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
On the detection of stellar wakes in the Milky Way: a deep learning approach Sven Põder, Joosep Pata, María Benito, Isaac Alonso Asensio, Claudio Dalla Vecchia Abstract Context. Due to poor observational constraints on the low-mass end of the subhalo mass function, the detection of dark matter (DM) subhalos on sub-galactic scales would provide valuable information about the nature of DM. Stellar wakes, induced by passing DM subhalos, encode information about the mass (properties) of the inducing perturber and thus serve as an indirect probe for the DM substructure within the Milky Way (MW). Aims. Our aim is to assess the viability and performance of deep learning searches for stellar wakes in the Galactic stellar halo caused by DM subhalos of varying mass. Methods. We simulate massive objects (subhalos) moving through a homogeneous medium of DM and star particles, with phase-space parameters tailored to replicate the conditions of the Galaxy at a specific distance from the Galactic center. The simulation data is used to train deep neural networks with the purpose of inferring both the presence and mass of the moving perturber. We then investigate the performance of our deep learning models and identify limitations of our current approach. Results. We present an approach that allows quantitative assessment of subhalo detectability in varying conditions of the Galactic stellar and DM halos. We find that our binary classifier is able to infer the presence of subhalos in our generated mock datasets, showing non-trivial performance down to a mass of 5 × 10 7 M ⊙ 5\times 10^{7}\rm\,M_{\odot} . In a multiple-hypothesis case, we are also able to discern between samples containing subhalos of different mass. By simulating datasets describing subhalo orbits at different Galactocentric distances, we test the robustness of our binary classification model and find that it performs well with data generated from different initial physical conditions. Out of the phase-space observables available to us, we conclude that overdensity and velocity divergence are the most important features for subhalo detection performance. 1 Introduction The standard \Lambda CDM scenario succesfully describes the behaviour of dark matter (DM) on extra-galactic scales ( Einasto 2010 ; Zavala & Frenk 2019 ) . Studies of structure formation ( Gramann 1990 ) , galaxy clustering ( Darragh-Ford et al. 2023 ) , supernova luminosities ( Perivolaropoulos & Skara 2022 ) and CMB correlation functions ( Planck Collaboration et al. 2020 ) have left little room for deviations from the CDM model on these scales. A key prediction of \Lambda CDM, that is yet to be confirmed, is the abundance of dark matter subhaloes on sub-galactic scales. In fact, studies of MW-like galaxy simulations show that the subhalo mass function (SHMF), that is the abundance of subhalos per unit mass, follows a power-law well below the galactic scale (e.g. Springel et al. 2008 ). In the absence of convincing observational evidence for small-scale dark matter clustering below subhalo masses of \approx 10^{9}\rm\,M_{\odot} , other alternative DM models (i.e. warm dark matter, self-interacting dark matter, fuzzy dark matter, etc.) are also allowed. These models impose a cutoff in the SHMF below a specific mass threshold and thus change the expected abundance of dark subhaloes orbiting galaxies ( Ostdiek et al. 2022 ; Zavala & Frenk 2019 ) . Constraining the low-mass end of the SHMF is therefore an important test of the CDM scenario as deviations from the expected SHMF behaviour could be explained by alternative DM models (e.g. Benito et al. 2020 ). This however is not an easy endeavour as subhaloes less massive than 10^{8-9}\rm\,M_{\odot} are not expected to host any stars due to their small mass and reionization effects ( Sawala et al. 2015 ; Benitez-Llambay & Frenk 2020 ) - they are dark subhaloes. In recent years, the expected count of low-mass subhalos inside a Milky Way-sized galaxy in the CDM scenario has been revised, although uncertainties remain. Garrison-Kimmel et al. 2017 report that the inclusion of baryonic physics actually supresses the size of the expected subhalo population when compared to DMO (dark matter only) simulations. The previous work has been improved on by Barry et al. 2023a , who used the FIRE-2 simulations (described in Hopkins et al. 2018 ), and found that at least 20 dark subhalos of mass >10^{6}\rm\,M_{\odot} should exist within <=30\rm\,kpc of the Galactic center. Given that these theoretical predictions are yet to be robustly validated by empirical observation, the Milky Way presents itself as an ideal laboratory for probing the low-mass end of the SHMF. Beyond the Local Group, investigating dark subhalos can be effectively pursued by observing the perturbations they impart on strongly lensed images of distant galaxies and quasars (e.g. Wagner-Carena et al. 2024 ). In recent years, deep learning methods have proven to be valuable tools in this endeavor, owing to their effectiveness in image classification tasks (for a thorough overview see e.g. Varma et al. 2020 and the references therein). Inside the Galaxy, a promising method to probe the low-mass end of the SHMF involves to search for gaps or density fluctuations in the distribution of cold stellar streams ( Bonaca & Price-Whelan 2024 ) . For example, Bonaca et al. 2019 studied the interaction of the GD-1 stream with a massive perturber whose mass range they found to be in a range of 10^{6}-10^{8}\rm\,M_{\odot} . With improved measurement data, the mass detection limit via stream perturbations could be as low as 10^{5}\rm\,M_{\odot} ( Bovy et al. 2016 ) . Another method to detect DM substructure in the Galaxy via pulsar timing array measurements, proposed in Siegel et al. 2007 , promises the detection of even lower masses. The current work focuses on the detection of stellar wakes in the MW - arguably the least studied phenomenon for DM detection in the literature. The underlying concept is built upon the notion that when a massive object moves through a field of stars, it experiences dynamical friction ( Chandrasekhar 1943 ) as it perturbs the phase-space of the surrounding stellar medium ( Mulder 1983 ) . Through gravitational interactions with the perturber, stars are pulled toward it and in time cause a relative overdensity opposite to the direction of the perturber’s movement (see e.g. Weinberg 1986 , who described this effect analytically in the context of infalling satellites). In recent years, there has been growing interest in exploring the effects of dynamical friction-induced wakes as a promising avenue for investigating dark matter substructure. In the work of Buschmann et al. 2018 , the authors develop an analytical likelihood formalism to use these perturbations in the stellar phase-space to infer the mass of the DM halo passing through the stars. A popular testbed for the detection of stellar wakes has been the MW’s largest satellite - the Large Magellanic Cloud (LMC) (see e.g. Garavito-Camargo et al. 2019 ; Tamfal et al. 2021 ; Rozier et al. 2022 ; Foote et al. 2023 ). Rozier et al. 2022 studied the response of a static Milky Way to the LMC’s infall using linear response theory. More recently, Foote et al. 2023 studied the wake produced by the infall of the LMC using idealised windtunnel simulations in the context of both CDM and FDM. Notably, they observed