Jonah Rose

About

I am a graduate researcher at the University of Florida's Department of Astronomy. I study dark matter models that differ from cold dark matter to see if they better explain our observations of the dark matter in our Universe. More recently I am training machine learning models to determine the dark matter model from field-level images.

I am from Portland OR and got my bachelor in Physics, Math and Computer Science from the University of Oregon in 2019. I got my Masters of Astronomy from the University of Florida in 2020 and am currently working toward obtaining my PhD. In fall 2023 I will be a predoc fellow at the Center for Computational Astrophysics in NYC.

Research

I run cosmological simulations of alternative dark matter models with the IllustrisTNG galaxy formation model. Specifically, I work with models within the Effective Theory of Structure Formation (ETHOS) which allows a direct mapping from particle physics models onto inputs for cosmological simulations. The models I use contain both suppression to the initial matter power spectrum and dark matter particle self-interactions. These extra physical properties cause the dark matter to behave differently than cold dark matter and offer potential solutions to many of the small scale problems currently vexing astronomers.

Recently I have explored a new method, developed by the CAMELS collaboration, that utilizes machine learning to constrain large and uncertain parameter spaces, like the properties of dark matter. To do this, I train a CNN on images from over 1,000 cosmological simulations to determine dark matter properties from an image taken from one of the simulations. I find that these models can determine dark matter properties to much better accuracy than other state-of-the-art methods.

For my Masters project I used the galaxy catalogs (>100,000 galaxies) from the IllustrisTNG cosmological simulations with the BPASS stellar population models to determine which galaxies are most likely to host binary neutron star mergers. I found that most BNS mergers occur in Milky Way-like galaxies, but on a galaxy-by-galaxy search, more massive galaxies are most likely to host a binary neutron star merger.

Publications

Contact

Elements

i = 0;

while (!deck.isInOrder()) {
    print 'Iteration ' + i;
    deck.shuffle();
    i++;
}

print 'It took ' + i + ' iterations to sort the deck.';