As the next generation of massive galaxy surveys come online, it is becoming increasingly important to develop and understand the machine learning tools that analyze big astronomical data. Neural networks are powerful and capable of probing deep patterns in data, but must be trained carefully on large and representative data sets. As a part of the predoctoral program at the Center for Computational Astrophysics, I developed and generated a new wing of the CAMELS project encompassing one thousand large dark-matter only simulations run through a sophisticated semi-analytic model for galaxy formation. In the paper introducing the suite, I measured the clustering across the 1000 generated galaxy catalogs in order to train a simple neural network, with the goal of answering: what summary statistic gives the best constraints for the most important cosmological parameters? This suite of simulations offers enormous potential to many applications of machine learning in astrophysics.
There's lots of interesting things in progress: for example, working with other SC-SAM collaborators to efficiently store the star formation and metallicity history across all 1000+ catalogs, and being able to generate SEDs and photometry that spans a broad parameter space of cosmology and astrophysics. I'm also working to expand the astrophysical parameter space covered, as well as run other types of semi-analytic models on the CAMELS-SAM merger trees. Reach out if you'd like to collaborate!
Check out the data release website: https://camels-sam.readthedocs.io , and the updated Arxiv release below. The official ApJ publication can be found here.
There's lots of interesting things in progress: for example, working with other SC-SAM collaborators to efficiently store the star formation and metallicity history across all 1000+ catalogs, and being able to generate SEDs and photometry that spans a broad parameter space of cosmology and astrophysics. I'm also working to expand the astrophysical parameter space covered, as well as run other types of semi-analytic models on the CAMELS-SAM merger trees. Reach out if you'd like to collaborate!
Check out the data release website: https://camels-sam.readthedocs.io , and the updated Arxiv release below. The official ApJ publication can be found here.
Here is a recent talk, from August 2023, discussing the CAMELS-SAM project considering a broader perspective, and upcoming parts of the project. Many thanks to Ricardo Ogando and others at the Observatorio Nacional of Brazil for the honor of giving this seminar!
Here's a great oversight of the base CAMELS-SAM project, shared on the Cosmology Talks youtube! Thanks, Shaun. From January 2022.
Here is an introduction to the CAMELS project as a whole, for those who aren't familiar:
