Accelerated Bayesian SED Modeling¶
State-of-the-art SED analyses use a Bayesian framework to infer the physical properties of galaxies from observed photometry or spectra. They require sampling from a high-dimensional space of SED model parameters and take >10-100 CPU hours per galaxy. This makes them practically infeasible for analyzing the billions of galaxies that will be observed by upcoming galaxy surveys (e.g. DESI, PFS, Rubin, Webb, and Roman).
SEDflow enables scalable Bayesian SED modeling using Amortized Neural Posterior Estimation
(ANPE), a simulation-based inference method that employs neural networks
to estimate the posterior over the full range of observations.
Once trained, SEDflow requires no additional model evaluations to estimate the posterior.
SEDflow takes ∼1 second per galaxy to obtain the posteriors of the Hahn et al. (2022a) SED
model parameters, all of which are in excellent agreement with traditional Markov Chain Monte
Carlo sampling results. For more details, check out Hahn & Melchior (2022).
PROVABGS SED Model¶
SEDflow applies ANPE to SED modeling using the recent Hahn et al. (2022a) SED model,
the state-of-the-art SPS model of the DESI PRObabilistic Value-Added Bright
Galaxy Survey (PROVABGS) catalog.
The SED of a galaxy is modeled as a composite of stellar populations defined by
stellar evolution theory, its star formation and chemical enrichment histories
(SFH and ZH), and dust attenuation.
The Hahn et al. (2022a) model utilizes a non-parametric SFH with a starburst, a
non-parametric ZH that varies with time, and a flexible dust attenuation
prescription.
NSA SEDflow Catalog¶
We apply SEDflow to 33,884 galaxies in the NASA-Sloan Atlas and construct
a probabilistic value-added catalog.
For more details on the catalog and how to download it, see [NSA SEDflow Catalog]
Authors¶
Questions or Feedback¶
If you have any questions or feedback, please feel free to reach out at changhoon.hahn@princeton.edu
