Milky Way, Machine Learning, Big Data

  • Yuan-Sen Ting, Hubble Postdoctoral Fellow, Princeton University
  • When: February 07, 2019, 15:00
  • Where: Homewood, Bloomberg Physics & Astronomy Building, Rm 462
    Johns Hopkins University


Understanding physical processes responsible for the formation and evolution of galaxies like the Milky Way is a fundamental but unsolved problem in astrophysics. Fortunately, most stars are long-lived. As such, using the stars as “fossil records” (what is known as Galactic archaeology) can offer unparalleled insight into the assembly of galaxies.  In recent years, the landscape of Galactic archaeology is rapidly changing thanks to on-going large-scale surveys (astrometry, photometry, spectroscopy, asteroseismology) which provide a few orders of magnitude more stars than before. In this talk, I will discuss new “phenomenological” opportunities enabled by large surveys.  I will also discuss how machine-learning tools could leverage the big data about the Milky Way by maximally harnessing information from low-resolution stellar spectra as well as the time-series photometric fluxes of stars. I will also present the new opportunities in Galactic archaeology in the era of deep photometry and spectroscopy, such as LSST, JWST, PFS, and MSE.

Neural networks have gained much attention in recent years due to their applications in our daily life including facial and voice recognition as well as data mining. Despite their remarkable ability, neural networks are severely underutilized in physical sciences. In this talk, I will briefly explain the basic concepts as well as some exciting frontier ideas in neural networks. I will discuss the opportunities for applying this simple yet interesting idea to physical sciences, in particular, the study of the Milky Way (what is known as Galactic archaeology). In particular, I will describe how a combination of data-driven models and neural networks can be an effective tool to harness information from low-resolution spectra and to relate various fields in physical sciences — such as the studies of spectroscopy and asteroseismology in astronomy.

View All Events