NAM2019
  • NAM2019
    • Registration
    • Key Dates & Outline Schedule
    • Practical Information
    • Exhibitors
    • Grants & Bursaries
    • Contacts
  • Science
    • Science Programme
    • Parallel Sessions
    • Plenary Talks
    • Community Session
    • Special Lunches
    • Posters
    • Presenter Guidelines
  • Social
    • What's On
    • Welcome Reception
    • RAS Awards Dinner
  • Media
  • Outreach
    • Outreach and Education Day
    • Fringe Event
    • School Visit Day
  • Lancaster
    • Travel
    • Accommodation
    • Childcare
    • Campus Map
    • About Lancaster
    • Code of Conduct
  • NAM2019
    • Registration
    • Key Dates & Outline Schedule
    • Practical Information
    • Exhibitors
    • Grants & Bursaries
    • Contacts
  • Science
    • Science Programme
    • Parallel Sessions
    • Plenary Talks
    • Community Session
    • Special Lunches
    • Posters
    • Presenter Guidelines
  • Social
    • What's On
    • Welcome Reception
    • RAS Awards Dinner
  • Media
  • Outreach
    • Outreach and Education Day
    • Fringe Event
    • School Visit Day
  • Lancaster
    • Travel
    • Accommodation
    • Childcare
    • Campus Map
    • About Lancaster
    • Code of Conduct

Thursday

Schedule

id
date time
AM
10:15
Abstract
Learning the Relationship between Galaxy Spectra and their Star Formation Histories
Thursday

Abstract details

id
Learning the Relationship between Galaxy Spectra and their Star Formation Histories
Date Submitted
2019-02-28 17:45:05
Chris
Lovell
University of Sussex
Machine Learning in Astrophysics
Talk
C. C. Lovell (University of Sussex), Viviana Acquaviva (City University New York), Peter A. Thomas (University of Sussex), Kartheik G. Iyer (Rutgers), Eric Gawiser (Rutgers), Stephen M. Wilkins (University of Sussex)
I will present a new method for inferring galaxy star formation histories (SFH) using machine learning methods coupled with two cosmological hydrodynamic simulations, EAGLE and Illustris. We train Convolutional Neural Networks to learn the relationship between synthetic galaxy spectra and high resolution SFHs. To evaluate our SFH reconstruction we use Symmetric Mean Absolute Percentage Error (SMAPE), which acts as a true percentage error in the low-error regime. On dust-attenuated spectra we achieve high test accuracy (median SMAPE = 12.0%). Including the effects of simulated experimental noise increases the error (13.2%), however this is alleviated by including multiple realisations of the noise, which increases the training set size and reduces overfitting (11.4%). We also make estimates for the experimental and modelling errors. To further evaluate the generalisation properties we apply models trained on one simulation to spectra from the other, which leads to only a small increase in the error (~16%) and recovers the input star forming sequence. We apply each trained model to SDSS DR7 spectra, and find smoother histories than in the VESPA catalogue. This new approach complements the results of existing SED fitting techniques, providing star formation histories directly motivated by the results of the latest cosmological simulations.

RAS Logo

Lancaster University Logo

STFC logo

Science Programme

  • Monday
  • Tuesday
  • Wednesday
  • Thursday
  • Posters

All attendees are expected to show respect and courtesy to other attendees and staff, and to adhere to the NAM Code of Conduct. To report harassment or violation of the code of conduct please click here.

© 2022 Royal Astronomical Society

Login