NAM2019
  • NAM2021
    • Contacts
  • Science
    • Science Programme
    • Plenary Talks
    • Parallel Sessions
    • Special Lunches/Discussion Sessions
    • Poster Session
    • NAM Community Session
  • Social
    • Presidential Address
    • Herschel Concert
    • RAS Awards Ceremony
    • Virtual Stonehenge Tour
  • Media
  • Public Engagement
    • Public engagement opportunities
    • Public talk
    • Writing Skyscapes
  • Venue
    • Code of Conduct
    • Accessing the conference
    • Gather.town
    • NAM2021 Slack
    • About Bath
  • Monday
  • Tuesday
  • Wednesday
  • Thursday
  • Friday
  • Posters
  • NAM2021
    • Contacts
  • Science
    • Science Programme
    • Plenary Talks
    • Parallel Sessions
    • Special Lunches/Discussion Sessions
    • Poster Session
    • NAM Community Session
  • Social
    • Presidential Address
    • Herschel Concert
    • RAS Awards Ceremony
    • Virtual Stonehenge Tour
  • Media
  • Public Engagement
    • Public engagement opportunities
    • Public talk
    • Writing Skyscapes
  • Venue
    • Code of Conduct
    • Accessing the conference
    • Gather.town
    • NAM2021 Slack
    • About Bath
  • Monday
  • Tuesday
  • Wednesday
  • Thursday
  • Friday
  • Posters

Thursday

Schedule

id
date time
PM2
16:15
Abstract
Paying Attention to Astronomy Data
Thursday

Abstract details

id
Paying Attention to Astronomy Data
Date Submitted
2021-03-30 15:58:00
Micah
Bowles
The University of Manchester
Machine Learning Applications in Astronomy
Contributed
M. Bowles (The University of Manchester), A. M. M. Scaife (The University of Manchester, The Alan Turing Institute), F. Porter (The University of Manchester), H. Tang (The University of Manchester), D. Bastien (The University of Manchester, SKAO)
In this talk I will introduce attention as a state of the art mechanism for classification of radio galaxies using convolutional neural networks. I will present multiple attention-based convolutional neural network models, that perform on par with previous state of the art classifiers whilst using 50% fewer parameters than the next smallest traditional CNN application in this field and producing attention maps. These attention maps provide a level of interpretability for otherwise potentially daunting deep learning models. By investigating different approaches to normalisation and aggregation methods, I will present local selections that can maximise the impact of the model as a whole. The resulting attention maps can be used to interpret the classifications made by the model and I demonstrate how the most salient regions identified by our attention model align well with the regions an expert human classifier would make use of for equivalent classifications. I will also demonstrate the value of the interpretability of attention, by using the attention maps as a diagnostic tool to investigate individually mis-classified sources in the labelled test set and broadly identify if the errors arose from the model or the data.

NAM 2020 Logo AWRAS Logo

 

Bath University LogoUKRI STFC new

All attendees are expected to show respect and courtesy to other attendees and staff, and to adhere to the NAM Code of Conduct.

© 2022 Royal Astronomical Society

Login