Loading Events

Masters student from the University of Trier presenting at Pawsey

You are invited to join Pawsey who is hosting Masters student, Tabea Rettelbach from the University of Trier.  Tabea will provide a short talk as follows:

Date:  Tuesday 4th December, 2018
Time: 3.00pm
Venue: Pawsey Supercomputing Centre, Burvill Court, Kensington WA 6151

Neural Networks for Flagging Noisy MWA Images

With the imminent deployment of the Square Kilometre Array, the amount of data that will be produced by the radio telescope is ever-increasing. It is beyond the scope of what can be evaluated manually, underlining the paramount importance of automated data analysis in Astronomy. For developing GLEAM-X, a survey of the Southern radio sky, data needs to be of highest quality. However, due to external influences, this cannot always be guaranteed and some images will show signs of corruption. In order to identify affected data, deep learning techniques for image classification are applied and show that they cannot only outperform experts concerning the accuracy of predictions, but with powerful GPU architectures, they are also orders of magnitude.

All are welcome, please complete the form below to register for the presentation.