Machine learning for transient recognition in difference imaging with minimum sampling effort

Yik-Lun Mong, Kendall Ackley, Duncan Galloway, Tom Killestein, Joe Lyman, Danny Steeghs, Vik Dhillon, Paul O'Brien, Gavin Ramsay, Saran Poshyachinda, Rubina Kotak, Laura Nuttall, Enric Pall'e, Don Pollacco, Eric Thrane, Martin Dyer, Krzysztof Ulaczyk, Ryan Cutter, James McCormac, Paul ChoteAndrew Levan, Tom Marsh, Elizabeth Stanway, Ben Gompertz, Klaas Wiersema, Ashley Chrimes, Alexander Obradovic, James Mullaney, Ed Daw, Stuart Littlefair, Justyn Maund, Lydia Makrygianni, Umar Burhanudin, Rhaana Starling, Rob Eyles, Spencer Tooke, Christopher Duffy, Suparerk Aukkaravittayapun, Utane Sawangwit, Supachai Awiphan, David Mkrtichian, Puji Irawati, Seppo Mattila, Teppo Heikkil"a, Rene Breton, Mark Kennedy, Daniel Mata-Sanchez, Evert Rol

Research output: Contribution to journalArticlepeer-review

59 Downloads (Pure)


The amount of observational data produced by time-domain astronomy is exponentially in-creasing. Human inspection alone is not an effective way to identify genuine transients from the data. An automatic real-bogus classifier is needed and machine learning techniques are commonly used to achieve this goal. Building a training set with a sufficiently large number of verified transients is challenging, due to the requirement of human verification. We present an approach for creating a training set by using all detections in the science images to be the sample of real detections and all detections in the difference images, which are generated by the process of difference imaging to detect transients, to be the samples of bogus detections. This strategy effectively minimizes the labour involved in the data labelling for supervised machine learning methods. We demonstrate the utility of the training set by using it to train several classifiers utilizing as the feature representation the normalized pixel values in 21-by-21 pixel stamps centered at the detection position, observed with the Gravitational-wave Optical Transient Observer (GOTO) prototype. The real-bogus classifier trained with this strategy can provide up to 95% prediction accuracy on the real detections at a false alarm rate of 1%.
Original languageEnglish
Pages (from-to)6009-6017
Number of pages9
JournalMonthly Notices of the Royal Astronomical Society
Issue number4
Early online date9 Oct 2020
Publication statusPublished - 1 Dec 2020


  • astro-ph.IM
  • methods: data analysis
  • methods: statistical
  • techniques: image processing


Dive into the research topics of 'Machine learning for transient recognition in difference imaging with minimum sampling effort'. Together they form a unique fingerprint.

Cite this