Embedded Files
A new era in Big Data astronomy
A new era in Big Data astronomy
Big data, supercomputers and artificial intelligence have transformed every aspect of science today, and astronomy is no exception. New highly sensitive telescopes are gathering an overwhelming amount of data that has to be analysed.
One of these telescopes is known as the Vera C. Rubin Observatory, which is undertaking the 10-year Legacy Survey of Space and Time (LSST). This groundbreaking study has diverse, ambitious science goals that range from studying asteroids to fundamental cosmology. LSST will be deeper, wider and faster than any survey before: not only will it produce an image catalogue of 20 billion galaxies, it will also create a movie of variable astrophysical sources, detecting 10 million of these every single night.
The Vera C. Rubin Observatory, built in Chile, will undertake the deepest, widest, fastest survey of the sky. Credit: Rubin Observatory/NSF/AURA/B. Quint
I was fortunate enough to visit MeerKAT in 2018. Credit: Michelle Lochner.
On the other side of the spectrum and closer to home, we have the world-leading radio telescope MeerKAT, built in South Africa by South Africans, which is a pathfinder telescope for the world’s biggest telescope, the Square Kilometre Array (SKA). SKA-Mid, now under construction in the Karoo, will produce 62 exabytes of data per year and the data from a single MeerKAT observation will easily fill a laptop hard drive. MeerKAT has repeatedly demonstrated how new technology can uncover a plethora of new phenomena due to its high sensitivity and resolution. This deluge of data has necessitated a radical shift in analysis techniques employed by astronomers. Machine learning algorithms have rapidly been adopted into the processing pipelines of many world-leading telescopes. However, one key question remains unanswered: amongst 20 billion galaxies and 62 exabytes of data, how will we make new, unexpected scientific discoveries? This is the research question driving me and my group.
Artist's impression of the Square Kilometre Array, under construction in both South Africa and Australia, which will be the largest radio telescope ever built. Credit: Square Kilometre Array Observatory
Without enough astronomers in the world to manually inspect every galaxy in the LSST and SKA datasets, our only recourse is to turn to the incredible discovery power of artificial intelligence. The astronomy community depends on these crucial methods to fully realise the scientific potential of the SKA and LSST. Research programmes in machine learning will ensure that these observatories achieve their transformative scientific goals, unlocking new discoveries that would otherwise remain hidden in the data deluge. My research spans anomaly detection, transient classification, and observing strategy optimisation, all unified by the goal of maximising scientific discovery with these new surveys. You can also visit my publications page for a list of publications and the student supervision page to meet my research group.
Page updated
Google Sites
Report abuse