Astronomers at the University of Hertfordshire in Hatfield, UK, have taught a machine to automatically recognize and classify galaxies in astronomical images using an unsupervised learning algorithm known as Growing Neural Gas.
This Hubble image shows the cluster of galaxies MACS0416.1-2403. Bright yellow ‘elliptical’ galaxies can be seen, surrounded by numerous blue spiral and amorphous galaxies. Gravitational arcs can also be seen. This image forms the test data that the machine learning algorithm is applied to, having not previously seen the image. Image credit: NASA / ESA / J. Geach / A. Hocking.
The scientists, co-led by Dr James Geach and Alex Hocking, demonstrated their algorithm using data from the Hubble Space Telescope ‘Frontier Fields’ – images of galactic clusters containing a mixture of galaxy types that would easily be recognized and classified by a professional astronomer.
“By training the algorithm using one field (Abell 2744) and applying the result to another (MACS 0416.1-2403), we show how the algorithm can cleanly separate image features that a human would associate with early and late type galaxies,” the scientists wrote in a paper submitted for publication in the Monthly Notices of the Royal Astronomical Society (arXiv.org preprint).
“The important thing about our algorithm is that we have not told the machine what to look for in the images, but instead taught it how to see,” Mr Hocking said.
Image showing the MACS0416.1-2403 cluster, highlighting parts of the image that the algorithm has identified as ‘star-forming’ galaxies. Image credit: NASA / ESA / J. Geach / A. Hocking.
“A human looking at these images can intuitively pick out and instinctively classify different types of object without being given any additional information. We have taught a machine to do the same thing,” Dr Geach said.
He added: “our aim is to deploy this tool on the next generation of giant imaging surveys where no human, or even group of humans, could closely inspect every piece of data. But this algorithm has a huge number of applications far beyond astronomy, and investigating these applications will be our next step.”
The astronomers are now looking for collaborators, making good use of the technique in applications like medicine, where it could for example help doctors to spot tumors, and in security, to find suspicious items in airport scans.
_____
Alex Hocking et al. 2015. Teaching a machine to see: unsupervised image segmentation and categorisation using growing neural gas and hierarchical clustering. MNRAS, submitted for publication; arXiv: 1507.01589
