How do people first learn to spot patterns?

How people separate things into objects that look the same and objects that look different? How do people identify which particular characteristics are important for discriminating between different types of objects? In this project we will consider machine learning and its application to the observations of transients, that are now being produced by the first generation of survey telescopes and will become progressively important in the next decade with the advent of the Large Synoptic Survey Telescope (LSST) and the European-Extremely Large Telescope (E-ELT). At Sheffield, we have a head start in the field, being a partner in GOTO (Gravitational Wave Optical Transient Observatory), which will detect a large number of brand new types of transients. This project is primarily concerned with developing unsupervised learning algorithms to mimic the way in which humans learn when given a new set of information they’ve never seen before. In the case of supernovae, the classification scheme has been developed over the last 80 years, with new branches being proposed on a regular basis. This project aims to answer the question: what is the most natural way to differentiate and classify different types of transient?
This project also comes with the possibility to collaborate with computer scientists based both in Sheffield and Thailand.


Science Graduate School:
As a PhD student in one of the science departments at the University of Sheffield, you’ll be part of the Science Graduate School – a community of postgraduate researchers working across biology, chemistry, physics, mathematics and psychology. You’ll get access to training opportunities designed to support your career development by helping you gain professional skills that are essential in all areas of science. You’ll be able to learn how to recognise good research and research behaviour, improve your communication abilities and experience technologies that are used in academia, industry and many related careers. Visit www.sheffield.ac.uk/sgs to learn more.