Einstein's Theory of General Relativity predicts a rare phenomenon called "gravitational lensing", where a massive body causes curvature in spacetime and acts as a giant cosmic lens. This work explores the application of a machine learning technique, convolutional neural networks, to finding images of gravitational lenses amongst large databases of astronomical data. Due to their rarity, we chose to train the machine using simulations. Using this technique we were able to make over five hundred new discoveries.
History
Thesis type
Thesis (PhD by publication)
Thesis note
Presented in fulfillment of the requirements of the degree of Doctor of Philosophy, Faculty of Science, Engineering and Technology, Swinburne University of Technology, 2019