Domain adaptation for cross-domain use of deep learning models

In astronomy, neural networks are often trained on simulated data with the prospect of being applied to real observations. Unfortunately, simply training a deep neural network on images from one domain does not guarantee satisfactory performance on new images from a different domain. I am interested in the study of different domain adaptation techniques to improve the performance of deep learning models in astronomy and also to better understand which features are shared across different domains to harness all available data for discovery. Some of the methods that can be used include Maximum Mean Discrepancy (MMD) and adversarial training with Domain Adversarial Neural Networks (DANN). I studied these methods for a test case of the classification of distant galaxy mergers from the Illustris-1 simulation, where the two domains presented differ only due to the inclusion of observational noise. Both MMD and adversarial training greatly improve the performance of the classifier on the target domain when compared to conventional machine learning algorithms, thereby demonstrating great promise for their use in astronomy.