These are the slides for a short talk with the title “Deep learning domain translation between mock observations and hydrodynamical simulations” at the Swiss SKA Days 2023 conference in Zurich. It presents a diagnosis for the reason we had a hard time training our generative image-to-image translation models on projection images of IllustrisTNG galaxies.
In order to make the presentation easier to follow (and for archival),
I used the reveal.js framework to turn my presentation slides into a
HTML/javascript website.