The Future of AI and Deep Learning Models

5min Snip

00:00

Play full episode

Summary

Transcript

Episode notes

The exact way that we use deep learning and AI in astronomy and cosmology is evolving incredibly fast right now. One thing you can imagine doing is a very detailed simulation of how does a black hole eat some gas, for example. And then get a deep learning model or something like that to try and predict based on those simulations. Now the question is how do you start getting a eyes that not only make predictions but explain them.

It's somewhat amazing that cosmology, the study of the universe as a whole, can make any progress at all. But it has, especially so in recent decades. Partly that's because nature has been kind to us in some ways: the universe is quite a simple place on large scales and at early times. Another reason is a leap forward in the data we have collected, and in the growing use of a powerful tool: computer simulations. I talk with cosmologist Andrew Pontzen on what we know about the universe, and how simulations have helped us figure it out. We also touch on hot topics in cosmology (early galaxies discovered by JWST) as well as philosophical issues (are simulations data or theory?).

Support Mindscape on Patreon.

Blog post with transcript: https://www.preposterousuniverse.com/podcast/2023/06/19/240-andrew-pontz…and-the-universe/

Andrew Pontzen received his Ph.D. in astronomy from the University of Cambridge. He is currently Professor of Cosmology at University College London. In addition to his research in cosmology, he frequently writes popular articles and appears in science documentaries. His new book is The Universe in a Box: Simulations and the Quest to Code the Cosmos.

See Privacy Policy at https://art19.com/privacy and California Privacy Notice at https://art19.com/privacy#do-not-sell-my-info.