Exploring the cosmic puzzles with gravitational lensing
and machine learning
The measured expansion rate of the universe is faster than
expected from the standard model of cosmology.
Independent of this ‘cosmic expansion puzzle’, the growth
of large-scale structure also appears slower than expected. I
will summarize the status of these cosmic puzzles and
possible resolutions, focusing on measurements of weak
gravitational lensing. The question of how much information
is contained in galaxy surveys is still wide open. I will
describe the state of play using both analytical statistics
that theorists love and deep learning approaches
that evoke mixed feelings. I will also introduce
transformers, a promising
new
architecture
for
astronomical data (as well as chatGPT).
Exploring the Cosmic Puzzles with Gravitational Lensing and Machine Learning