AI for Dark Matter  & Gravitational Lensing

Jenny Wagner (Academia Sinica Institute of Astronomy and Astrophysics/Helsinki Institute of Physics, University Helsinki)

Can Artificial Intelligence tell us how much dark matter really matters?

Strong gravitational lensing has advanced as a standard probe to map mass densities of cosmic structures or to try and infer parameters of the cosmological concordance model, like the Hubble Constant.

Recently, the heuristic, simulation-based mass density models that used to be fitted to the strong-lensing data have been replaced with those from AI/ML techniques.

Do the latter increase the precision and accuracy at which we can infer the distribution and nature of dark matter from strong-lensing observations?

To answer these questions, this talk will introduce a new way of interpreting strong-lensing observations, separating the local properties of the light-deflecting mass densities that are directly constrained by the data from the model assumptions added by fitting mass density models to the data.

References

[1] A model-independent characterisation of strong gravitational lensing by observables (arXiv:1906.05285 [astro-ph.CO] review paper of the core method)
[2] ‘Double’ Galaxy Mystifies Hubble Astronomers (press release for a unique example analysis)
[3] The Case Against the Cosmological Principle (and/or FLRW) (integration into the community of observational cosmology)
[4] Chaos im Kosmos • Dunkle Energie Dunkle Materie (outreach
contribution in German language)

---

The colloquium will be streamed but not recorded.
Zoom link: https://uni-frankfurt.zoom.us/j/2848286010?pwd=VmtCY1RCc1hpVStKd0RibFBpc1IzZz09
Meeting ID: 284 828 6010
Password: 068695