8–13 Jun 2025
America/Winnipeg timezone
Welcome to the 2025 CAP Congress Program website! / Bienvenue au siteweb du programme du Congrès de l'ACP 2025!

Learning to See: Applying Inverse Recurrent Inference Machines to See through Refractive Scattering

9 Jun 2025, 15:15
15m
Oral Competition (Undergraduate Student) / Compétition orale (Étudiant(e) du 1er cycle) Atmospheric and Space Physics / Physique atmosphérique et spatiale (DASP/DPAE) (DASP) M2-2 General Topics: Planetary Physics and Astronomy | Thèmes généraux: Physique planétaire et astronomie (DPAE)

Speaker

Mr Arvin Kouroshnia (University of Waterloo)

Description

The Event Horizon Telescope (EHT) has captured images resolving the horizon of Sagittarius A (Sgr A∗), unveiling a new window into understanding strong gravity and cosmology. Scattering from the turbulent plasma of the interstellar medium distort the appearance of Sgr A, limiting the practical angular resolution of Sgr A. We utilized a recurrent neural network trained on samples ignorant to General Relativistic Magnetohydrodynamics (GRMHD), due to the non-birefringence of Sgr A, we were able to leverage the use of multiple polarizations in the scattering mitigation scheme, demonstrating that it is possible to mitigate scattering well below the instrumental resolution of EHT, 24 µas.

Keyword-1 Super massive black holes
Keyword-2 Convolutional Neural Networks
Keyword-3 Galactic center

Author

Mr Arvin Kouroshnia (University of Waterloo)

Co-authors

Kenny Nguyen (University of Waterloo) Dr Chunchong Ni (University of Waterloo) Mr Ali SaraerToosi (Vector Institute) Prof. Avery Broderick (University of Waterloo)

Presentation materials