+ 
Parameterization of the M87* blackhole using Generative Adversarial Networks
-People: Lily , Pavlos Protopapas
-Accurate parameterisation of the M87* blackhole is - challenging as the simulations are computationally expensive resulting in - sparse training datasets. In order to increase the size of the training grid, - we propose a data augmentation methodology based - on Conditional Progressive Generative Adversarial Networks to generate a - variety of synthetic black hole images based on its - spin and electron distribution parameters.
+Reconstructing M87* Black Hole Images using Multi-Conditional Diffusion Models
+People: Yuqing Pan et al.
+Diffusion models have become popular neural network architectures + due to their success in various computer vision applications. This paper bridges the gap between advanced + deep learning techniques and astronomical imagery. We present a multi-conditional diffusion model, + InstructPix2Pix-M87, trained on general relativistic magnetohydrodynamic (GRMHD) simulated images of the + M87* black hole to enhance the quality of observations from the Event Horizon Telescope (EHT). + Our approach demonstrates improved de-blurring and parameter inference capabilities for M87*, + laying the groundwork for further studies in astrophysics..
+ + Paper + + Code + +
+ 