Rising Stars Hackathon: Contributing to current R&D for the largest astronomical telescopes

Rising Stars Hackathon: Contributing to current R&D for the largest astronomical telescopes

The Rising Stars sponsored AI Hackathon "Optimizing AI-driven workflows within a mission-critical cyber-physical system" brought together participants eager to enhance a denoising module as part of comprehensive AO RTC pipeline. The Hackathon began with an introduction to the context of the challenge. Hackathon leaders Damien Gratadour from coordinator Observatoire de Paris-CNRS and Bartomeu Pou Mulet from the Barcelona Supercomputing Center presented the challenges faced by the largest ground-based telescopes that are able to provide angular resolution and a collecting area to perform the measurements required to detect fainter exoplanets but that struggle to overcome optical distortion induced by atmospheric turbulence.

The presenters introduced the current deep learning techniques being used to pre-process images for denoising purposes. The proposed architecture consists of convolution layers together with maxpool operations. While the initial work on the denoising autoencoder and convolution autoencoder consisted of learning more robust representations in the hidden layers and use those as an initialisation point for other networks in tasks such as image classification, it can be applied for denoising purposes.

 

Bartomeu hackathon
Bartomeu Pou Mulet from the Barcelona Supercomputing Center presents the Hackathon challenge

 

feedback
Attendees get feedback on their solutions

 

The hackathon proposed enhancing the denoising module as part of the comprhenesive AO real-time controller pipeline. It relied on the use of COMPASS (COMputing Platform of Adaptive opticS System), an end-to end simulation tool for AO system, allowing the users to generate syntehtic but realistic AO loop data as if they were operating an actual AO system.

This event provided an opportunity for attendes to: 1) actively contribute to current research and development for the largest astronomical telescopes, aiming to evidence new exoplanets and habitable worlds, 2) learn more on applying AI/DNNs to real world use cases, and 3) to have fun with advanced technologies for simulating complex cyber-physical systems.

This one day Hackathon was held at the Barcelona Supercomputing Center on 11 June 2024 and was co-located within the International Conference on Reliable Software Technologies (AEiC 2024). More information about the conference can be found on this page.