Toutatis

For this challenge, we invite you to become "virtual contributors" to the Asteroid Grand Challenge and develop a hypothetical method, concept note or simple prototype that demonstrates how Machine Learning could be used to help us avoid the same fate as the dinosaurs.

Given the volume of deep space imagery gathered by telescopes, astronomers and space enthusiasts all around the world, the task of detecting Near Earth Objects, including potentially hazardous asteroids, become more and more tough.

Recently, several advances in the field of machine learning and particularly in image classification have been made boosting the number of research projects on new applications, for example, identification of cancerous cells, context classification or NEO identification.

Our proposal is to create a massively distributed Deep Learning classification system. To achieve this goal, we have outlined a processing architecture that takes advantage of the computation capabilities of personal computers (and maybe mobile devices) throughout the globe, all of them coordinated by a central server located in NASA headquarters.

Using a server, written in Python, and serving a Tensor Flow instance, all this clients will be able to to use, share and help in deep space image analysis. On the other hand, clients will use a browser extension (Chrome for our proof of concept) that will allow them to send their images (for astro-enthusiasts) or just observe images sent by other users all in exchange of some computation time.

A final thought, imagine you use this service for your own deep space images, wouldn't be great to know instantaneously whether you've discovered a new asteroid, and what's more, to get to name it after you?