The Eclipse AI, Cloud, & Edge (AICE) Working Group’s OpenLab — a vendor neutral experimental test environment — is working on a use case with the potential to transform more than 10 million lives: The AURA Demonstrator.
AURA is a French non-profit organization developing a patch for people with drug-resistant epilepsy to detect seizures in advance and warn patients.
The team’s first goal was to train their model on the Temple University Hospital dataset. But the complexity of the model’s setup, combined with long execution times, made it difficult to train.
To tackle the full data set, the team had to:
- Make their model more sustainable, reliable, and reproducible
- Simplify its setup
- Improve the visualization import process
Check out the full article to find out how they overcame these challenges and are taking their model even further.
To learn more about AURA and the AICE Working Group, and get involved: