Thomas Unterthiner

AI Research Engineer at Helsing

Unterthiner co-created FID, one of the most widely used ways to evaluate AI-generated images, and helped develop self-normalizing neural networks, which keep information more stable as it passes through a deep model.