I read about application of neural networks to fluid dynamics. It ended up being... | Hacker News

Hacker Newsnew | past | comments | ask | show | jobs | submit

scotty79 on Feb 1, 2017 | parent | context | favorite | on: The Convex Geometry of Inverse Problems [video]

I read about application of neural networks to fluid dynamics. It ended up being faster than usual approaches. At least some matemathical solutions might eventually be superseeded by pretrained nn at least in some contexts.

SmooL on Feb 1, 2017 [–]

I think the key was that the neural network didn't compute the fluid model accurately so much as a compute a simulation that looked accurate to humans

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact