Hacker News new | past | comments | ask | show | jobs | submit login

A pet peeve of mine is that differentiable programming is co-opted almost entirely by deep learning + neural networks. The idea of differentiable programming is much bigger than SGD, and in fact neural networks are typically a simple program to differentiate. Full differentiable programming requires solving much more involved problems around control flow than just implementing numerical forward/reverse mode for math operations with well defined and understood gradients.



Join us for AI Startup School this June 16-17 in San Francisco!

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

Search: