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

If it works, it won't go unnoticed.


Ideally.

According to fast.ai there's this (very simple) technique to automate optimizing the learning rate parameter for deep neural networks, that went unnoticed for a couple years, because it was discovered by a no-name researcher, while all the big wig researchers were spending tons of manual hours finding the optimal learning rate.




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

Search: