Y
Hacker News
new
|
ask
|
show
|
jobs
by
eru
1418 days ago
Lots of linear functions with \x -> max(0, x) thrown in between.
(That's literally what neural nets with relu as activation unit do.)