Differentiable approximations to the min and max operators

Posted by aidanrocke on Thursday February 13, 2020 @05:29PM
from the github.com dept.

( 2 comments | 3 points | on HN )

Re:Differentiable approximations to the min and max operators by heartbeats on Monday February 17, 2020 @03:47PM (#22350579)

How does this stack up to the softmax[0] function, log(exp(n_1) + exp(n_2) + ... + exp(n_i))? Are they analytically equivalent?

[ Parent | on HN ]

Re:Differentiable approximations to the min and max operators by aidanrocke on Thursday February 13, 2020 @05:30PM (#22322076)

tl;dr

1. Within the context of optimisation, differentiable approximations of the min and max operators on R^n are very useful.

2. However, in order for these approximations to be useful they must also be numerically stable.

[ Parent | on HN ]