Re:A visual proof that neural nets can compute any function
by grafs50 on Sunday March 06, 2022 @03:51PM
(#30581141)
With the top comments (now) all talking about the fact that this universal approximation theorem doesn't really have much impact in the real world. I wonder, is this interesting outside of theory? Has this motivated any techniques that have created (or may create) real-world, empirical results? Could it even?
[ Parent
| on HN ]
Re:A visual proof that neural nets can compute any function
by mr_toad on Sunday March 06, 2022 @07:24PM
(#30582689)
Plenty of problems in the area of computation have little impact on the real world, but they contribute to our fundamental understanding of ‘understanding’. The Entscheidungsproblem is probably the most significant of these.