Skip to content

Comment on Theoretical Motivations for Deep Learningparent

Comments

Because I've seen lots of other handwriting, even if in another language. I have very strong priors.

The problem is that a computer comes in without knowing anything about tangential phenomenon. So it needs lots of data to catch up to me and my years of forming associative connections about other handwriting I've seen.

If I showed you alien (ie not human) handwritten samples, you'd probably stuggle too.

AboutSource Built by g1lg1l

Hackerly is an independent reader for Hacker News, built on the public HN API. Not affiliated with Y Combinator.