[jinfiesto]

> It’s definitely hard to explain an idea like SVM, it’s applications, and how it works/what it does without a background in some linear algebra.

I don't actually think this is the case. The basic idea is that you can represent data as points in n-dimensional space and draw decision boundaries in that space. I think most people should at least be able to understand this geometrically for n=2/3 and then accept that it possibly extrapolates to n > 3.

[bashinator]

What is a "decision boundary?" What does it mean to represent data as a point in space? Why would the number of dimensions of space matter? How does any of that relate to AI or machine learning?

[jammygit]

Those details matter a lot, you're right. However, pictures help a lot imho, and I bet my parents could get the jist of this:

https://qph.fs.quoracdn.net/main-qimg-968b9df8608f76d41586a0...

[bashinator]

Is that teaching a computer to distinguish between males and females based on height and weight?

[jinfiesto]

More or less, yes.

[jdietrich]

>most people should at least be able to understand this geometrically

Most people can't reliably read a bus timetable, calculate a 10% tip, or multiply 537 by 12. A concept like SVM is absolute voodoo to the overwhelming majority of the population and always will be, no matter how you try and explain it.

https://nces.ed.gov/naal/sample.asp

[bigchewy]

huh?