16. Learning: Support Vector Machines

It's first time i understand abot SVM

what an amazing lecture and professor. This is what its all about. Some people are just blessed lecturers and he is for sure one of those people.

I have a question. At the start of the lecture, the professor says that we can imagine W to be a normal vector of any length and that the distance of a point from the hyperplane is the projection of any point onto W. But how can the dot product of WX be equal to the length of the projection of X onto W if W is any length? This can only work if W is a unit vector unless I'm mistaken. WX = ||W||*||X||*cos(theta) and the projection length of X onto W is ||X||*cos(theta). So WX can only equal ||X||*cos(theta) if ||W|| = 1. Otherwise ||X||*cos(theta) = WX/||W||.

I've been watching a lot of MIT, Stanford, Harvard, Princeton lectures, but this... This was phenomenal, hands down the best lecture I've ever seen.

Rest in Peace Prof

What the fuck with these people just come in the middle of the class..

Thank you for the great video. How do you know which X_+ and x_- will have f(x)=1 "In the gutter"? Is this something that is recursively calculated in practice?

I love Prof Winston! Thank you Prof!

All my respects to the late great Prof. Winston, by far my favorite AI teacher of all time.

SU'PORtts' - Already Tribal invent. Artz = kaldean teacher! Hence; he's racist - evoluchist(demonic women) - Tribal LA'(law). Only Tribes: Own any land - products therin.

❤️❤️❤️❤️❤️❤️❤️

It's the Magic Flute right here

saved me for my endsems. thank you.

How beautiful a proof can be if taught in the correct way! Brilliant!

amazing

This is how things should always be taught. Patience, deep understanding and passion to teach. I wish I would've had a professor like him as a graduate student.

dot product dont give projection we need to multipy it either by |x| or |y| to give us the projection im i right ?

genius explanation here. very simple to understand, I love digging into the underlining mathematics thats involved with these algorithms. The idea that a kernel function can transform the nonlinear inseparable data into a higher dimensional space that is linear is simple mind boggling to me. So cool of a concept and really simple to realize once you get the mathematics behind it.

RIP Patrick Winston! What an amazing teacher

🥰

Primeiro ele tem que cuidar da saúde dele esta barrigudo propenso a várias doenças.

The saddest thing is that these good professors never take their health wellness seriously. 😢 They suddenly leave us with no replacement

Merci

나한테 이게 왜 뜨냐..

бред

ita too easy as ima genius cuz 1+1is three

cool

Kernel trick, that has to be one of the most beautiful ideas I've seen (so far) in any branch of mathematics.

for the longest time I thought he was walking around with a half-eaten sandwich in his hand.

That was hands down the most amazing lecture i have ever attended online

Simply the best <3

nice,

You made my day :)

Great lecture 👏👏👏

Saw this video long time back when i started learning ML.. Years later, I am revisiting this masterpiece again 👍

coolest professor, he seems like bruce banner explaining concepts

/\