Gradient Descent for Support Vector Machines and Subgradients

[@tolgauye7924]

Hocam allah razi olsun. Bir tane subgradient anlatan duzgun video bulana kadar gotumuz catladi. Tesekkur ederim

[@CamilleKaze57]

Thanks a lot, it really helped me! You're a much better teacher than mine at uni lol

[@sanjaykrish8719]

From Fourier to Z transform to SVM's... Prof, you are great.

[@tolgauye7924]

adam gibi adam barry van veen. Aday ol ,oy verelim!!!

[@marofe]

Can the subgrad result in a no descent direction?

[@赵宇川-n3q]

after we get the optimized w, how to get the optimized b?

[@morisjohn4167]

Thank you so much sir, after studying svm i wondered how actually can we find optimal W , your video gave me an answer for that.

[@faridzabihian1694]

hi what happens when we compute the gradient for one data point?does it finsih all the point or pick a random point?

[@anouarlahmdani3718]

is the function continuous in 1? 5:13

[@farhanhyder7304]

Simple and easy to understand

[@alonbegin8044]

great video! I just don't get your intuition for choosing any value you want within your constraints.. whats stooping me to choose -d*x_i instead of zero in 9:37 and maybe get different result in the loss function that can potentially dramatically change my result in for example to classify corona sick people or not

[@isaacnewton1545]

That's called subgradient NOT subderivative. The subderivative is a taken name for the dini derivative look at Rockafellar and Wets book.