2016 Problem 2 subgradients

This forum is inactive. Browsing/searching possible.

CS-433 Machine Learning

Connect your moderator Slack workspace to receive post notifications:

Hi, I am confused about the solution for this problem.

What is the h() function and what is the \Nabla y_n ? Thank you.

7 Jan '21 ·

Hi,

To my understanding there are 2 small typos in the solution:

h(y_n - f(w, x_n)) is actually g(y_n - f(w, x_n))
\(nabla\) y_n - f(w, x_n) is \(nabla\) (y_n - f(w, x_n))
i.e. there is a () missing

There is a detailed explanation of this exercise in lab 2 btw.

7 Jan '21 ·

Page 1 of 1

How to style: strictly use the or click here. E.g., \(\alpha + \beta\) gives (inline) \(\alpha + \beta\). No \(\LaTeX\) preview (yet).