First Order Methods Subgradient Projected

Hope you will enjoy this video. I know my voiceover is lacking some emotion but i will try my best to improve that for my next video.

Gradient Descent (unconstrained optima of differentiable function)

In this video we start looking at non-smooth optimization. We take a look at the

Author: Hongyi Zhang, Suvrit Sra.

Note: sound cuts out for last 20 minutes or so, sorry!

We formulate the

lecture4 04 subgradient algorithm

Of g zero right that's all non-smooth

Okay so this is our

Okay so these are kind of two classic results on the

Neither the lasso nor the SVM objective function is differentiable, and we had to do some work for each to optimize with ...

This video derives the oracle complexity of the

I recommend you watch in 1.25x or 1.5x to not waste time.

So we can define subgrading descent like before we just replace the gradient by

First

proof of gradient descent.

Now at

In this video, we look at the various properties of

Subgradients

Subgradient methods