DPRG
DPRG List  



[DPRG] Stanford Online Free Engineering Courses

Subject: [DPRG] Stanford Online Free Engineering Courses
From: Chris Jang christopher.jang at yahoo.com
Date: Mon Jul 20 22:31:41 CDT 2009

> Machine Learning CS229 
> Convex Optimization I EE364A 
> Convex Optimization II EE364B 

Hello, the single most illuminating tutorial I've seen on machine learning is the one by Christopher Burges about support vector machines. It was after reading this that machine learning made sense for me. The discussion of model capacity and structural risk is good and applies to machine learning generally. This is a great paper.

http://research.microsoft.com/en-us/um/people/cburges/papers/SVMTutorial.pdf

SVM is only one approach. However, if you know only one way, I believe this is the one to know - the linear classifier (with kernel trick). It is probably the way most engineers prefer to see things. It's close to vector spaces and numbers. Statisticians see things in terms of abstract distributions which I find difficult to follow.

If you wonder why you didn't see any of this in school, one reason is that much of the theory was only invented in the mid-1990s. It's relatively new.


      

More information about the DPRG mailing list