4
1

I'm looking for a brief and concise introductory paper/book that explains the main and the most used optimization techniques in machine learning. Some thing equivalent to "Gibbs Sampling for the Uninitiated" and "Bayesian Inference with Tears", but for optimization.

This question is marked "community wiki".

asked Sep 26 '12 at 20:23

theDever's gravatar image

theDever
16125

wikified Sep 27 '12 at 19:23

I'm going to answer my own question and add some useful resources that I found in the last day.

(Sep 28 '12 at 17:42) theDever

5 Answers:

Just read openopt docs http://openopt.org/Problems

answered Sep 26 '12 at 22:30

marshallp's gravatar image

marshallp
8391016

How about "Optimization for Machine Learning" from Sra, Nowozin and Wright? There is also "Foundations of Machine Learning" which apparently also talks about optimization. I haven't read it yet, though.

answered Sep 27 '12 at 02:44

Andreas%20Mueller's gravatar image

Andreas Mueller
2686185893

The Sra, Nowozin, and Wright book is great for someone interested in researching optimization for machine learning and wanting to get a peek at some deep places where interesting ideas lie. It is not really an introduction, and it assumes a lot of familiarity with the topic.

(Sep 27 '12 at 09:30) Alexandre Passos ♦

Mark Schmidt wrote a very nice paper listing the most common methods used in Machine Learning and the links between them. I strongly encourage you, and anyone else, to read it.

Notes on Big-N problems

answered Sep 27 '12 at 02:53

Nicolas%20Le%20Roux's gravatar image

Nicolas Le Roux
7652912

edited Sep 27 '12 at 02:54

Complementing Nick's answer, it might help to hear these algorithms being explained out loud, and for that I strongly recommend Stephen Wright's nips tutorial on optimization in ML.

answered Sep 27 '12 at 09:38

Alexandre%20Passos's gravatar image

Alexandre Passos ♦
2554154278421

Smola's machine learning book has a good self-contained chapter for optimization.

This answer is marked "community wiki".

answered Sep 28 '12 at 17:52

theDever's gravatar image

theDever
16125

Your answer
toggle preview

powered by OSQA

User submitted content is under Creative Commons: Attribution - Share Alike; Other things copyright (C) 2010, MetaOptimize LLC.