- Handout 0 -- About the course
-
Lecture 1 -- Statistical prediction -- an overview (11 pages)
-
Lecture 2 -- Optimization in statistics (9 pages)
- Lecture 3-part I -- Unconstrained optimization (part I, 11 pages)
- Lecture 3-part II -- Unconstrained optimization (part II, 17 pages)
- The perceptron algorithm lecture notes by Avrim Blum
- Lecture 4 -- Boosting and other methods for averaging classifiers (26 pages) 2/7/12 Updated with extensions to Boosting pages 23-26. typos fixed on pages 13-14
- A Boosting Approach to Machine Learning by Rob Schapire (an introductory article)
-
"AdaBoost is consistent" by Peter Bartlett and Mikhail Traskin
- ''Convexity, classification, and risk bounds'' by Peter
L. Bartlett, Michael I. Jordan, Jon D. McAuliffe, in Journal of the
American Statistical Association 101(473), March 2006: 138-156.
- "Evidence Contrary to the Statistical View of Boosting" by David Mease, Abraham Wyner, JMLR 2008
(a set of experiments highlighting unanswered questions about
boosting)
- Large margins using the perceptron algorithm" by Yoav Freund and Rob Schapire, (The Voted Perceptron Algorithm) in Machine Learning 37(3):277-296, 1999
- Maryam Fazel's slides on convex optimization. Look at the Convex sets and Convex functions lectures.
- Lecture 5 Convex sets (a stump. Small additions to Boyd Chapter 2)
- Lecture 6 Convex functions (a very small stump. Small additions to Boyd Chapter 3)
- Lecture 7 Convex conjugate, Bregman divergences, exponetial family models (7 pages)
- Generalized linear models: A unified approach by Jeff Gill
"Clustering with Bregman divergences" by Banerjee et al. (45 pages; of main interest sections 2, 3 and appendix A)
- Lecture 8 Maximum entropy (20 pages)
- Lecture 9 Support Vector Machines (16 pages)
-
"A tutorial on support vector machines for pattern recognition" by Chris Bur
gess (appeared in Data Mining and Knowledge Discovery, 2, 121-167, 1998.
- www.kernel-machines.org The SVM and Kerne
l machines resource page (not up to date)
- E. J. Candes and M. Wakin. An introduction to compressive sampling IEEE Signal Processing Magazine, March 2008 21-30.
|