10-702: Statistical Machine Learning
Syllabus, Spring 2010
http:/www.cs.cmu.edu/~10702
Statistical Machine Learning is a second graduate level course in machine learning, assuming
students have taken Machine Learning (10-701) and Intermediate Statistics (

10-702 Statistical Machine Learning: Assignment 3
Due Friday, February 19
Hand in to Sharon Cavlovich, GHC (Gates Hillman Center) 8215 by 3:00. Use R for all
numerical computations.
1. Undirected graphical models
(a) Let X = (X1 , . . . , Xd )T where each

10-702 Statistical Machine Learning: Assignment 2
Due Friday, February 5
Hand in to Sharon Cavlovich, GHC (Gates Hillman Center) 8215 by 3:00. Use R for all
numerical computations.
1. (Convex sets and functions)
(a) For x Rn dene the p norm
1/p
n
x
p
=
j=

10-702 Statistical Machine Learning: Assignment 2 Solutions
1. Convex sets and functions
(a)
First, we prove that if p 1, then C is convex set. Let g(x) = |x|, where x
R. g(x) is convex function. Let h(y) = y p , where y > 0 and p 1. Since
h (y) = py p1

10-702 Statistical Machine Learning: Assignment 3 Solution
1. (a)
d
d
log p(x) = 0 +
j Xj +
j=1
d
jk Xj Xk X + +
jk Xj Xk + +
(1)
j<k<
j<k
Since A = 0 whenever cfw_1, 2 A, all the terms in the above linear right hand side can be partitioned
into 3 parts :

10-702 Statistical Machine Learning: Practice Midterm Exam
Submit solutions to any four of the following seven problems. Clearly indicate
which problems you are submitting solutions for. Write your answers in the space provided;
additional sheets are atta

10-702 Statistical Machine Learning: Assignment 1 Solutions
1. Review of Maximum Likelihood.
Let X1 , . . . , Xn be a random sample where Xi cfw_1, 2, . . . , k. Let [0, 1] and suppose
that P(Xi = 1) = and P(Xi = j) = for j > 1 where = (1 )/(k 1).
(a) Fin

Special Topics Lecture
ROBOT VISION
Peter Hansen
Spring 2011
Outline
Cameras/lenses
Overview of some applications
and image processing methods
Stereo vision
Image formation
Scene reconstruction
Visual odometry and mapping
Examples from LNG pipe inspection

Brett Browning
and
M. Bernardine Dias
Spring 2011
Additional
ideas and teams for final
project due today
Any questions?
Slide 2
Slide 4
Introduction to navigation
Reactive navigation
References
Next
Goal
Rough
Slide 5
Our
starting definition:
How to mo

Brett Browning
and
M. Bernardine Dias
Spring 2011
Guest
lecture on Sunday
Robot Vision
Reminder:
Lab #1 due in 1 week from today
Slide 2
Slide 4
Final Project Assignment
Lidar recap
Processing points
Occupancy grids
Slide 5
Light
Detection and Rangin

10-702 Statistical Machine Learning: Assignment 1
Due Friday, January 22
Hand in to Sharon Cavlovich, GHC (Gates Hillman Center) 8215 by 3:00. Use R for all
numerical computations.
1. (Review of Maximum Likelihood.) Let X1 , . . . , Xn be a random sample