CIS 520: Machine Learning


CIS 520: Machine Learning Fall 2007
	COURSE DESCRIPTION: CIS 520 provides a fundamental introduction to the mathematics and practice of machine learning. Probabilistic and statistical methods for prediction and clustering are covered in depth. Topics covered include linear and logistic regression, support vector machines, neural networks, EM, k-means, graphical models, dimensionality reduction, and reinforcement learning. For details, see the course schedule. CIS 520/001 is for CIS PhD students; CIS 520/002 is for everyone else. AUDIENCE: The course is aimed broadly at advanced undergraduates and beginning graduate students in computer science, electrical engineering, mathematics, physics, and statistics. Undergraduates who meet the prerequisites are particularly encouraged to enroll, as are students from other departments. PREREQUISITES: Multivariable calculus Linear algebra Elementary probability Programming experience in a language such as C, Java, or Matlab INSTRUCTOR and TEACHING ASSISTANTS Prof. Lyle Ungar ungar@cis.upenn.edu Office Hours: Mondays 3:30-5:00 p.m in 504 Levine Jinsong Tan (TA) jinsong@seas.upenn.edu Office Hours: Tues 11:30-1:30, in GRW 571 Cheryl Hickey (staff) 502 Levine For picking up or dropping of homeworks. COURSE LOCATION AND TIME: Tuesdays and Thursdays 9:00 - 10:30AM Heilmeier Hall, Towne Building (ground floor) TEXTBOOKS: There is no required textbook, however there are many readings (many of which can be accessed via the course schedule) and a highly recommended book: The Elements of Statistical Learning: Data Mining, Inference, and Prediction by Trevor Hastie, Robert Tibshirani, Jerome Friedman We will be using Matlab for the course. It can be purchased at the bookstore ($99) purchased online from Mathworks used on the computers in SEAS and a variety of other places on campus The precise version does not matter You might well be able to get away using an open source version such as octave or scilab (see their descriptions), but I have not tried any of the code there, and so make no guarantees. On reserve in the Towne library: The Elements of Statistical Learning: Data Mining, Inference, and Prediction by Trevor Hastie, Robert Tibshirani, Jerome Friedman Artificial Intelligence: A Modern Approach (Second Edition) by Stuart Russell and Peter Norvig MATLAB guide by Desmond J. Higham and Nicholas J. Higham The key readings have been scanned and are available from the schedule page. Further references are avaliable on the ML wiki COURSE REQUIREMENTS AND GRADING: There will be (roughly) weekly homeworks, which will constitute 50% of your grade, a midterm (20%) and a final (30%) Homeworks are posted on the web page. They are generally due Thursdays at class time. To protect my sanity, late homeworks will be penalized: Late Homework Penalty in class on the next class after HW is due: 50% penalty in class on the second class after HW is due: 75% penalty Homework is currently submitted in hardcopy only. Later ones will be submitted electronically as well. We will also use pair programming which some argue is an effective teaching technique. Information re: CIS graduate students who can tutor other CIS grad students can be found at www.cis.upenn.edu/grad/tutorial.shtml Other courses of potential interest Ben Taskar is teaching CIS 521 Artificial Intelligence; there should be little enough overlap that students can take both courses without fear of undue repetition. see also the list of machine learning courses at Penn return home

CIS 520: Machine Learning
Fall 2007

COURSE DESCRIPTION:

CIS 520 provides a fundamental introduction to the mathematics and practice of machine learning. Probabilistic and statistical methods for prediction and clustering are covered in depth. Topics covered include linear and logistic regression, support vector machines, neural networks, EM, k-means, graphical models, dimensionality reduction, and reinforcement learning.

For details, see the course schedule.

CIS 520/001 is for CIS PhD students; CIS 520/002 is for everyone else.

AUDIENCE:

The course is aimed broadly at advanced undergraduates and beginning graduate students in computer science, electrical engineering, mathematics, physics, and statistics. Undergraduates who meet the prerequisites are particularly encouraged to enroll, as are students from other departments.

PREREQUISITES:

Multivariable calculus
Linear algebra
Elementary probability
Programming experience in a language such as C, Java, or Matlab

INSTRUCTOR and TEACHING ASSISTANTS

Prof. Lyle Ungar
ungar@cis.upenn.edu
Office Hours: Mondays 3:30-5:00 p.m in 504 Levine
Jinsong Tan (TA)
jinsong@seas.upenn.edu
Office Hours: Tues 11:30-1:30, in GRW 571
Cheryl Hickey (staff) 502 Levine
For picking up or dropping of homeworks.

COURSE LOCATION AND TIME:

Tuesdays and Thursdays 9:00 - 10:30AM
Heilmeier Hall, Towne Building (ground floor)

TEXTBOOKS:

There is no required textbook, however there are many readings (many of which can be accessed via the course schedule) and a highly recommended book:

The Elements of Statistical Learning: Data Mining, Inference, and Prediction by Trevor Hastie, Robert Tibshirani, Jerome Friedman

We will be using Matlab for the course. It can be

purchased at the bookstore ($99)
purchased online from Mathworks
used on the computers in SEAS and a variety of other places on campus

The precise version does not matter

You might well be able to get away using an open source version such as octave or scilab (see their descriptions), but I have not tried any of the code there, and so make no guarantees.

On reserve in the Towne library:

The Elements of Statistical Learning: Data Mining, Inference, and
Prediction by Trevor Hastie, Robert Tibshirani, Jerome Friedman
Artificial Intelligence: A Modern Approach (Second Edition) by
Stuart Russell and Peter Norvig
MATLAB guide by Desmond J. Higham and Nicholas J. Higham

The key readings have been scanned and are available from the schedule page. Further references are avaliable on the ML wiki

COURSE REQUIREMENTS AND GRADING:

There will be (roughly) weekly homeworks, which will constitute 50% of your grade, a midterm (20%) and a final (30%) Homeworks are posted on the web page. They are generally due Thursdays at class time. To protect my sanity, late homeworks will be penalized:

Late Homework Penalty

in class on the next class after HW is due: 50% penalty
in class on the second class after HW is due: 75% penalty

Homework is currently submitted in hardcopy only. Later ones will be submitted electronically as well.

We will also use pair programming which some argue is an effective teaching technique.

Information re: CIS graduate students who can tutor other CIS grad students can be found at www.cis.upenn.edu/grad/tutorial.shtml

Other courses of potential interest

Ben Taskar is teaching CIS 521 Artificial Intelligence; there should be little enough overlap that students can take both courses without fear of undue repetition.
see also the list of machine learning courses at Penn

return home