University of Utah

School of Computing

Machine Learning

CS 5350/6350

Fall Semester 2012

WEB 2250 TH 12:25-1:45

Instructor: Thomas C. Henderson

Overview of Course

Course Objectives

Machine Learning (Marsland ):

is an enterprise that uses statistical methods to disentangle data using models constructed with the aid of computer science, mathematics and engineering
making computers modify or adapt their actions to perform better
making computers learn from data
making computers remember and generalize
making computers get better at some task through practice.
has a wide variety of applications:

understanding images
fraud detection
game playing
robotics behavior synthesis.

The following lists the goals for the machine learning course:

Experience with machine learning algorithms (in Matlab)
Explore machine learning in terms of the issues described above
Learn appropriate reporting techniques

Prerequisites

The prerequisite is successful completion of CS 3505 (for CS5350), CS5300 or CS6300 (for CS6350).

Course Description

We will work on the problems and solutions to machine learning using modern machine learning techniques.

Software Used to Support Class

Students will develop codes in Matlab or Python.

Required Materials

We will use:

Marsland . Machine Learning, CRC Press, 2009 (required)

Assignments

There are 2 major types of assignments (use the Lab Report Format for Problem and Project reports) as well as quizzes:

Problem Assignments: These must be completed by the due date and turned in as PDF report files.
Project (CS6350 only): Requires periodic reports on progress with related codes, images, graphs, etc.
Quizzess: Weekly (short) quizzes will be held during the semester.

Class Syllabus

The lectures will cover the text on the following schedule (may vary some during semester to accommodate progress):

Date	Topic	Material (Marsland )	Problem Assignments

August 21-23 August 28-30	Intro & Neural Networks (Matlab,Project)	Chapters 1,2,3	Assigned: A1 Assigned: CS6350 Project
September 4-6 September 11-13	Radial Basis Functions and Splines;Support Vector Machines	Chapters 4-5	Assigned: A2
September 18-20	Learning with Trees	Chapter 6	Assigned: A3
September 25-27 October 2-4	Ensemble Learning & Probability and Learning	Chapters 7-8	Assigned: Project and A4 Assigned: Project
October 9-11	Fall Break
October 16-18 October 23-25	Unsupervised Learning and Dimensionality Reduction	Chapters 9-10	Assigned: Project and A5 Assigned: Project P3
October 30 - November 1 November 6-8	Evolutionary Learning, Reinforcement Learning	Chapters 12,13	Assigned: Project and A6
November 13-15 November 20 November 27-29 December 4-6	Reinforcement Learning, MCMC	Chapters 13-14	Assigned: A7 Assigned: A8

Class Schedule and Assignments

The lectures and assignments will cover the text as we progress through the semester. Assignments will usually be handed out on Tuesday and due on a Thursday after the material is covered.

Instructor

Instructor:

Thomas C. Henderson , Professor; Anshul Joshi (TA)

E-Mail:

tch@cs.utah.edu, anshul.v.joshi@utah.edu

Phone:

801-581-3601

Fax:

801-585-3743

Office Hours (2781 WEB): By appointment.
Anshul: TH 2-4pm (and by appointment; Office 3115 MEB)

Grading Information

The grading distribution will be as follows:

Project: 30% (CS6350)
Problems: 50% (CS6350), 70% (CS5350)
In-Class Tests and Participation: 20% (CS6350), 30% (CS5350)

You are expected to make a good effort on all assignments and in-class discussion based on a careful reading of the assigned material. I will assign a grade based on how reasonable your solution is given the difficulty of the assignment, the time required, and the style and content of the solution. My goal is to look at all your work, and to assign a grade based on your participation, effort and results. It's better to ask questions before and during an assignment, than to try and understand what went wrong after it's due. The proportions given above delineate how I intend to apportion the weight of the various work in the course.

Assignment Due Time

Unless otherwise stated in an assignment, all assignments will be due by classtime on the assignment due date. You should handin PDF’s. The time that we use for an assignment is the submit time. I may ask for supporting material as well (Matlab codes, images, math analysis, etc.).

Appeals Procedure

See the Code of Student Rights and Responsibilities, located in the Class Schedule for more details.

Appeals of Grades and other Academic Actions

If a student believes that an academic action is arbitrary or capricious he/she should discuss the action with the involved faculty member and attempt to resolve. If unable to resolve, the student may appeal the action in accordance with the following procedure:

Appeal to Department Chair who should be notified in writing within 40 working days; chair must notify student of a decision with 15 days. If faculty member or student disagrees with decision, then,
Appeal to Academic Appeals Committee (see flyers posted in MEB and EMCB for members of committee). See II Section D, Code of Student Rights and Responsibilities for details on Academic Appeals Committee hearings.

Assignment Late Policy

No late work is accepted.

Individual Work

The purpose of the assignments is to improve your skills at solving problems and demonstrating that you understand the class material. Collaboration with other class members is acceptable in understanding problems or software tools. For any individual assignments or work turned in, you must do your own work. Using someone else's work or giving someone else your work is considered plagiarism and will be dealt with using standard College and University procedures (i.e., failure of assignment and class). The SoC policy states: "As defined in the University Code of Student Rights and Responsibilities, academic misconduct includes, but is not limited to, cheating, misrepresenting one’s work, inappropriately collaborating, plagiarism, and fabrication or falsification of information. It also includes facilitating academic misconduct by intentionally helping or attempting to help another student to commit an act of academic misconduct. A primary example of academic misconduct would be submitting as one's own, work that is copied from an outside source." (See cheating_policy.pdf and SoC_ack_form.pdf in Link to Class Info and Docs.)

Registration

See university web page for the full academic calendar (Calendar web page). See the university web page for a copy of the withdraw guidelines as well, or see the Student Code.

See the college web page for more Guidelines.

American with Disabilities Act (ADA)

The University conforms to all standards of the ADA. If you wish to qualify for exemptions under this act, notify the Center for Disabled Students Services, 160 Union.
The University of Utah seeks to provide equal access to its programs, services and activities for people with disabilities. If you will need accommodations in the class, reasonable prior notice needs to be given to the Center for Disability Services, 162 Olpin Union Building, 581-5020 (V/TDD). CDS will work with you and the instructor to make arrangements for accommodations.

All written information in this course can be made available in alternative format with prior notification to the Center for Disability Services.