1. CS 535 / CMPE 535 Machine Learning
Instructor’s Name: Asim Karim Year: 2005-06
Office No. & Email: 429, akarim@lums.edu.pk Quarter: Winter
Office Hours: 11.00 to 12.30 MW Category: Graduate
TA for the Course: TBA
Course Code
(Units)
CS 535 / CMPE 535 Machine Learning
(3 Units)
Course
Description
Machine learning investigates the mechanisms by which knowledge is acquired through learning.
It studies adaptive computational systems that improve their performance with experience.
Machine learning methods have been applied to a diverse number of problems ranging from
learning strategies for game playing to recognizing human speech and learning to drive an
autonomous vehicle. This course covers the primary approaches to machine learning, including
inductive inference of decision trees, computational learning methods, Bayesian learning methods,
and reinforcement learning.
Core/Elective This is an elective course.
Pre-requisites MATH 131 Probability and Statistics
Students are expected to have proficiency in algorithm implementation in C/C++/JAVA.
Goals 1. Provide a comprehensive introduction to machine learning methods
2. Build mathematical foundations of machine learning and provide an appreciation for its
applications
3. Provide experience in the implementation and evaluation of machine learning algorithms
4. Develop research interest in the theory and application of machine learning
TextBooks,
Program
ming
Environ
ment,
etc.
Primary Text:
1. Machine Learning, Tom Mitchell, McGraw-Hill International Edition, 1997.
Supplementary Text (strongly recommended):
2. Introduction to Machine Learing, Ethem Alpaydin, Prentice Hall of India, 2004.
2. CS 535 / CMPE 535 Machine Learning
Year: 2005-06
Quarter: Winter
Lectures,
Tutorials &
Attendance
Policy
There will be 19 sessions of 75 minutes each, one in-class midterm exam and one final exam.
There are no points for attendance; however, students who miss lectures will find it very difficult to
make up for the content covered, with the possible additional penalty of missing quizzes.
Grading 25% Assignments (hand + programming)
10% Quizzes
30% Midterm exam
35% Final Exam (Comprehensive)
Additional
Details
The course website will be the primary source for announcements and reading material including
lecture slides, handouts, and web links. http://suraj.lums.edu.pk/~cs535w05
Late policy: 1-day late – 10% reduction; 2-day late – 20% reduction; > 2-day late – assignment not
accepted. Cheating and plagiarism: All cases will be referred to the disciplinary committee for
appropriate action. If an assignment is discussed among students, it is required that each student
writes up the solution independently, and without looking at notes from the discussion. Downloading
code segments from the internet and presenting them as your own is considered plagiarism..
3. CS 535 / CMPE 535 Machine Learning
Year: 2005-06
Quarter: Winter
Topics Sessions Readings
1. Introduction: Overview of machine learning and its applications
2. Concept Learning: Boolean target functions, Inductive learning,
Version spaces, Generalization, limitations
3. Decision Tree Learning: Decision tree learning algorithm,
Discrete valued functions, Inductive bias, Over-fitting of data,
Noisy data
4. Evaluating Hypotheses: Accuracy of a hypothesis over training
and unseen data; Sampling theory basics; Comparing
Hypothesis and Learning algorithms; Issues when data is limited
MIDTERM EXAM
5. Bayesian Learning: Evaluating learning algorithms in Bayesian
framework, Learning algorithms that manipulate probabilities,
Minimum description length; Bayesian belief networks and
conditional independence
6. Computational Learning Theory: PAC learnability; Sample
complexity of hypothesis spaces; VC dimension
7. Reinforcement Learning: Q-learning, Temporal difference
learning, Relationship to dynamic programming
8. Wrap-up: Make-up and introduction to selected advanced topics
FINAL EXAM
1
2-4
5-7
8-10
12-14
15-16
17-18
19-20
Ch 1
Ch 2
Ch 3
Ch 5
Ch 6
Ch 7
Ch 13