Lecture 01

CS-13410
Introduction to Machine Learning
by
Mudasser Naseer
 Course information
Instructor: Dr. Mudasser Naseer
Email: [email protected]
Room # G03
Lectures: Sec-7B: Wednesday - 02:00 – 03:30 (CS-001)
Friday - 11:00 – 12:30 (CS-001)

Office Hrs: Monday, Tuesday, Thursday : 08:30 – 12:30

Wednesday: 11:00 – 12:30
Friday: 08:30 – 11:00
(or by appointment)
Recommended Prerequisite: Statistics, Programming (Java, C+
+ or Python) 2
 Travel Advisory
You are about to take the “Introduction to Machine
Learning” course.
It is intended primarily for senior semester Computer
Science students.
The course is mainly a theoretical and mathematical course.
Broad coverage of different machine learning topics (some
important topics in depth).
We will discuss numerical examples so that the students
will be able to implement different algorithms for their
FYPs.
The course will require a substantial amount of work by
yourselves to digest and pass successfully. 3
Aim and Objectives
There are four primary objectives for the course:
To provide a broad survey of approaches and
techniques in machine learning;
To develop a deeper understanding of some major
topics in machine learning;
To develop the design and programming skills that will
help you to build intelligent, adaptive artifacts;
To develop the basic skills necessary to pursue
research in machine learning.

4
 Learning outcomes
On completion of the course students will be expected to:
Have a good understanding of the fundamental issues
and challenges of machine learning: data collection and
pre-processing, model selection, model complexity, etc.
Have an understanding of the strengths and weaknesses
of many popular machine learning approaches.
Appreciate the underlying mathematical relationships
within and across Machine Learning algorithms and the
paradigms of supervised and un-supervised learning.
Be able to design and implement various machine
learning algorithms in a range of real-world applications.
5
Resources
Lecture slides
Text book:
Machine Learning by Tom M. Mitchell
Introduction to Machine Learning by Ethem
Alpaydin 2nd Ed
Reference Book:
An Introduction to Machine Learning by Miroslav
Kubat 2nd Edition

6
Administrative details
Assessment:
Final numeric score will be weighted as follows:

Homework/Assignments (4): 10%

Quizzes (4): 15%
Midterm Exam: 30%
Final Exam: 45%

(This is a tentative distribution and might change slightly)

Your final grade will follow University of Lahore conventions,

i.e. A, A-, B+, etc. 7
Administrative details
Grade determined by mid, final exams, quizzes,
and assignments.
The assignment/homework should be submitted
individually.
Each “normal” week (not counting holidays,
strikes, ceremonies, wars, etc.) there will be 2
lectures of theory.

8
Collaboration on Assignments, etc.
Preparing assignments/homeworks independently is a key
ingredient for understanding the material (and,
consequently, passing the exam :-). So it is highly
recommended you make a serious effort to solve the
problems on your own.
You may discuss, consult, collaborate with other students
on the problem sets, but your solutions must be written up
independently, by you.
You are welcome to consult online and offline sources for
your solutions, but you are (a) expected to give clear cites
of your references, and (b) use a write up of your own.
Recall that Google (or any alternative search engine) is a
double edge sword. 9
Collaboration on Assignments, etc.
Cases of plagiarism that will be detected will be
dealt with severely. (For example, reducing your
grades for the whole course, not just the relevant
assignment, canceling the assignment and/or
reporting the incident to the appropriate
university authority.)
 If I suspect X had copied from Y, both will be
regarded as cheaters.

10
Course Structure
We will start from the basic concept of machine
learning and move towards advance topics using
multiple examples from each topic.
This should hopefully expose you to some of the
beautiful topics and ideas in the field.
 Naturally we will get into some of the topics in
great depth and their corresponding applications.
Each topic will be accompanied by some examples
from real world (given by me) and tasks (to be done
by you).

11
Course overview
 What is this course Introduction to Machine Learning?

 Covers basic concepts of ML, in particular focusing on the core

concepts of supervised and unsupervised learning.

 In supervised learning we will discuss algorithms which are trained

on input data labelled with a desired output.

 Unsupervised learning aims to discover latent structure in an Input

signal where no output labels are available.

 Students will learn the algorithms which underpin many popular

Machine Learning techniques, as well as developing an

understanding of the theoretical relationships between these
algorithms. 12
How to succeed?
Put the effort in
Make good use of the available resources
Attend lectures
Read books and online materials
Lot of practice
ASK!!!
Aim to fully understand new concepts

13
HOW TO FAIL?
Relax too much
Don’t assimilate material given in lectures
Absent from lectures
Don’t seek help
“Borrow” from “others”
They can’t help in the learning, exam or class tests!!!!

14
Why “Learn” ?
Machine learning is programming computers to optimize
a performance criterion using example data or past
experience.
There is no need to “learn” to calculate payroll
Learning is used when:
Human expertise does not exist (navigating on Mars),
Humans are unable to explain their expertise (speech
recognition)
Solution changes in time (routing on a computer network)
Solution needs to be adapted to particular cases (user
biometrics)

15
What We Talk About When We Talk
About“Learning”
Learning general models from a data of particular
examples
Data is cheap and abundant (data warehouses, data
marts); knowledge is expensive and scarce.
Example in retail: Customer transactions to consumer
behavior:
People who bought “Blink” also bought “Outliers”
(www.amazon.com)
Build a model that is a good and useful approximation to
the data.
16
A Few Quotes
“A breakthrough in machine learning would be worth
ten Microsofts” (Bill Gates, Chairman, Microsoft)
“Machine learning is the next Internet”
(Tony Tether, Director, DARPA)
Machine learning is the hot new thing”
(John Hennessy, President, Stanford University)
“Web rankings today are mostly a matter of machine learning”
(Prabhakar Raghavan, Dir. Research, Yahoo)
“Machine learning is going to result in a real revolution” (Greg
Papadopoulos, CTO, Sun)
“Machine learning is today’s discontinuity”
(Jerry Yang, CEO, Yahoo) 17
So What Is Machine Learning?
Automating automation
Getting computers to program
themselves
Writing software is the bottleneck
Let the data do the work instead!

18
Traditional Programming
Data
Computer Output
Program

Machine Learning

Data
Computer Program
Output

19
Magic?
No, more like gardening

Seeds = Algorithms
Nutrients = Data
Gardener = You
Plants = Programs

20
Sample Applications
Web search
Computational biology
Finance
E-commerce
Space exploration
Robotics
Information extraction
Social networks
Debugging
[Your favorite area]

21