THE UNIVERSITY OF LAHORE

Course Outline

CS-13410 Introduction to Machine Learning

Spring-2022
Credit Hours (3+1)
Weekly tuition pattern 2 session lectures (90 min each)
Artificial Intelligence, Design & Analysis of Algorithm, Statistics,
Prerequisites: Programming (Java, C++ or Python)
Course Instructor Mr. Waqar Ashiq

Syllabus Designed By Dr. Mudasser Naseer

1- Course Description
This course is designed to introduce the field of Machine Learning, 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, and
learn a function mapping from the input to the 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. The students will be exposed to the application of machine learning to a
range of real-world problems.
2- Objectives
The objective of this course is:-
1. To provide a broad survey of approaches and techniques in machine learning;
2. To develop a deeper understanding of several major topics in machine learning;
3. To develop the design and programming skills that will help the students to build
intelligent, adaptive artifacts;
4. To develop the basic skills necessary to pursue research in machine learning.

3- Course Learning Outcomes (CLOs)

At the end of the course the students will be able Domain BT Level* Graduating
to: Attributes
(GAs) of
PLOs
1. Have a good understanding of the fundamental C Understand, GA1, GA2

Page 1 of 7
The University of Lahore
 issues and challenges of machine learning: data
collection and pre-processing, model selection, (2)
model complexity, etc.
2. Have an understanding of the strengths and GA1, GA2
Understand,
weaknesses of many popular machine learning C
(2)
approaches.
3. Apply pattern classification algorithms to Application
classify multivariate data. C , GA2
(3)
4. Be able to analyze, evaluate, design and Analyze, GA3,
implement various machine learning algorithms in Evaluate, GA4, GA5
C
a range of real-world applications. Create,
(4, 5, 6)
* BT= Bloom’s Taxonomy, C= Cognitive domain, P= Psychomotor domain, A=
Affective domain

4- Course Contents
Basic concepts of Machine Learning; Supervised learning; Supervised learning setup.
Decision tree, KNN, Baysian Learning, Naïve Bayes Classifier, Simple and Logistic
regression; Perceptron; Generative learning algorithms; Gaussian discriminant analysis;
Support vector machines; Model selection and feature selection; Evaluating and
debugging learning algorithms; Learning theory; Bias/variance tradeoff; Unsupervised
learning; K-means Clustering; K-Medoid Clustering; Herarchichal Clustering; EM
algorithm. PCA (Principal components analysis); Reinforcement learning and control;
5- Teaching Methodology / Course Structure
Lectures, Written Assignments, Practical Assignments, Semester Project, Presentations
6- Course Assessment
Sessional Exam, Home Assignments, Quizzes, Project, Presentations, Final Exam
7- Text
1. BookName: Machine Learning by Tom M. Mitchell
8- Reference Materials
1. Introduction to Machine Learning by Ethem Alpaydin 2nd Edition
2. An Introduction to Machine Learning by Miroslav Kubat 2nd Edition
9- Course Duration
This course will be held twice a week of 1.5 hours class duration.
10- Course style
The course will be delivered in a classroom environment.
11- Additional Course Requirement
In addition to the objectives of this course, students are expected to gain skills to model
and verify the real time case studies of the concurrent systems using automatic model
checking tools.

Page 2 of 7
The University of Lahore
 12- Course Outline
The lecturers are supposed to complete the following topics/sub-topics before the
mid/final term examination as prescribed in the course outline below:

Topics Covered
Week Lecture Topic Comments
1 Introduction to Machine Learning Course
 Course Outline
 Learning Objectives
 Course Learning Outcomes
1
2 Basics of Machine Learning
 Inductive and Deductive Learning
 ML Applications
 Major Challenges in ML
 Machine Learning VS Traditional Programming
1 Designing a ML problem Assignment
 Learning Input Output Functions General Settings 01 –
 Concept Learning / Function Learning Complete
 Learning is a Searching Problem Python
 Representation of Instance/Examples Learning
Tutorial
2 2 Steps to Build Efficient Machine Learning Models
 Data Understanding and preprocessing
 Represent Data into Machine Understandable
Format
 Treating Real World Problems as Learning Input
Output Functions in Machine Learning
 Categorization of Input and Output Data
3 1 Types of Learning in Machine Learning
 Supervised Learning
 Unsupervised Learning
 Semi supervised Learning
Machine Learning Cycle
 Training, Testing, and Application Phase
 Balanced and Unbalanced Data
 Data Split Approachs (Class balanced Split and
Random data Split)
Machine Learning Training Regimes
 Batch Method
 Incremental Method
 Online Method
2 Basics of Concept Learning
 Instance Space and Computing Size of Instance
Space

Page 3 of 7
The University of Lahore
  Concept Space and Computing Size of Concept
Space
 Hypothesis Space and Using Search Bias to Reduce
Size of Concept Space
1 Hypothesis and Hypothesis Space Representation
 Concept Representation as Conjunction (AND) of
Input Attributes
 Concept Representation as Disjunction (OR) of
Conjunction (AND) of Input Attributes
 Comparison of Concept Representations
 Hypothesis in the form of Rules
4  Reduction in Size of Concept Space(C)
 Syntactically Distinct vs Semantically Distinct
Hypothesis
 General to Specific Ordering of Hypotheses
2 Find S Algorithm Quiz 1
 Representing Training Examples and Hypothesis for
FIND S Machine Learning Algorithm
 FIND-S Algorithm Machine Learning Cycle
1 Version Space Algorithms Assignment 2
 Basics of Version Space
 List Then Eliminate Algorithm
 Candidate Elimination Algorithm
5 2 Candidate Elimination Algorithm – Machine Learning
Cycle
 Evaluation using Single model approach
 Evaaluation using Ensemble Approach
 Voting based method
1 Evaluating Hypothesis (Models) Quiz 2
 Why Evaluate Hypotheses (Model)?
 Train Test and K-Fold Methodology
 Comparing Machine Learning Algorithms
 Evaluation Measures for Classification Problems
6
 Evaluation Measures for Regression Problems
 Evaluation Measures for Sequence to Sequence
Problems
2  Inplementation of training, testing phase of machine
learning in python using structured dataset
7 1 Decision Tree Learning
 Decision Trees
 ID3 - Basic Decision Tree Learning Algorithm
 Entropy and Information Gain
2 Decision Tree Learning Conti…
 Inductive Bias in Decision Tree Learning
 Refinements to Basic Decision Tree Learning
 Overfitting Training Data

Page 4 of 7
The University of Lahore
  Methods to avoid overfitting
1
8
2
1
9 Mid-Term Exams
2
1 KNN for Classification
10
2 Bayesian Learning, Naïve Bayes Classifier
1 Introduction to Linear regression, Least Square Method
11 2 Gradiant Descent, Multiple Linear Regression, Logistic
regression
1 Unsupervised Learning, Dimensionality Reduction, Quiz-3
PCA
12
2 PCA – with example Assignment
#3
1 Support Vector Machine
13
2 Clustering, k-Mean clustering
1 K-Medoid Clustering, CLARA
14
2 Hierarchical Clustering-MST, BIRCH Quiz-4
1 Artificial Neural Networks (ANN) – Introduction, Assignment
15 Perceptrons #4
2 Multilayered Networks and Backpropagation Algorithm
1 Backpropagation Algorithm -Examples
16
2 Deep Learning Models
17,18 Final Term Exam

13- Assessment Criteria

ASSESSMENT CRITERIA
No. Assessment Percentage
1. Quiz 10%
2. Assignments + Project 25%
3. Midterm 25%
4. Final 40%
5.
Total 100%

14- Attendance Requirements

You are expected to attend all lectures, seminars, tutorials, and lab sessions or any other
classroom activity announced. Where you fail to attend classes, you cannot expect the
lecturer to brief you on what you have missed. You are responsible for your attendance,
not the academic staff. Attendance at tutorials and lab sessions will be strictly monitored,
and failure to attend will be taken into account.
Note: Minimum of 75% Attendance in lectures/lab sessions/seminars (if any) are
required for a student to sit in the Final-Term examination.

Page 5 of 7
The University of Lahore
Etiquette: Please keep all cell phones turned OFF during class. If your activities during
class are deemed disruptive, you will be asked to leave.
15- Submission and Collection of Assignment
All assignments should be handed in at the beginning of the class sessions when they are
due. All assignments may be handed back during scheduled classes.
16- General Information
Students are required to be familiar with the university code conduct, and to abide by its
terms and conditions.
16.1 Copying of Copyright Material by Student
A condition of acceptance as a student is the obligation to abide by the University’s
policy on the copying of copyright material. This obligation covers photocopying of any
material using the University’s photocopying machines, and the recording off air, and
making subsequent copies, of radio or television broadcasts, and photocopying textbooks.
Students who flagrantly disregard University policy and copyright requirements will be
liable to disciplinary action under the Code of Conduct.
16.2 Academic Misconduct
Please refer to the Code of Conduct for definitions and penalties for Academic
Misconduct, plagiarism, collusion, and other specific acts of academic dishonesty.
Academic honesty is crucial to a student's credibility and self-esteem, and ultimately
reflects the values and morals of the University as a whole. A student may work together
with one or a group of students discussing assignment content, identifying relevant
references, and debating issues relevant to the subject. Academic investigation is not
limited to the views and opinions of one individual, but is built by forming opinion based
on past and present work in the field. It is legitimate and appropriate to synthesize the
work of others, provided that such work is clearly and accurately referenced. Plagiarism
occurs when the work (including such things as text, figures, ideas, or conceptual
structure, whether verbatim or not) created by another person or persons is used and
presented as one’s own creation, unless the source of each quotation or piece of borrowed
material is acknowledged with an appropriate citation. Encouraging or assisting another
person to commit plagiarism is a form of improper collusion and may attract the same
penalties. To prevent Academic Misconduct occurring, students are expected to
familiarize themselves with the University policy, the Subject Outline statements, and
specific assignment guidelines. Students should also seek advice from Subject Leaders on
acceptable academic conduct.
16.3 Guidelines to Avoid Plagiarism
Whenever you copy more than a few words from any source, you must acknowledge that
source by putting the quote in quotation marks and providing the name of the author. Full
details must be provided in your bibliography. If you copy a diagram, statistical table,
map, etc., you must acknowledge the source. The recommended way is to show this
under the diagram. If you quote any statistics in your text, the source should be
acknowledged. Again full details must be provided in your bibliography. Whenever you
use the ideas of any other author you should acknowledge those, using the APA

Page 6 of 7
The University of Lahore
(American Psychological Association) style of referencing.
Students are encouraged to co-operate, but collusion is a form of cheating. Students may
use any sources (acknowledged of course) other than the assignments of fellow students.
Unless your Subject Leader informs you otherwise, the following guideline should be
used: Students may work together in obtaining references, discussing the content of the
references and discussing the assignment, but when they write, they must write alone
16.4 Referencing For Written Work
Referencing is necessary to acknowledge others' ideas, avoid plagiarism, and allow
readers to access those others’ ideas. Referencing should:
1. Acknowledge others' ideas
2. Allow readers to find the source
3. Be consistent in format and
4. Acknowledge the source of the referencing format
To attain these qualities, the school recommends use of either the Harvard or American
Psychological Association style of referencing, both of which use the author/date.
16.5 Referencing Standards
APA style referencing
17- Approval
Checked by, Approved by,

Page 7 of 7
The University of Lahore