Machine Learning (CAI-761) - Theory
Credit Hours: 3(3-0) Prerequisites: None
Teacher: [Link] Office: R-22 Noureen_zafar@[Link].p
Zafar k
Course Introduction:
Machine learning is one of the fastest growing areas of computer science, with far-
reaching applications. The aim of this course is to:
a) Present the basic machine learning concepts;
b) Present a range of machine learning algorithms along with their strengths and
weaknesses;
c) Apply machine learning algorithms to solve problems of moderate complexity.
Reference Materials:
Ian H. Witten, Data Mining: Practical machine learning tools and techniques, Morgan
Kaufmann, 2016
EthemAlpaydin, Introduction to Machine Learning, 2nded.
SergiosTheodoridis, Konstantinos Koutroumbas, Pattern Recognition, 4thed.
Richard O. Duda, Peter E. Hard, Pattern Classification, 2nded.
LaureneV. Fausett, Fundamental of Neural Networks: Architectures, Algorithms and
Applications.
Course Objectives:
Students will be able to use machine learning techniques to solve complex real-
world problems, by identifying the right method for your task, implementing an
algorithm, assessing and improving the algorithm’s performance.
Understand the formulation of well-specified machine learning problems
Learn how to perform supervised and reinforcement learning, with images and
temporal sequences.
Week # Topics
Course Introduction
Week 1
Machine Learning (ML) course overview, Overview of ML Tools
Machine Learning Examples, Supervised and Unsupervised
Learning
Week 2
Supervised Learning: Formulation, Setup, Train-test split, Generalization
Discriminative vs. Generative Learning,
Regression: Linear Regression and Regression Model, Train the model
with gradient descent
Week 3
Bayesian Decision Theory
Parametric and Non Parametric Methods
Dimensionality Reduction
Introduction
Density Estimation
Histogram Estimator
Kernel Estimator
k-Nearest Neighbor Estimator
Week 4
Multivariate Data
Nonparametric Classification
k-NN Classifier
Condensed Nearest Neighbor
Nonparametric Regression
Selection of Smoothing Parameters
Clustering Approaches
Introduction
k-Means Clustering
Data description and clustering
Week 5
Hierarchical Clustering
Criterion function for clustering
The number of cluster problem and cluster validation
Classification Techniques
Week 6 Artificial Neural Networks
Multilayer Perceptron
Competitive Learning:
Support Vector Machines
Week 7
Decision Trees, Random Forests, and XGBoost
Introduction
Decision Node
Leaf Node
Class
Univariate trees
Classification trees
Regression trees
Pruning
Rule extraction from trees
Multivariate trees
Combining Multiple Classifiers
Rationale
Generating Diverse Learners
Model Combination Schemes
Voting
Week 8 Bagging
Boosting
Mixture of Experts
Stacked Generalization
Fine-Tuning an Ensemble
Cascading
Midterm Exam
Stacked Generalization introduction
Combined different models and implement in python
Week 9
Reinforcement Learning
K-armed Bandit Problem
Define reward
Week 10 Understand the temporal nature of the problem
Define k-armed bandit
Transfer Learning
Develop Model Approach
Week 11
Pre-trained Model Approach
Adversarial Learning
Deep Learning
Adversarial Attacks
Week 12 Defending Against Adversarial Attacks
Memory based learning
Week 13
Transformation based learning
What is the Transformer model?
Transformer model: general architecture
What is the Transformer neural network?
Functioning in brief
Week 14
Famous Transformers models
BERT
GPT-2, GPT-3
Limitations of the Transformer
Inductive logic programming
Learning from entailment
Week 15
Learning from interpretations
Approaches to ILP
Week 16 Introductory Machine Learning packages
Final Term Exam
ProgrammingAssignments DoneintheCourse:
Students will be assigned a series of in-class and homework tasks designed to develop
and enhance their programming and problem-solving skills.
Teaching Methodology:
Lectures, Written Assignments, Practical labs, Semester Project, Presentations
Courses Assessment:
Mid Exam, Home Assignments, Quizzes, Project, Presentations, Final Exam
Grading for 3(3-0) A-48-above, B=39-47, C=30-38, D=24-29, F= less than 24
ClassTimeSpenton(incredithours):
Theory ProblemAnalysis Solution Designand SocialandEthicalIssue
PracticalDevelopment s
40% 15% 40% 5%
Course Learning Outcomes (CLOs)/PLOs Mapping
At the end of course the students will be able to: Bloom Taxonomy
1. Describe basic machine learning concepts, theories and applications. C1 (Knowledge)
2. Apply supervised learning techniques to solve classification problems C3 (Apply)
of moderate complexity
3. Apply unsupervised learning techniques to solve C3 (Apply)
4. Apply reinforcement learning algorithms to environments C3 (Apply)
C=Cognitive domain
CLO – PLO Mapping
CLOs
CLO:1 CLO:2 CLO:3
PLOs
PLO-1 √ √
(Knowledge)
PLO-2 √
(ProblemAnalysis)
PLO-3 √
(System Design)