2024 Faculty Courses School of Engineering Undergraduate major in Information and Communications Engineering

Machine Learning (ICT)

Academic unit or major: Undergraduate major in Information and Communications Engineering
Instructor(s): Ze Jin / Kenji Suzuki
Class Format: Lecture (Face-to-face)
Media-enhanced courses: -
Day of week/Period (Classrooms): 3-4 Tue / 3-4 Fri
Class: -
Course Code: ICT.S311
Number of credits: 200
Course offered: 2024
Offered quarter: 3Q
Syllabus updated: Mar 14, 2025
Language: Japanese

Syllabus

Course overview and goals

In this course, you will learn the mathematics, basic principles, methods of machine learning and their applications. To make those machine learning methods your own knowledge, you will practice programming of them and how to implement them by using machine learning libraries. First, you will learn the fundamental knowledge on and basic principles of machine learning including supervised & unsupervised learning, classification & regression, and ill-posed problem & optimization. Next, you will learn major machine learning models and methods and the basic principles, features, and implementations of them. Those models and methods include linear discriminant analysis, principal component analysis, training and validation methods, clustering, support vector machines, neural networks, and deep learning models.

Course description and aims

To learn the fundamentals and basic principles of machine learning and make them your own knowledge and skills to solve machine learning problems. To learn major machine learning methods and make them your own knowledge to apply them properly. To learn how to implement machine learning methods by programming.

Keywords

Least square method, Linear discriminant analysis (LDA), Principal component analysis (PCA), K-means method, Bayes estimation, Neural networks, Backpropagation, Support vector machine (SVM), Deep learning, Convolutional neural network (CNN).

Competencies

Specialist skills
Intercultural skills
Communication skills
Critical thinking skills
Practical and/or problem-solving skills

Class flow

In the first three classes, you will learn the fundamental mathematics and basic principles of machine learning. In the following eight classes, you will learn major machine learning models and methods. In each class, you will practice the implementations of the methods by programming them with machine learning libraries. In the last three classes, you will learn and practice deep learning.

Course schedule/Objectives

	Course schedule	Objectives
Class 1	Introduction to the course	To learn the summary of the course, and the basic knowledge on, the field of, and applications of machine learning.
Class 2	Principle of learning 1	To understand supervised & unsupervised learning, classification & regression, ill-posed problem & optimization, and the least square method.
Class 3	Principle of learning 2	To understand the degrees of freedom in a model & different types of errors, over-fitting & regularization, and the cause of dimensionality.
Class 4	Linear discriminant analysis (LDA)	To learn the linear discriminant, within-class variance & between-class variance, and Fisher’s linear discriminant analysis.
Class 5	Principal component analysis (PCA)	To learn the mathematics behind PCA, PCA for dimensionality reduction, data compression with PCA, and differences from LDA.
Class 6	Methodology of training, validation, and testing	To study the methodology of training, validation, and testing, and performance metrics in machine learning.
Class 7	Unsupervised learning and Bayesian methods	To learn clustering algorithms including the K-means algorithm, maximum likelihood estimation, and expectation-maximization (EM) algorithm.
Class 8	Neural networks 1	To learn biological neural networks and their models, the history of neural networks, a neuron model, and the perceptron.
Class 9	Neural networks 2	To learn the multilayer perceptron, activation functions, and the error back propagation and its derivation.
Class 10	Neural networks 3	To learn how to interpret neural networks, how to design neural networks, and neural network applications.
Class 11	Support vector machines (SVMs), pre-processing of data, and fusing multiple models	To learn linear and nonlinear SVMs.
Class 12	Overview of deep learning	To learn convolutional neural networks (CNNs), LeNet, AlexNet, VGG-16, ResNet, and MTANN.
Class 13	Computation of deep learning	To learn the convolution, pooling, and softmax layer in convolutional neural networks (CNNs).
Class 14	Applications of deep learning	To learn deep learning applications to image processing, object detection, segmentation, super-resolution, voice recognition, natural language processing, medical image processing, and medical image diagnosis.

Study advice (preparation and review)

To enhance effective learning, students are encouraged to spend approximately 100 minutes for preparing for a class and another 100 minutes for reviewing the class contents (and assignments) after each class by using the textbook, reference books, and other course materials.

Textbook(s)

A reference book to study neural network and deep learning lectures and practices in the latter half of the classes in this course.
J. Krohn (author), K. Suzuki (supervisor of a translation), M. Shimizu (translator),「Python, TensorFlowで実践する深層学習入門: しくみの理解と応用」* , Tokyo Kagaku Dojin Publishing, Ltd., 278 pp., ISBN：9784807920389, 2022.
*This book is the Japanese edition of “Deep Learning Illustrated”.

Reference books, course materials, etc.

A reference book to some machine learning methods in this course.
M. Sugiyama, 「イラストで学ぶ　機械学習」最小二乗法による識別モデル学習を中心に, Kodansha ltd., ISBN：978-4-06-153821-4, 2013

Evaluation methods and criteria

Final examination and reports for assignments and programming exercises.

Related courses

LAS.M102 ： Linear Algebra I / Recitation
LAS.M101 ： Calculus I / Recitation
LAS.M105 ： Calculus II

Prerequisites

Basic knowledge on differential and integral calculus.