AI506: Advanced Machine Learning

The Study Board for Science

Teaching language: Danish or English depending on the teacher, but English if international students are enrolled
EKA: N400006102
Assessment: Second examiner: External
Grading: 7-point grading scale
Offered in: Odense
Offered in: Spring
Level: Bachelor

STADS ID (UVA): N400006101
ECTS value: 7.5

Date of Approval: 08-10-2024


Duration: 1 semester

Version: Approved - active

Entry requirements

This course cannot be chosen by students, who have passed DM568, DM873 or DS809

Academic preconditions

Academic preconditions. Students taking the course are expected to:

  • Have basic understanding of linear algebra
  • Have basic programming skills in python, including management of virtual environments, utilization of libraries and basic data science operations

Course introduction

Machine learning is currently transforming the world we are living in and has demonstrated successful applications in almost all domains of science as well as everyday life. As useful and impressive advanced machine learning techniques are, they all also come with limitations, drawbacks and pitfalls. For everyone aiming for proficiency in Artificial Intelligence, it is absolutely crucial to not only apply basic machine learning techniques but also understand them and use them correctly to avoid errors, biases and achieve optimal performance.

In this course, we will discuss the fundamentals of advanced techniques in machine learning like deep neural networks, transformers, graph neural networks, and probabilistic modeling, as well as their applications to practical problems. We will learn about the power but also the limitations of these methods. At the end of the course, the students will have significant familiarity with the subject and will be able to apply the learned techniques to a broad range of different fields.

Expected learning outcome

In order to achieve the purpose of the course, the learning objective for the course is for the student to demonstrate the ability to:

  • Describe the design and function principles of the presented algorithms in a mathematically precise language.
  • Analyse the various types of advanced machine learning techniques and argue for their ideal field of application and make an informed choice of method for a concrete problem at hand.
  • Reflect on the different machine learning approaches, their advantages/disadvantages and comment on their most appropriate application to specific problems.
  • Adapt advanced machine learning algorithms to domain-specific tasks.
  • Utilize libraries to implement, train and apply all presented methods in practice.





Content

  • Probabilistic Modeling
  • Deep Neural Networks
  • Optimization and Regularization
  • Transformers
  • Graph Neural Networks
  • Autoencoders
  • Fundamentals of interpretability in machine learning

    Literature

    See itslearning for syllabus lists and additional literature references.

    Examination regulations

    Exam element a)

    Timing

    June

    Tests

    Portfolio

    EKA

    N400006102

    Assessment

    Second examiner: External

    Grading

    7-point grading scale

    Identification

    Full name and SDU username

    Language

    Normally, the same as teaching language

    Duration

    Oral exam - 30 minutes incl. grading, no preparation

    Examination aids

    No aids allowed

    ECTS value

    7.5

    Additional information

    The portfolio exam consists of two elements:
    1. Two larger assignments during the course
    2. Oral exam

    Indicative number of lessons

    54 hours per semester

    Teaching Method

    At the faculty of science, teaching is organized after the three-phase model ie. intro, training and study phase.
    • Intro phase: 36 hours
    • Training phase: 18 hours comprised of 18 hours tutorials
    Activities during the study phase:
    • Solution of small take-home exercises in order to discuss these in the exercise sections.
    • Solving the project assigments
    • Self study of various parts of the course material.
    • Reflection upon the intro and training sections.

    Teacher responsible

    Name E-mail Department
    Lukas Paul Achatius Galke galke@imada.sdu.dk Department of Mathematics and Computer Science
    Melih Kandemir kandemir@imada.sdu.dk Concurrency

    Timetable

    Administrative Unit

    Institut for Matematik og Datalogi (datalogi)

    Team at Registration

    NAT

    Offered in

    Odense

    Recommended course of study

    Transition rules

    Transitional arrangements describe how a course replaces another course when changes are made to the course of study. 
    If a transitional arrangement has been made for a course, it will be stated in the list. 
    See transitional arrangements for all courses at the Faculty of Science.