Advanced Machine Learning MethodsLaajuus (5 cr)
Code: R504D108
Credits
5 op
Teaching language
- English
Objective
- Knowledge and skills to understand beyond-basic contemporary machine learning (ML) models and methods, and to choose and apply them in a principled and sound way
- Abilities for understanding connections to, and dependencies between, model/method properties and timely topics (e.g., ethics, sustainability, explainability).
- Abilities to solve a computational problem via machine learning without using a high-level ML application programming interface
Content
- Theory and practice of the beyond-basic contemporary machine learning (ML) models and methods
- Use of suitable tools (e.g., an ML application programming interface enabling both high and low level expression) for building solutions
Qualifications
Introduction to Machine Learning Methods
Assessment criteria, satisfactory (1)
Grade 1: The student knows the theory on the considered ML models and methods. The student is able to solve beyond-basic contemporary ML problems, using the considered tools.
Assessment criteria, good (3)
Grade 3: The student understand the theory on the considered ML models and methods. The student is able to solve a variety of beyond-basic contemporary ML problems, using the considered tools, suitably.
Assessment criteria, excellent (5)
Grade 5: The student understand the theory on the considered ML models and methods. The student is able to solve a variety of beyond-basic contemporary ML problems, using the considered tools, most suitably.
Enrollment
01.10.2024 - 31.12.2024
Timing
03.02.2025 - 25.04.2025
Credits
5 op
Mode of delivery
Contact teaching
Unit
Bachelor of Engineering, Information Technology
Teaching languages
- English
Teachers
- Tuomas Valtanen
Responsible person
Tuomas Valtanen
Student groups
-
R54D22S
Objective
- Knowledge and skills to understand beyond-basic contemporary machine learning (ML) models and methods, and to choose and apply them in a principled and sound way
- Abilities for understanding connections to, and dependencies between, model/method properties and timely topics (e.g., ethics, sustainability, explainability).
- Abilities to solve a computational problem via machine learning without using a high-level ML application programming interface
Content
- Theory and practice of the beyond-basic contemporary machine learning (ML) models and methods
- Use of suitable tools (e.g., an ML application programming interface enabling both high and low level expression) for building solutions
Location and time
Lapland University of Applied Sciences, Rantavitikka Campus, 13.1.2025 - 15.5.2025.
Teaching methods
Lectures, workshops, examples, exercises and self-supervised work.
Exam schedules
The course will be graded based on personal work and exercises.
Content scheduling
- Natural language processing machine learning models
- Advanced specialized machine learning algorithms
- Machine learning model optimization tools
- Error metrics and ML model performance evaluation
- Use of suitable tools (e.g., a high-level ML application programming interface, like scikit-learn and TensorFlow) for building solutions
Evaluation scale
H-5
Assessment criteria, satisfactory (1)
Grade 1: The student knows the theory on the considered ML models and methods. The student is able to solve beyond-basic contemporary ML problems, using the considered tools.
Assessment criteria, good (3)
Grade 3: The student understand the theory on the considered ML models and methods. The student is able to solve a variety of beyond-basic contemporary ML problems, using the considered tools, suitably.
Assessment criteria, excellent (5)
Grade 5: The student understand the theory on the considered ML models and methods. The student is able to solve a variety of beyond-basic contemporary ML problems, using the considered tools, most suitably.
Assessment methods and criteria
The course will be graded on the scale of 1 - 5 and failed (0). The grading will be based on the submitted exercises/assignments.
Qualifications
Introduction to Machine Learning Methods