VY CROSS test 2Laajuus (5 cr)

Objective

Machine learning is related to the technologies of making computing devices learn and extract "hidden" information from input-data patterns. Extracted information could be used to make reasonable output (it can be in the form of suggestions, conclusions, or decisions), or to gain deep knowledge (by exploring data) about specific behavior. The student who successfully completes this course will be: 1. Able to explain the manifestation of machine learning and its possible applications. Furthermore, they will be familiar with several concepts like data modelling, overfitting, underfitting, generalisation, memorisation, learning data, and validating data. 2. Aware of supervised learning algorithms and their different kinds and applications 3. Able to apply different regression methods as well as neural networks to capture hidden relations in supervised learning 4. Able to explain probabilistic models and Bayesian-based machine learning algorithms. 5. Aware of data quality in machine learning and how to improve and clean data. 6. Able to explain classification algorithms as well as apply them in simple scenarios. 7. Aware of unsupervised learning concepts and clustering. 8. Able to define reinforcement learning and its main differences with supervised and unsupervised machine learning. 9. Aware of the applications as well as limitations of machine learning algorithms. 10. Finally, the course develops lifelong learning, Oral, written, and interpersonal skills (Group Work, English), critical and analytical thinking, problem modeling and solving skills, IT skills, and optimized decisions. The issue of professional ethics, norms of handling big data and data protection protocol are considered as an integral part of machine learning process.

Content

1. Introduction to machine learning and data modelling. 2. Supervised learning algorithms and neural networks 3. Probability theory and Bayesian-based algorithms. 4. Parametric Algorithms 5. Enhance data quality and Principal component analysis 6. Unsupervised machine learning and clustering 7. Kernel machines and SVM 8. Hidden Markov Models 9. Reinforcement learning 10. Combine Algorithms and Applications

Qualifications

It is recommended to know: the fundamentals of probability theory, linear algebra, optimisation theory, matrix calculus, and some programming skills.

Materials

1. Lecturer Notes 2. E. Alpaydin: Introduction to Machine Learning, 3rdEdition, MIT Press, 2014 3. S. Rogers and M. Girolami, "A First Course in Machine Learning", 2nd Edition, CRC Press 2017

Further information

Responsible Unit: School of Technology and Innovations

Execution methods

All lectures are recorded. The teaching style is based on Flipped Learning. Lectures 32 h, independent work 103 h

Accomplishment methods

Online 15 Quizzes (about 150 questions). A Written report could be submitted for Bouns (optional). Grading: On a scale 1-5 or fail