Foundations of Machine Learning

Information

Teachers: Pétia Georgieva

Duration: One semester

Work hours: 162

Contact hours: 45

ECTS: 6

Scientific area: Computer Science

Objectives

Understand the fundamentals of machine learning, namely supervised and unsupervised approaches.
Know how machine learning techniques can be applied in the context of regression models, classification, clustering, deep learning.

Learning outcomes

Knowledge of basic concepts and machine learning techniques.
Digital skills to implement ML algorithms in computer programs.
Reference search and literature review.
Reinforce the experience of working in a team.
Written and oral communication skills.
Problem definition and resolution.

Requirements

For the proper functioning of the course, admitted students must have a good foundation in mathematics (mainly linear algebra)
Know how to program in Python
Understand, read and write in English

Grading

50% Practical assignments

50% Written exam

Methodology

Theoretical-practical classes with exposition of the subject in the theoretical component; solving exercises in the practical component based on Python. Projects allocated to groups of two students. Monitoring of projects outside the classroom.

Syllabus

Fundamental concepts: data-driven models, gradient-based optimization.
Supervised ML: Regression, Classification.
Unsupervised ML: Clustering, k-means; Principal Component Analysis (PCA).
Anomaly Detection.
Deep learning: introduction. Deep Neural Networks.
Convolutional Neural Networks. Object detection and recognition (YOLO neural network).
Recurrent Neural Networks (GRU, LSTM). Time series processing with Long Short Term Memory (LSTM) network.
Reinforcement learning. Markov Decision Process (MDP). Bellman's equation; Q-learning.