Advanced Machine Learning
This theme broadens the fundamentals, and deepens with advanced, and current research in Machine Learning. It draws upon ideas and techniques from various disciplines, such as statistics, decision theory, optimization, and physics.
The courses cover classical statistical learning to modern alternative learning strategies, emphasizing a fundamental understanding of these techniques and their application, including model tuning, selection, validation, testing procedures, and the risks and benefits associated with applying these methods in real-world scenarios.
Year 1 |
|||
Quarter 1 |
Quarter 2 |
Quarter 3 |
Quarter 4 |
Data management and Engineering | Software Engineering and Testing for AI Systems | Responsible Data Science and AI Engineering | Research course |
Machine and Deep Learning | Theme 1 | Theme 1 | Theme 1 |
Probabilistic AI and Reasoning | Theme 2 | Theme 2 | Theme 2 |
Credits: each course in a theme is 5EC, so each theme is 15EC.
Students choose 2 themes, each of which has 3 courses in the 2nd, 3rd and 4th quarters of the 1st year. For this theme, you will take the following courses:
Q2 - Elements of Statistical Learning
This course covers all the basic concepts of Statistical Learning, focusing on the classical techniques of Machine Learning (ML) before the era of Deep Learning. These concepts include classification, (ridge) regression, (hierarchical) clustering, feature reduction and extraction, model selection and bootstrapping, fairness in ML. The emphasis is on the concepts rather than the mathematical details.
Q3 - Alternative Learning Strategies
Many applications of machine learning do not fit exactly into the classical supervised learning setting. This course covers machine learning scenarios and strategies beyond the classical setting, which are relevant to contemporary machine learning research and applications. Examples are causal, meta, adversarial, multiple instance and semi-supervised learning.
Q4 - Generative Modeling
The course will primarily focus on generative modelling and inference techniques. Regarding modelling, it explores the classification of generative models and conducts a comprehensive study of fully observed models, including Flow-Based Models, Autoregressive models, Random Fields, Log-linear models, Markov Models, and more. It also delves into the latent variable models such as probabilistic PCA, diffusion models, variational autoencoders and relevant reparameterization techniques. On the inference side, the course will encompass various learning methods related to generative models. Furthermore, this course examines the practical applications that utilize generative modelling, particularly in the domains of Image Synthesis, Natural Language Processing (NLP), and multimodal learning.