Tom (Tomáš) Pecher

Supervised Learning: Course Introduction

Arguably the most influential subfield of machine learning, the breakthroughs in supervised learning sparked the AI revolution of the 20th century. At its core, supervised learning is the study of functions that map inputs to outputs: specifically, using functions to approximate analytical mappings between datasets. This goal of these approximations is, of course, so that they can be used to predict future outcomes, even for data we have not seen before. Sometimes, we have an analytical formula that we hope to fit to the data, such as in linear regression. However, more recently with the rise of deep learning, we forego known analytic functions and instead focus on learning from data using neural networks.

Course Structure:

• Section 1: Mathematical Foundations
• Section 2: Numerical Optimisation Methods
• Section 3: Supervised Learning Framework
• Section 4: Bias-Variance Decomposition
• Section 5: PAC Learning Theory
• Section 6: Linear Models and Regression
• Section 7: Generalised Linear Models
• Section 8: K-Nearest Neighbours
• Section 9: Kernel Methods
• Section 10: Non-parametric Regression
• Section 11: Decision Trees and Ensembling
• Section 12: Support Vector Machines
• Section 13: Neural Networks
• Section 14: Deep Learning Architectures
• Section 15: Advanced Deep Learning
• Section 16: Bayesian Machine Learning
• Section 17: Advanced Deep Learning
• Section 18: Multilabel and Multiclass Learning
• Section 19: Meta Learning
• Section 20: Online and Continual Learning
• Section 21: Evaluation and Model Selection
• Section 22: Specialised Applications

Back to AI Courses...