COURSE SYLLABUS
Section 1: Introduction to Learning Theory / Εισαγωγικές Έννοιες Θεωρίας Εκμάθησης
- Learning machines
- Pattern recognition systems
- The system design cycle
- Learning and adaptability
Section 2: Bayesian Decision Theory / Μπεϋζιανή Θεωρία Αποφάσεων
- Classifiers, discrimination functions, decision surfaces
- Classification of minimum probability of error
- (Gaussian) Bayes classifier for binary classification
- Classification error probability
- Bayesian belief networks
Section 3: Parameter Estimation and Supervised Learning / Εκτίμηση Παραμέτρων και Εκμάθηση με Επιτήρηση
- Maximum likelihood estimation
- Bayesian estimation
Section 4: Non-Parametric Techniques / Μη Παραμετρικές Τεχνικές
- Distribution estimation and Parzen windows
- k-Nearest neighbors
- Methodologies for reducing the number of patterns
Section 5: Linear Discrimination Functions / Γραμμικές Συναρτήσεις Διάκρισης
- Linearly separable classes
- Perceptron Algorithm
- Least squares method
Section 6: Artificial Neural Networks (ANN) / Τεχνητά Νευρωνικά Δίκτυα (ΤΝΔ)
- Introduction and structure
- Recursive ANN
- Back-propagation algorithm
- Multilayer Perceptron
Section 7: Stochastic Methods and Unsupervised Learning / Στοχαστικές Μέθοδοι και Εκμάθηση Χωρίς Επίβλεψη
- Sample density and recognizability
- Recursive ANN
- Unsupervised learning with Gaussian sample density
- Data grouping