This information sheet indicates how the course will be organized at pandemic code level yellow and green.
If the colour codes change during the academic year to orange or red, modifications are possible, for example to the teaching and evaluation methods.

Artificial Neural Networks

Course Code :2000WETANN
Study domain:Physics
Bi-anuall course:Taught in academic years starting in an odd year
Academic year:2020-2021
Semester:2nd semester
Contact hours:30
Study load (hours):84
Contract restrictions: No contract restriction
Language of instruction:Dutch
Exam period:exam in the 2nd semester
Lecturer(s)Paul Scheunders

3. Course contents *

The course is about the use of artificial neural networks, in particular in the domain of statistical pattern recognition. After a thorough introduction on the domain of statistical pattern recognition, feedforward neural networks are treated.

1. Introduction: biological neural networks
2. Statistical Pattern Recognition
2.1 An example: character recognition
2.2 Classification and regression
2.3 Pre-processing and feature extraction
2.4 The curse of dimensionality
2.5 Polynome fitting
2.6 Model complexity
2.7 Multivariate non-linear functions
2.8 Bayes theorem
2.9 Decision planes

3. Probability density estimation
3.1 Parametric methods
3.2 Maximum Likelihood
3.3 Bayesian Inference
3.4 Sequential parameter estimation
3.5 Non-parametric methods
3.6 Mixture models

4. The single layer network
4.1 Linear discriminant functions
4.2 Linear separability
4.3 Least Squares techniques
4.4 The perceptron
4.5 Fisher's linear discriminant

5. Multilayer networks
5.1 Forward network mappings
5.2 Threshold neurons
5.3 Sigmoid neurons
5.4 Weight space symmetry
5.5 Higher ord networks
5.6 Projection Pursuit
5.7 Error back-propagation
5.8 Jacobian
5.9 Hessian