Syllabus for |
|
| FFR135 - Artificial neural networks |
| |
| Owner: FCMAS |
|
|
5,0 Credits (ECTS 7,5) |
| Grading: TH - Five, Four, Three, Not passed |
| Level: C |
|
Department: 16 - PHYSICS
|
Teaching language: English
| Course module |
|
Credit distribution |
|
Examination dates |
|
Sp1 |
Sp2 |
Sp3 |
Sp4 |
|
No Sp |
| 0100 |
Examination |
5,0 c |
Grading: TH |
|
5,0 c
|
|
|
|
|
|
|
Contact examiner, |
Contact examiner |
In programs
TDATA COMPUTER SCIENCE AND ENGINEERING - Algorithms, Year 4 (elective)
TTFYA ENGINEERING PHYSICS, Year 4 (elective)
TELTA ELECTRICAL ENGINEERING, Year 4 (elective)
TITEA SOFTWARE ENGINEERING, Year 4 (elective)
TITEA SOFTWARE ENGINEERING, Year 3 (elective)
TAUTA AUTOMATION AND MECHATRONICS ENGENEERING, Year 4 (elective)
FCMAS MSc PROGRAMME IN COMPLEX ADAPTIVE SYSTEMS, Year 1 (compulsory)
Examiner:
Professor
Bernhard Mehlig
Eligibility:
For single subject courses within Chalmers programmes the same eligibility requirements apply, as to the programme(s) that the course is part of.
Course specific prerequisites
Basic programming skills
Aim
Neural networks are distributed computational models inspired by the structure of the human brain, consisting of many simple processing elements, connected in a network. Neural networks are increasingly used in many different fields of engineering for tasks such as pattern recognition, prediction and control. The theory of neural networks is a crossdisciplinary field which includes neurobiology, computer science and statistical physics.
The course gives an overview and a basic understanding of neural-network algorithms. Topics covered: associative memory models (Hopfield model), algorithms for learning from examples (perceptrons, back-propagation), and models for self-organisation (Hebbian learning). By learning methods from statistics and computer science students can develop an understanding of when neural networks are useful in application problems.
Content
Course home page
Introduction to neurobiology
Associative memory. The Hopfield model. Optimi-zation problems.
The perceptron
Multilayer perceptrons. Back-propagation and other learning algorithms. Radial basis network.
Introduction to learning theory. Generalization.
Reinforcement learning - learning in situations with infrequent information from the environment.
Self-organization in neural networks. Hebbian learning. Self-organizing feature maps.
Neural network applications.
Classification and prediction using methods from statistics and machine learning.
Organisation
Lectures
In order to develop a thorough understanding of the basic neural network algorithms, students are expected to develop their own implementations (in a language of their chocie), and apply them to simple application problems. This is done through a sequence of small projects (see below).
Literature
J. Hertz, A. Krogh, R.G. Palmer: Introduction to the Theory of Neural Computation (Addison-Wesley, 1991). S. Haykin, Neural Networks: A Comprehensive Foundation, Maxmillan/IEEE Press, 1994.
Examination
Through a sequence of small projects organised into five examples sheets, where students implement the basic models in the course, and use them in application problems.Written presentation of projects.