 |
Unsupervised learning |
| |
|
10,5h L / 3h R / 1 WE + 1 OE / 4 credits ECTS in common with IIC_OAD2 and IIC_OAD3 / IIC_OAD1 |
| |
|
Michel IANOTTO and Hervé FREZZA-BUET |
| |
|
The aim of this course is to introduce the techniques of unsupervised learning, where the information learned is extracted from data distribution, without any external input. In the preliminaries, a general introduction for the different learning paradigms is given so that the next courses about “Supervised Llearning” and “Finite State Automata for Probablistic AI” may be linked up with this one.
|
| |
|
Introduction (1h30) |
| |
|
In this part, we will expose the generic problem of learning and we will come with three different flavors of it: unsupervised learning, supervised learning and reinforcement learning. The concept of overfitting will be presented in an informal manner too.
|
| |
|
Data distribution analysis (9h) |
| |
|
The unsupervised learning paradigm covers cases where learning occur without any external input. In this case, learning comes down to describing data. Methods presented will include neural networks (self-organizing maps, incremental networks), principal components analysis and decision trees.
|
| |
| |
| |
| |
|
References
B. Fritzke, http://www.neuroinformatik.ruhr-uni-bochum.de/ini/VDM/research/gsn/JavaPaper/. |
| |
| |
| Last update 06/07/2007 by Cl.M. |
