Reti neurali

Durata: 28 ore

Periodo didattico: secondo semestre



Settore disciplinare: INF/01


Unico modulo (Prof. Andrea Fusiello).


1. Introduction: Basics of Neural Networks and Support Vector Machines, Basics of Fuzzy Logic Modeling. Approximation of Multivariate Functions, Learning and Statistical Approaches to Regression and Classification

2. Single - Layer Networks: The Perceptron, Convergence Theorem and Learning Rule. The Adaptive Linear Neuron and the Least Mean Square Algorithm.

3. Multilayer Perceptrons: The Error Backpropagation Algorithm, The Generalized Delta Rule, Heuristics or Practical Aspects of the Error
Backpropagation Algorithm.

4. Radial Basis Function Networks:Ill - Posed Problems and the Regularization Technique, Stabilizers and Basis Functions, Generalized Radial Basis Function Networks.

5. Fuzzy Logic Systems: Basic Set Operations, Fuzzy Relations,   composition of Fuzzy Relations, Fuzzy Inference,Compositional Rule of Inference,Defuzzification.

6. Support Vector Machines: Risk Minimization Principles and the Concept of Uniform Convergence, The VC Dimension, Structural Risk Minimization, Support Vector Machine Algorithms.


Gli argomenti 5 e 6 verranno svolti opzionalmente in funzione di tempo ed



V.Kecman. Learning and Soft Computing: Support Vector Machines, Neural

Networks and Fuzzy Logic Models. MIT Press, 2001.




Elenco corsi 2013/2014