学术报告(Fabio Roli)

Short Course

Introduction to Multiple Classifier Systems

Lecturer：Prof. Fabio Roli
University of Cagliari
Dept. of Electrical and Electronics Eng., Italy
 

Brief Description of course contents and objectives：

In the field of pattern recognition, fusion of multiple classifiers is current
ly used for solving difficult recognition tasks and designing high performance
 systems. From a theoretical viewpoint, fusion of multiple classifiers allows
overcoming the well-known limitations of classical approach to design a patter
n recognition system that focuses on the search of the best individual classif
ier. From a practical viewpoint, the concept of multiple classifiers derives n
aturally from the context and requirements of many applications. This 6 hours
course will illustrate the basic concepts and motivations of multiple classifi
er systems and will present the main methods and algorithms for creating and c
ombining classifier ensembles. Some applications of multiple pattern classifie
rs will be discussed, with a particular focus on security applications (e.g.,
spam filtering). In the last two hours of the course, the basic concepts of ad
versarial pattern recognition will be introduced, and the use of multiple clas
sifier systems to make a pattern recognition system harder to evade will be di
scussed.
 
For details: http://lamda.nju.edu.cn/opense/CourseOfFabioRoli.htm