Accés ràpid intranet

Més informació...

a a a

Deim Seminar


Application of machine learning methods in computer and network security


Saradha Ravi

Professor/a organitzador/a

Josep Domingo Ferrer


Indian Institute of Science, Bangalore


19-12-2011 12:00


The area of computer and network security is of prime importance today with the growth in the rate and the sophistication of the threats employed by the attackers. A lot of data mining and machine learning have been employed in the past to problems in computer security. We specifically aim to look at the area of malware analysis in the talk. The techniques of code obfuscation render most of the signature based detection methods useless. We look at ways of using dynamic analysis to classify and cluster malware into families. The use of profile HMMs for this problem is discussed here. We also look at the anomaly detection methods employed in network security. Newer methods like active learning and other feature selection methods are discussed for the same problem to address speed and other issues like change in the concept of normality.


Laboratori 231