Sites Inria

Version française

Rencontres Inria Industrie

Équipe Tamis (Inria Rennes-Bretagne Atlantique) -

Data Thieves Detection via semantic analysis

© INRIA Sophie Auvin - T comme Temps

One of the current threat on the privacy and confidentiality of data is the possibility of data theft through malware infection. In our demonstration, we show how our malware detection tool chain can be used to detect malwares and prevent exfiltration of data after an infection. Methods based on syntactic properties become ineffective as soon as the binary is slightly mutated or obfuscated. We'll show that our method, based on behavioral analysis, is able to detect even mutated and obfuscated malware, thus protecting user data from avanced infections.

Keywords: Learning Malware Symbolic analysis