"Today we generate multiple quintillion bytes of data per day. The rate of the data creation is so high that 90% of the data in the world today has been created in the last two years alone. This rapid acceleration in the information production information has lead to a two fold need: a) protecting the sensitive data; and b) creating new technologies to filter and analyze this data. Machine learning techniques has long been used to analyze large amounts of data in myriad domains including for security and privacy (spam filtering, anomaly and intrusion detection, and so on).
In this talk, I am going to talk about a large-scale data analytics framework for behavioral malware classification. It comprises program analysis, fast feature hashing, and SVM-based learning techniques to model malware behavioral patterns. The system combines continuous data collection from heterogeneous sources (the end host systems, in-house and publicly available malware data) with frequent and automated retraining to efficiently and accurately classify incoming malware samples.
Rekha is Research Scientist at Intel Science and Technology Center for Secure Computing. She received her PhD from Rutgers University in 2011, and her Bachelors of Engineering from National Institute of Technology, Surat in 2000. Her research interests are in the areas of system security, secure analytics, operating systems and distributed systems."