Malware Clustering - Static and Dynamic Behavior Analysis

Explore a novel approach to malware clustering in this 49-minute conference talk from LASCON 2017. Dive into unsupervised similarity search techniques that group similar malwares together based on their static and dynamic behavior. Learn about the preprocessing stages involving classic machine learning approaches, and discover how this method proves to be robust, scalable, and repeatable on large datasets. Cover topics such as Jaccard similarity, min hashing and encoding, locality-sensitive hashing, and hybrid approaches. Gain insights into cluster evaluation and feature evaluation techniques for effective malware analysis.

Introduction
Need for clustering malware
What is done today
Jaccard similarity
Min hashing and encoding
Localitysensitive hashing
Algorithm
Summary
Hybrid Approach
Cluster Evaluation
Feature Evaluation