WhiteRabbit Combining Threat Intelligence Public Blockchain
Machine learning model using Bitcoin transaction data to identify illegal cryptocurrency addresses.
Summary
Cryptocurrency Crime
Bitcoin transactions involving illegal activities can be identified through analysis of transaction patterns and connected addresses.
Machine Learning Approach
A machine learning model merges cyber-related information with BTC transaction data. Both private and public sector security professionals can use it to identify criminal activity or build legal cases for stolen coin recovery.
Dataset Construction
Identifying BTC addresses involved in illegal activity provides starting points. Navigation along the chain reconstructs clusters of connected "dirty" addresses using FIFO rules to label them.
Behavioral Classification
Characteristic behavioral analysis determines malicious features for machine learning classification of new addresses. A three-part framework collects clean/dirty address datasets, tests classification models, and predicts address classification.
Related talks
RSAC Cybersecurity·53 min
Analysts investigate attack surfaces and defenses concerning Machine Learning and the Cloud: Disrupting Threat Detection and Prevention.
AI Village·27 min
Andy Applebaum
Using automated planning and AI to improve red team offensive operations.
AI Village·15 min
Ivan Torroledo
Threat detection using defensive machine learning techniques.