Multi-modal Deep Learning Based Fusion Approach to Detect Illicit Retail Networks from Social Media

Abstract

Illicit drug trafficking poses a significant threat in today’s society in general and in the younger population in particular. Drug abuse has been known to be correlated with car-accidents, crimes, diseases, deaths and many other negative aspects of society. With the advancement of social media being an open platform to express all kinds of social activities, drug use can be encouraged on this platform to lead other people towards drug abuse. Where people abuse drugs on social media, it is a prominent platform for drug dealing as well. Drug dealers post different types of drug images along with their contact information on social media. Tracking drug dealers among millions of social media users can be very challenging for law enforcement agencies. Therefore, automatic detection of drug dealers (along with the type of drugs they sell and their contact information) is crucial. In this article we have presented a state-of-the-art social media analytic algorithm which does multi modal analysis in order to detect drug-related posts and drug dealers from social media. We propose to detect different types of drugs from social media posts which include: pills, mushrooms, LSD, cannabis, cocaine, syrup, hookah and cigars using our drug type detection model. We also propose to extract drug dealer’s information from social media and create novel AI algorithms to improve understanding of the operations of Illicit Retail Networks (IRN) that will help detect, disrupt and ultimately dismantle these networks. Our approach is generalizable to detect different illicit networks from social media such as human trafficking, illegal gun sales, money laundering and others.