Tracking Illicit Drug Dealing and Abuse on Instagram using Multimodal Analysis

TL;DR

This paper presents a multimodal analysis framework using multitask learning and decision-level fusion to automatically detect and analyze drug-related posts and user behavior on Instagram, aiding law enforcement and public health efforts.

Contribution

It introduces a novel multimodal analysis approach combining multitask learning and fusion techniques for drug detection on social media, improving scalability and reproducibility.

Findings

Effective detection of drug-related posts demonstrated

Framework scalable to large social media datasets

Outperforms conventional labor-intensive methods

Abstract

Illicit drug trade via social media sites, especially photo-oriented Instagram, has become a severe problem in recent years. As a result, tracking drug dealing and abuse on Instagram is of interest to law enforcement agencies and public health agencies. In this paper, we propose a novel approach to detecting drug abuse and dealing automatically by utilizing multimodal data on social media. This approach also enables us to identify drug-related posts and analyze the behavior patterns of drug-related user accounts. To better utilize multimodal data on social media, multimodal analysis methods including multitask learning and decision-level fusion are employed in our framework. Experiment results on expertly labeled data have demonstrated the effectiveness of our approach, as well as its scalability and reproducibility over labor-intensive conventional approaches.