Interpretable Graph-Language Modeling for Detecting Youth Illicit Drug Use

2025-10-11

AI Summary

Existing adolescent and young adult (TYA) illicit drug use detection methods treat survey variables as independent features, ignoring their underlying semantic interdependencies. To address this limitation, we propose LAMI—a novel framework that synergistically integrates graph structure learning with large language models (LLMs) for youth substance use risk modeling. LAMI employs a learnable graph layer to automatically construct a semantic association graph among individual survey responses, fuses heterogeneous data from the Youth Risk Behavior Survey (YRBS) and the National Survey on Drug Use and Health (NSDUH), and jointly leverages graph neural networks (GNNs) and LLMs for behavioral pattern discovery and natural-language explanation generation. Evaluated on multiple real-world datasets, LAMI significantly outperforms state-of-the-art baselines, achieving substantial improvements in classification accuracy. Moreover, it provides interpretable identification of critical risk pathways—including family environment, peer influence, and academic stress—yielding actionable insights for public health interventions.

Technology Category

• Machine Learning: Large Multimodal Models (LMMs)
• Natural Language Processing: (Large) Language Models
• Data Mining & Knowledge Management: Graph Mining, Social Network Analysis & Community

Application Category

• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation

Abstract

Illicit drug use among teenagers and young adults (TYAs) remains a pressing public health concern, with rising prevalence and long-term impacts on health and well-being. To detect illicit drug use among TYAs, researchers analyze large-scale surveys such as the Youth Risk Behavior Survey (YRBS) and the National Survey on Drug Use and Health (NSDUH), which preserve rich demographic, psychological, and environmental factors related to substance use. However, existing modeling methods treat survey variables independently, overlooking latent and interconnected structures among them. To address this limitation, we propose LAMI (LAtent relation Mining with bi-modal Interpretability), a novel joint graph-language modeling framework for detecting illicit drug use and interpreting behavioral risk factors among TYAs. LAMI represents individual responses as relational graphs, learns latent connections through a specialized graph structure learning layer, and integrates a large language model to generate natural language explanations grounded in both graph structures and survey semantics. Experiments on the YRBS and NSDUH datasets show that LAMI outperforms competitive baselines in predictive accuracy. Interpretability analyses further demonstrate that LAMI reveals meaningful behavioral substructures and psychosocial pathways, such as family dynamics, peer influence, and school-related distress, that align with established risk factors for substance use.

Problem

• Detecting youth illicit drug use from survey data
• Modeling latent connections between survey variables
• Interpreting behavioral risk factors with explanations

Innovation

• Joint graph-language modeling for drug use detection
• Learns latent connections via graph structure learning
• Generates explanations using graph structures and semantics