AI maps disease across your everyday social worlds

What is it about?

Imagine living in multiple social worlds every day. At your residence, family members are present, in your career, you work with a group, in your educational environment, friends are there, and within your neighborhood, you create connections at various places. Every universe follows its distinct regulations, includes diverse people, and links to others meaningfully for your wellness and health. Our research developed multilayer modular fusion graph attention network (MMF-GAT), an AI framework that predicts how diseases spread through these overlapping social layers. It works by mapping your memberships in different groups (e.g., living with family members or attending the same school) and accounting for the roles of individuals who act as brokerages between separate worlds, potentially carrying infections from one circle to another.

Featured Image

Photo by Dynamic Wang on Unsplash

Photo by Dynamic Wang on Unsplash

Why is it important?

The framework represents overlapping group affiliations with their distinct interaction patterns across different social contexts, then quantifies how much each factor contributes to prediction performance. This approach extends beyond epidemic forecasting to any complex system with multilayered, modular network structures, opening up applications across various fields, including biological systems, healthcare networks, organizational dynamics, and consumer behavior.

Perspectives