Google Research Introduces ME-POIs: A Mobility-Informed Framework that Adds "How a Place Is Used" to Text-Based POI Embeddings

Google Research and USC introduced ME-POIs, which combines text-based POI embeddings with aggregated mobility data.

The framework captures real usage patterns—such as visit intent and busyness—rather than place descriptions alone.

It reported strong gains over text-only embeddings on map-enrichment tasks in Los Angeles and Houston.

The approach could improve map quality, business intelligence, and place prediction, though no code or model weights were released.