RESEARCH · RESEARCH · #1594
Google Earth AI research prototypes (PDFM, Geospatial Reasoning agent) used to support disease forecasting and Ebola response
Google Research describes new papers and research prototypes under Google Earth AI — including the Population Dynamics Foundation Model (PDFM), a Geospatial Reasoning agent, and a planetary prediction engine — that combine satellite imagery, mobility and environmental data, and foundation models to augment public health forecasting. Partners including WHO AFRO and DRC’s INRB used the prototypes to map exposed settlements and forecast Ebola spread, while academic collaborators applied PDFM to predict cardiovascular mortality, vaccination coverage, dengue, and cholera hotspots.
KEY POINTS
- Google Research describes new papers and research prototypes under Google Earth AI — including the Population Dynamics Foundation Model (PDFM), a Geospatial Reasoning agent, and a planetary prediction engine — that combine satellite imagery, mobility and environmental data, and foundation models to augment public health forecasting.
- Partners including WHO AFRO and DRC’s INRB used the prototypes to map exposed settlements and forecast Ebola spread, while academic collaborators applied PDFM to predict cardiovascular mortality, vaccination coverage, dengue, and cholera hotspots.
- These research prototypes show how geospatial foundation models can fill reporting gaps and give health authorities faster, actionable forecasts — demonstrated by WHO AFRO and DRC partners during an active Ebola response.
WHY IT MATTERS
These research prototypes show how geospatial foundation models can fill reporting gaps and give health authorities faster, actionable forecasts — demonstrated by WHO AFRO and DRC partners during an active Ebola response.