Public articles linked to the same research event.
The latest research from Google Using Google Earth AI's Population Dynamics Foundation Model (PDFM) as a proof-of-concept, this work integrates self-supervised pretrained representations of place as plug-and-play inputs into existing epidemiological workflows, independently evaluated by external partners across five public-health tasks: cross-border measles vaccination coverage modeling rose from 16% to 22% explained variation, current-year cardiovascular death nowcasting showed no statistically significant difference from census inputs, one-month-ahead dengue forecasting improved accuracy in up to 72% of active transmission municipalities, postpartum depression risk prediction gained AUC +0.0038 in unseen states, and cholera hotspot forecasting helped at four to eight weeks ahead.
Using Google Earth AI's Population Dynamics Foundation Model (PDFM) as a proof-of-concept, this work integrates self-supervised pretrained representations of place as plug-and-play inputs into existing epidemiological workflows, independently evaluated by external partners across five public-health tasks: cross-border measles vaccination coverage modeling rose from 16% to 22% explained variation, current-year cardiovascular death nowcasting showed no statistically significant difference from census inputs, one-month-ahead dengue forecasting improved accuracy in up to 72% of active transmission municipalities, postpartum depression risk prediction gained AUC +0.0038 in unseen states, and cholera hotspot forecasting helped at four to eight weeks ahead.
Using Google Earth AI's Population Dynamics Foundation Model (PDFM) as a proof-of-concept, this work integrates self-supervised pretrained representations of place as plug-and-play inputs into existing epidemiological workflows, independently evaluated by external partners across five public-health tasks: cross-border measles vaccination coverage modeling rose from 16% to 22% explained variation, current-year cardiovascular death nowcasting showed no statistically significant difference from census inputs, one-month-ahead dengue forecasting improved accuracy in up to 72% of active transmission municipalities, postpartum depression risk prediction gained AUC +0.0038 in unseen states, and cholera hotspot forecasting helped at four to eight weeks ahead.
Using Google Earth AI's Population Dynamics Foundation Model (PDFM) as a proof-of-concept, this work integrates self-supervised pretrained representations of place as plug-and-play inputs into existing epidemiological workflows, independently evaluated by external partners across five public-health tasks: cross-border measles vaccination coverage modeling rose from 16% to 22% explained variation, current-year cardiovascular death nowcasting showed no statistically significant difference from census inputs, one-month-ahead dengue forecasting improved accuracy in up to 72% of active transmission municipalities, postpartum depression risk prediction gained AUC +0.0038 in unseen states, and cholera hotspot forecasting helped at four to eight weeks ahead.