Using Ensemble Deep Learning to Deliver Aid Better in Post-Flood Pakistan

Summary

CEGA blog post summarising the poverty mapping research results and their implications. Critically notes that 'the SPSU launched an initiative with the World Bank to strengthen social protection that relies on a multidimensional poverty index instead of poverty prevalence estimates', indicating the government partner chose a different targeting approach rather than adopting the ML-generated poverty maps. Describes the ensemble model results as 'promising' but frames the work as a proof-of-concept rather than an operational deployment.