Machine Learning + Satellite Pics = Knowledge

Stanford University researchers calculated poverty down to the village level in five African countries – without using any poverty data.Poverty levels for Nigeria, Uganda, Tanzania, Rwanda and Malawi were predicted down to the village level Stanford University’s Sustainability and Artificial Intelligence Lab recently concluded estimates of poverty levels down to the village levels for five African countries: Nigeria, Uganda, Tanzania, Rwanda and Malawi (see map, with the most poor areas coloured red). The results were published on 19 August in Science magazine.    What is remarkable is the way the researchers arrived at their estimates. They did not use any of the conventional data covering incomes, consumption etc. as these data are not available at the village level of these countries in the first place. Instead, they used a combination of satellite imagery and machine…


Link to Full Article: Machine Learning + Satellite Pics = Knowledge