Poverty mapping has proven to be a difficult task in past years. Poor countries are often reluctant to account for poverty due to corruption or the inability to do so because of ongoing conflicts. The World Bank reports that only 20 African countries conducted two or more population surveys on poverty from 2000 to 2010.
A new study from Stanford University hopes to improve poverty mapping by combining high-resolution satellite imagery with artificial intelligence.
According to a feature article published by online tech magazine Motherboard, Neal Jean, a Ph.D. engineering student at Stanford, has designed a machine learning algorithm that can predict poverty in Malawi, Nigeria, Rwanda, Tanzania, and Uganda.
Using satellite imagery to determine “nightlights” and levels of economic activity as a method of poverty mapping is nothing new. What’s different about the algorithm designed by Jean and his team is that it looks at daylight images of infrastructure, such as roads and metropolitan areas, which it then uses to identify nighttime patterns.
“Our basic approach involved a machine learning technique called ‘transfer learning,’ which is the idea that you can solve a hard problem – in our case, predicting poverty from satellite images – by trying to solve an easier one,” Jean said.
According to Motherboard, the algorithm may prove to be a very effective method of poverty mapping, especially given the cost of traditional household surveys and the lack of viable alternatives. Another advantage of the machine learning model is its transparency, as it doesn’t rely on private or protected information.
Jean told Motherboard that he hopes to make the technology open-source and cooperate with NGOs to put the algorithm to use. “If we could provide them with high-resolution poverty maps, they could overlay them on regions where operations already exist, and ultimately inform where they distribute funding,” he argued.
Jean’s machine learning algorithm is not the only artificial intelligence tool that is providing better data for poverty alleviation efforts. South African computer scientist Muthoni Masinde developed a solution that can forecast droughts with 98 percent accuracy, combining traditional knowledge with new technologies. In recognition of her achievements, she received a Distinguished Young Women Researcher award at the 2016 South African Women in Science Awards.
Technological advance has been the greatest impetus for poverty reduction throughout history, and artificial intelligence is the future of poverty mapping. It provides economists and scientists with better data in order to pinpoint and resolve problems that are holding developing countries back.
– Philip Katz