Lecture | March 6 | 4:10-5:30 p.m. | 202 South Hall
Recent advances in deep learning and satellite imagery make it possible to remotely monitor economic and agricultural trends across the developing world, at high resolution. These advances are now being translated from research labs into the real world.
This seminar will discuss the technology and vision behind Atlas AI, a Bay Area start-up that spun out of Stanford University in 2018. Co-founded by Berkeley alum Marshall Burke (Ph.D. â14) and established in partnership with Rockefeller Foundation, the company has a unique focus on social impact and the sustainable development goals. It is creating a hybrid business model that supports the generation of public goods, including open datasets for use by researchers and non-profits. Atlas AI's analytics platform currently delivers village-scale estimates of household well-being, alongside plot-level forecasts of smallholder crop yields, for all of sub-Saharan Africa. Through beta testing and early deployments, the team is exploring issues related to privacy, bias, accountability, and transparency in machine learning.