AI on the farm: plant disease diagnosis anytime, anywhere

Sreebin’s ginger leaves are dying, Tanmay asks about rice pathogens, Leka’s tomato leaves are wilting and Adrienne’s fruits are disfigured. They gather at PlantVillage.org asking how to maintain their crops while fending off mites and blights. Most answers come within hours, or days. The site aspires to become a free comprehensive digital library of plant disease images and information stocked by voluntary virtual librarians experienced in plant disease from around the globe. But that is only step one in David Hughes’ and Marcel Salathé’s grand vision to aid growers in need. Using machine learning they plan to automate the whole process to put a plant pathologist in the pocket of every grower worldwide that can diagnose disease within seconds. With hundreds of crops and nearly two thousand different diseases it is an audacious idea. One that puts machine learning to work diagnosing disease and saving lives, rather than directing lethal drones. And there is another benefit of their free open access venture. As big ag increasingly consolidates and may one day control 90% of the available seed, says Hughes “Marcel and I consider the emergence of AI to be very good for global food security but a terrible problem if the data and algorithms are not open and so our food becomes further controlled by big companies.”

I wrote about Hughe’s and Salathé’s work in Natural Defense. Their work was recently awarded a 2017 Inspire Challenge grant along with other innovators. For more see this recent article in Wired magazine.

Advertisements
%d bloggers like this: