AI research applied to challenges like food security and health care.
Both IBM Research’s offices in Kenya and South Africa and Google’s AI lab in Ghana share the same mission as their parent organizations: to pursue fundamental and cutting-edge research. They focus on issues like increasing access to affordable health care, making financial services more inclusive, strengthening long-term food security, and streamlining government operations. The list is not unlike that for an AI research lab located anywhere else in the world, but the context adds nuance to the objectives.
“Research cannot be detached from the environment in which it is performed,” says Moustapha Cisse, the director of Google AI Ghana. “Being in an environment where the challenges are unique in many ways gives us an opportunity to explore problems that maybe other researchers in other places would not be able to explore.”
Before founding its AI lab in Ghana, for example, Google began working with farmers in rural Tanzania to understand some of the struggles they faced in maintaining consistent food production. The researchers learned that crop disease can significantly reduce yield, so they created a machine-learning model that could diagnose early stages of disease in the cassava plant, an important staple crop in the region. The model, which works directly on farmers’ phones without needing access to the internet, helps them intervene earlier to save their plants.
Wayua gives another example. In 2016, the Johannesburg team at IBM Research discovered that the process of reporting cancer data to the government, which used it to inform national health policies, took four years after diagnosis in hospitals. In the US, the equivalent data collection and analysis takes only two years. The additional lag turned out to be due in part to the unstructured nature of the hospitals’ pathology reports. Human experts were reading each case and classifying it into one of 42 different cancer types, but the free-form text on the reports made this very time-consuming. So the researchers went to work on a machine-learning model that could label the reports automatically. Within two years, they had developed a successful prototype system, and they are now striving to make it scalable so it can be useful in practice. [read more]
Source: MIT Technology Review