Machine Learning, along with the other technology that is coming out of the fourth industrial revolution, has consistently proven that they have great power and influence as their application can make or break a project’s success. While there are many industry experts, such as Stephen Hawking and Elon Musk, having completely polar opposite views on what these new technologies should be used for, there are many examples of these technologies being used for solving some genuinely important problems such as fighting air pollution, bringing efficiency in the processes that involve lots of non-renewable resources, and now to combat starvation and malnutrition.

US Agency for International Development has recently given a 3-year grant of $1 million to a group of researchers from the Cornell University who are working on a machine learning project to analyse agricultural and food market conditions more efficiently.

The aim of this project is to collect information directly from satellite and other channels to predict the situations that could lead to poverty and undernutrition in some of the world’s poorest countries.

To start off, the project will make use of open-source satellite data for measuring solar-induced chlorophyll fluorescence (SIF) which are photons that the plants in the target area emit. This information will be then used to monitor agricultural productivity. Information about land-surface temperature shall also be processed to gain insights on crop stress situations under the conditions of water deficit or excessive heat exposure.

This will all then be consolidated with market food-price to create a comprehensive report and forecast contributing conditions, inducing poverty, in areas where most of the economy depends upon agricultural outputs.

Chris Barrett, the principal investigator on the grant, has said, “A method that can use near real-time, low-cost or freely available remotely sensed data can speed up the delivery of this information, and sharply reduce the cost” He added, “If you are a humanitarian organization trying to really target your resources at the poorest rural areas, this seems a powerful diagnostic tool.”

This method, when completed, hopes to bring enormous efficiency in the research work of humanitarian organizations, which at present have to rely on the manual tasks of routinely drawing lines on the map and then collecting data through surveys. Although this method does have disadvantages as these surveys are time-consuming and unreliable on metrics that change more dynamically.  As well as this, without the digitization of the data collected, the processing of all the information takes so long that decisions that are made from their results are often considered irrelevant. But with a system that is digital from the ground up, with technologies like Big Data Analytics and Machine learning, efficiency can finally be brought into the system, which will further promote resource optimization.

Regarding these new methods, Mr. Barrett has said, “These sorts of measures can give you the underlying data with which to do rigorous impact evaluation.” He explained, there are a lot of policies involved in ensuring poverty reduction and improving child nutrition. Data from this method can also be used to evaluate how effective those policies are, and if newer ones need to be implemented.

Resources:

http://news.cornell.edu/stories/2018/10/harnessing-machine-learning-and-big-data-fight-hunger