Danilo Sarti*, Alessandra Lemos*, Rafael Moral*, Andrew Parnell*
Most of us use machine learning in our daily lives without ever realising it. We watch series or movies that have been automatically recommended to us, and use applications to choose restaurants or which concert to go to on the weekend. But what does this have in common with farmers’ decisions or the regulation of crop varieties?
Machine learning is the study of routines (algorithms) which improve automatically using data provided to the system. In other words, the data provided by other users for video streaming or restaurant recommendations helps the algorithm guess the most appropriate item to recommend to us. Machine learning techniques are crucial in the context of a world where data is becoming more and more available.
The InnoVar project uses machine learning techniques to develop new systems that use data collected in experiments conducted across multiple countries to evaluate newly developed wheat varieties. Such data is passed to machine learning applications to provide a recommendation of the variety to plant in a certain location to farmers.
InnoVar’s machine learning solutions will work as a recommendation application to enable farmers to know which wheat varieties most suit their needs and farm environment, including aspects of soil types and disease resistance. The same application can also serve regulators when deciding or not to approve new varieties.
Machine learning techniques are a substantial contribution of the InnoVar project for establishing digital agriculture in Europe, helping farmers reduce their risks of production and optimize their allocations of resources, minimising environmental impacts.
* Hamilton Institute/Mathematics and Statistics Department, National University of Ireland, Maynooth.