Towards Data Science author Oleksii Kharkovyna wrote an article about the evolving role of machine learning in agriculture. Kharkovyna referred to machine learning as the fourth agricultural revolution.
There are numerous applications of machine learning in this sector. These include:
- Crop monitoring
- Automated irrigation controls
- Greenhouse climate controller
- Animal health monitoring
The benefits of machine learning in the agriculture sector cannot be overstated. A recent report shows that the artificial intelligence market for farmers and ranchers around the world is expected to reach $2.9 billion within the next five years.
Machine Learning is the Fourth Agriculture Revolution
It is interesting to see how machine learning has played a role in the evolution of agriculture. If we could go back in time and ask the inhabitants of the first civilizations for an example of what a machine means to them, they would probably show us a couple of rudimentary agricultural tools. Today, modern machinery is far from that reality due to its constant technological evolution that would be unimaginable in the eyes of the first peoples that inhabited the Earth.
Today, the speed at which machines are perfected is such that they even have the capacity to learn autonomously the tasks for which they were designed. This technology is called Machine Learning.
The term was created in 1959 by Arthur Lee Samuel, a pioneer in the area of computer games and Artificial Intelligence (AI). At that time, specialists from different branches within AI sought to create systems that would learn automatically through data. However, due to some technological restrictions, the real growth of this discipline took place in the 1990s thanks to the advances of the Digital Revolution.
The field has millions of pieces of data that AI can transform into information in order to favor the accuracy of the producer’s decision making and, thus, maximize production in a sustainable way. Not long ago, data was just notes that were kept in a drawer, since it could not be transformed into concrete actions due to lack of technology. Now, producers can make the most of each hectare thanks to the precise information provided by innovative hardware and software systems.
Thus, today’s agricultural machines are capable of performing much more than the basic functions of sowing, harvesting and applying inputs. They are equipped with sensors through which soil moisture levels, light intensity and pest threats can be measured in real time. As more information is collected, it returns to the machine itself which has the ability to gradually adjust operations to optimize agricultural yields. As a result, mathematical prediction models in some contexts exceed the capacity of humans themselves, because they help provide the best options for increasing production.
An example of this is the nutritional supplementation of soil with nitrogen that can be done more precisely to care for the environment and ensure higher levels of production. The georeferencing of the machines comes from automatic learning that allows the customization of agricultural tasks. Today, the same person in front of a steering wheel has the option to make different use of his time, while a modern machine guarantees the care of his harvest.
It is believed that 70% of crop productivity is the result of decisions made by professionals in the field. Machines capable of learning from experience join with growers to help them achieve better levels of efficiency in their crops. In traditional agriculture, the farmer was guided by his intuition; in modern agriculture, he uses drones and sensors. What will the machines of the future be like?