WUR Leads European Project For AI That Better Understands Agricultural Systems

AI can help farmers, growers, and researchers make better decisions. However, the potential of AI in agriculture is still only being utilized to a limited extent. In the AgriScienceFM project, Wageningen University & Research is developing foundation models with European partners that combine data more effectively and make agricultural AI more reliable for practice, research, and policy.

Smart soil and water management, autonomous harvesting robots, disease and pest control: AI is already visible in agriculture. Yet, its application on a large scale is difficult. According to Ioannis Athanasiadis, Chair Holder of AI at WUR, there are several reasons for this. “Agricultural systems naturally differ significantly from one another. You are dealing with diversity in crops, soils, climatic conditions, and available tools. As a result, the AI ​​solutions that do exist often work only once or only under specific circumstances.”

Athanasiadis also points out the multidisciplinary nature of agriculture. “The unique and at the same time challenging aspect of agriculture is that it is an interaction between humans, nature, and biology. As a result, you have to connect knowledge and information from various fields, such as water and soil systems, genomics, the environment and climate, and diseases and pests. Currently, models are built and data collected separately in each discipline. Consequently, the complete picture needed for AI that works well under different agricultural conditions is often still missing.”

Foundation Models For Agriculture
According to Athanasiadis, foundation models are needed to break through this fragmentation: basic models that can combine large amounts of data from various sources. Within the AgriScienceFM project – launched in June – researchers will work on three coherent foundation models centered on the core of agricultural systems: biological material, such as plants and animals; the natural environment, such as soil, water, and climate; and human activity, such as management and crop choices.

Athanasiadis: “These models form the basis for tools that enable farmers to make well-founded decisions in a variety of circumstances based on combined datasets. This yields benefits not only for farmers themselves, but also for nature and society, such as sustainable and more efficient food production, a healthier environment, and better climate adaptation.”

European project
AgriScienceFM is a Horizon Europe project for AI in agricultural sciences. The project aligns with the European commitment to make stronger use of AI in scientific domains as well. WUR coordinates the project. According to Athanasiadis, WUR possesses both extensive domain knowledge and expertise in the field of AI. “That combination makes us unique in the world. At the same time, there are many other institutes with valuable expertise. That is why we are joining forces in a consortium. We are collaborating with universities and institutes from Greece, Germany, Spain, the UK, and Belgium, among others.”

Strategic Autonomy
Liesbeth Luijendijk, who is involved in the integration of Robotics & AI in the agro- and food sector from WUR, mentions another reason why this European cooperation is important: our strategic autonomy. “Due to all geopolitical developments, it is necessary that we, as Europe, are as little dependent as possible on major powers such as China and the US. Therefore, it is essential that we develop knowledge and technology in the fields of agriculture and AI within the EU.” According to her, AgriScienceFM shows that the European Commission recognizes this necessity and sees the importance of strong European cooperation on this important subject.

It is also important for the Netherlands to position itself in this area, Luijendijk continues. “In the national investment and innovation agenda Food 2040, in which we at WUR are involved, we agreed that we, as the Netherlands, want to remain a frontrunner when it comes to high-tech and sustainable food production. If we want to achieve this, we really must invest in AI. We cannot afford to rely solely on domain expertise and leave the AI ​​part to other parties or countries. We must ensure that we have both in-house.”

From Satellite To Soil Advice
In the first half of the three-year project, the researchers will gather and align existing public datasets: from satellite images and weather data to field measurements, livestock farm data, crop data, and genetic material. The models will be tested in concrete applications, such as satellite monitoring of crops and water scarcity, local soil advice for farmers, faster breeding of resilient crops, and precision agriculture regarding diseases, pests, and animal health.

AgriScienceFM thus focuses not only on new AI models, but also on the question of how well those models work for real agricultural challenges. To this end, the consortium is also developing so-called benchmarks: tests that allow researchers to assess whether a model yields usable results in the field. Athanasiadis: “AI only gains real value when the outcomes are reliable across different regions, cultivation systems, and practical situations.”

Bridge Between AI And Agricultural Research
With AgriScienceFM, Athanasiadis hopes not only to develop usable foundation models but also to build a bridge between AI and agricultural scientists. “Many scientists are now aware of the importance of AI but do not always have sufficient knowledge to make good use of its possibilities. That is why the project also includes a knowledge-sharing component. With AgriScienceFM, we offer an environment in which we create a shared understanding of why agrifood challenges and AI are inextricably linked and how AI developments can offer solutions to existing problems.”