.gif)
AgML, the Machine Learning team of the Agricultural Model Intercomparison and Improvement Project (AgMIP) will have its second workshop in Leipzig, Germany, on November 3-5, 2025!
The effects of climate change have already been felt by the agricultural sector around the world. These impacts are only expected to increase over the coming years with rising temperatures and the increasing frequency and magnitude of extreme climate events. Farmers and policy-makers need accurate yield forecasts, robust climate change impact projections, adaptation and mitigation strategies, crop health monitoring tools and decision support systems. Artificial intelligence (AI) is being rapidly adopted for all of these applications. To effectively make use of new developments in AI, transdisciplinary collaboration is needed to identify and share domain-specific best practices, provide datasets and evaluate novel methods according to diverse stakeholder-led criteria.
AgML launched in 2023, bringing together researchers and practitioners from around the world to share knowledge, to curate open benchmark datasets and standardised evaluation frameworks for different agricultural modelling applications, and to develop new methods adapted to the challenges of this domain. Since the start of AgML, we have published a perspective paper on the need for a transdisciplinary community to build better agricultural models using machine learning. Our first workshop in January 2024 in Wageningen focussed on two activities: a dataset and model evaluation framework for sub-national maize and wheat yield forecasting using observational data from over 25 countries across six continents, and a benchmark dataset challenge which uses process-based simulations to test the ability of data-driven models in projecting future climate change impacts to crop yields.
We want the workshop to be a moment in which we can further these collaborations, but also strengthen and widen our transdisciplinary community and identify new activities and collaborations to address issues not yet addressed. By bringing together data, models and expertise across disciplines, workshop participants will collaborate on dataset improvement and model analysis with the goal of developing protocols to assess the skill of machine learning models for reproducible, intercomparable, generalizable and interpretable modelling of agricultural and food systems.
The workshop will include an informal poster session, in-depth discussion sessions, lightning presentations, hands-on tutorials and a hackathon. We will share an agenda in early July. The workshop will be hybrid, but some sessions may not be possible to attend online, although we will try our best! There will be focussed sessions on subnational yield forecasting, projections of crop productivity under climate change, hybrid crop modelling and good practices for yield modelling with machine learning. Other topics are also of interest - please get in touch if you would like to discuss a potential session or topic! We will also have plenty of time for open-ended discussion and brainstorming.
Attendance is free, but registration is mandatory. In-person participants will be selected based on their profile and interest in the workshop, online participation is open to all. The deadline for registration for in-person attendance is 31st July. Participants are expected to arrange their own transport and cover their accommodation costs. We have a limited amount of funding available to support the travel and accommodation expenses of some participants, which will be allocated according to need and with priority given to those from the Global South.
Workshop committee: Lily-belle Sweet (Helmholtz Centre for Environmental Research - UFZ), Monique Oliveira (Embrapa Agricultura Digital), Oumnia Ennaji (Mohammed VI Polytechnic University), Peng Fu (Louisiana State University)
