Food safety is a growing concern, and the SemEval 2025 Task 9—The Food Hazard Detection Challenge—aims to address it through the power of AI. This challenge invites AI & NLP enthusiasts, researchers, and developers to create innovative systems that can potentially identify and explain food-related risks from online sources.
Participants will tackle two sub-tasks:
- Text Classification for Food Hazard Prediction (ST1): Predict the type of food hazard and the affected product based on text data.
- Food Hazard and Product “Vector” Detection (ST2): Provide precise predictions of the exact hazard and product involved.
A key aspect of this challenge is explainability, ensuring that models not only predict hazards but also clearly explain their reasoning. This transparency is crucial for trust and practical application in food safety.
Participants will progress through a well-structured timeline, from initial trials to the final evaluation. You’ll have access to a rich dataset and the chance to test your models against others in a competitive environment. At the end of the journey, your findings will contribute to cutting-edge research, with a final paper submission.
The challenge is part of the EFRA project and it is organized by our partners, Stockholm University and Agroknow. Don’t miss this opportunity to contribute to the field of food safety while advancing explainable AI systems.
