Last week, the EFRA consortium travelled to Pisa, Italy for the first plenary meeting of the project. The event was hosted by CNR (National Research Council) and it took place at the beautiful place of Officine Garibaldi. Representatives from all our nine partners came together to discuss the progress, challenges, and future plans of EFRA.
The primary objective of the event was to enable partners to collaboratively establish goals and determine their respective contributions for the upcoming months. During the first day, our use case leaders defined the next steps on their pilots, which mainly involves meaningful data and data sources sharing, related to pathogen prevention in poultry, pesticide use in agriculture, and anticipation of regulatory changes.
Agroknow discussed what steps it will take to develop a causal inference network for the identification of salmonella contamination root cause analysis and an AI-assisted web crawler that automatically detects web pages containing information about food safety incidents. Stockholm University aims to take the crawler a step further, by enhancing this for broader data collection and more inclusive information gathering. Apart from that, the plan for the upcoming months of Stockholm University also involves creating datasets and baselines for future EFRA deliverables in NLP (Natural language processing), collaborating with Agroknow, CNR, Maize and SGS Digicomply.
On the second day, the discussion commenced with WFSR (Wageningen Food Safety Research) taking the lead, explaining their next course of action in the federated learning AI training concept, ensuring the input needed for algorithm modelling is used at the data source ensuring data privacy. The final goal of WFSR is to develop models, which identify and predict long-term, systemic, unknown risks in the food supply chain, with limited data sharing. With the main focus on data, Maize shared with the consortium members the outcomes they aim to achieve in terms of data source reporting and the development of a video crawler, capable of scouting video channels and collecting data, relevant to food safety. In addition, Maize will work on the architecture design of both the Analytics Powerhouse and Data Hub of the EFRA platform.
Rainno, focused on communication, community building & dissemination, by sharing the marketing goals for the upcoming months and providing training on the proper DEC activities reporting process. Finally, CNR will investigate new AI models and techniques that are both explainable to the end user as well as very efficient in terms of energy consumption. Also, they will research alternative AI learning approaches, allowing more effective and energy-efficient predictions.
The EFRA consortium’s first plenary meeting in Pisa was a successful gathering of all nine partners. The event focused on establishing goals and contributions for the upcoming months, with the consortium members presenting their plans, highlighting initiatives and building fruitful collaborations. The meeting laid the groundwork for meaningful progress towards our goal to create the first analytics-enabled, secure-by-design, green data space for AI-enabled food risk prevention.
By Marilena Dimitrakopoulou, Research Project Manager, Agroknow. Last month, our project coordinator, Agroknow, travelled
By Tony Lindgren, Head of the Systems Analysis and Security Unit, Department of Computer
A Machine Learning Approach for automative fault nowcasting By John Pavlopoulos, assistant professor at