Explore the forefront of food safety innovation through an exclusive interview with Tony Lindgren, from Stockholm University. Discover the project’s objectives and achievements as we delve into their pioneering work in utilizing AI for proactive food safety measures.
Tony Lindgren is a senior lecturer and the Head of the Systems Analysis and Security Unit, at the Department of Computer and Systems Sciences, at Stockholm University (SU). His research interest focuses on Machine Learning and Constraint & Logic Programming, while he teaches Principles and Foundations of Artificial Intelligence (PFAI). For EFRA, he leads the SU team, which mainly works on the creation of hyper-targeted NLU micro-modules for the EFRA AI models.
How is the project coming along?
Good. We have so far recruited a Ph.D. student (Korbinian Randl) who is now working on the project and we have added an extra senior researcher (Aron Henriksson) assisting in the project. We have already finalized some research in the project, for example, we presented a paper, named “Automotive fault nowcasting with machine learning and natural language processing”, at the European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases (ECML PKDD) held in Turin last year, and the paper was published in the Journal of Machine Learning. In addition to that we have released a pre-print of our latest planned publication that has been submitted to a conference, but is not yet accepted.
"Automotive fault nowcasting with machine learning and natural language processing"
"Faster, better, safer – new machine learning methods revolutionize complex systems"
What is the overall goal of the project?
I would say that the overall goal of the project is to explore new technologies, especially various forms of AI methods, to get insights and actionable intelligence when it comes to food safety issues.
That is a little bit vague, could you elaborate?
Sure, there exist data sources related to food safety, be it from companies, authorities as well as the public, for example from social media. This data has the potential to be used for proactively enhancing food safety not as is usually the case today in a reactive context. The EFRA project, where EFRA stands for Extreme Food Risk Analytics, aims to push the boundaries of previous usage of such data by exploring novel, experimental, and promising approaches in extreme data mining, aggregation, and analytics technologies. Funded by the EU, we aim to thoroughly test methods and tools that may significantly help address the grand scientific, economic, and societal challenges associated with the safety and quality of the food that European consumers eat.
You work at the Department of Computer and System Sciences at Stockholm University. I’m curious about what your team is working on when it comes to the EFRA context and its goals. Data science is nothing I as a layman would associate with food safety.
Well, you know Data science is in every science these days! We do not bring any expert knowledge from the domain of food safety, but in the EFRA project, we have partners that have that expertise. What we bring to the table is state-of-the-art knowledge from the field of Data Science, especially in the field of Natural language Processing (NLP), which has had a huge rise in attention since the success of Large Language Models like Chat GPT.
What are you working on at the moment?
We are currently planning for a study where we aim to create a model for early warning of food hazards. The idea is to use public data sources, as I mentioned earlier, for example, authorities that cover outbreaks of different food hazards together with data from social media. By aligning the different datasets, we can hopefully find a model for the early detection that a food hazard is imminent. At the moment we have the data sources needed for this study and have begun the preliminary work of aligning the data.
